text
stringlengths 87
880k
| pmid
stringlengths 1
8
| accession_id
stringlengths 9
10
| license
stringclasses 2
values | last_updated
stringlengths 19
19
| retracted
stringclasses 2
values | citation
stringlengths 22
94
| decoded_as
stringclasses 2
values | journal
stringlengths 3
48
| year
int32 1.95k
2.02k
| doi
stringlengths 3
61
| oa_subset
stringclasses 1
value |
|---|---|---|---|---|---|---|---|---|---|---|---|
==== Front
Lipids Health DisLipids in Health and Disease1476-511XBioMed Central London 1476-511X-4-111594387710.1186/1476-511X-4-11ResearchComparison between swallowing and chewing of garlic on levels of serum lipids, cyclosporine, creatinine and lipid peroxidation in Renal Transplant Recipients Jabbari Abbas [email protected] Hassan [email protected] Amir [email protected] Reza [email protected] Clinical pharmacy laboratory, Drug Applied Research Center, Tabriz University of Medical Sciences, University Ave., Tabriz, Iran2 Clinical pharmacy laboratory, Drug Applied Research Center, Tabriz University of Medical Sciences, University Ave., Tabriz, Iran3 Biochemistry and Drug Metabolism Laboratory, Drug Applied Research Center, Tabriz University of Medical Sciences,, University Ave., Tabriz, Iran4 Nutrition Laboratory, Drug Applied Research Center, Tabriz University of Medical Sciences, University Ave., Tabriz, Iran2005 19 5 2005 4 11 11 1 3 2005 19 5 2005 Copyright © 2005 Jabbari et al; licensee BioMed Central Ltd.2005Jabbari et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Abstract Hyperlipidemia and increased degree of oxidative stress are among the important risk factors for Atherosclerosis in renal transplant recipients (RTR). The Medical treatment of hyperlipidemia in RTR because of drugs side effects has been problematic, therefore alternative methods such as using of Garlic as an effective material in cholesterol lowering and inhibition of LDL Oxidation has been noted. For evaluation of garlic effect on RTR, 50 renal transplant patients with stable renal function were selected and divided into 2 groups. They took one clove of garlic (1 gr) by chewing or swallowing for two months, after one month wash-out period, they took garlic by the other route. Results indicated that although lipid profile, BUN, Cr, serum levels of cyclosporine and diastolic blood pressure did not change, Systolic blood pressure decreased from138.2 to 132.8 mmHg (p=0.001) and Malondialdehyde (MDA) decreased from 2.4 to1.7 nmol/ml (p=0.009) by swallowing route, Cholesterol decreased from 205.1 to 195.3 mg/dl (p=0.03), triglyceride decreased from 195.7 to 174.8 mg/dl (p=0.008), MDA decreased from 2.5 to 1.6 nmol/ml (p=0.001), systolic blood pressure decreased from 137.5 to 129.8 mmHg (p=0.001), diastolic blood pressure decreased from 84.6 to 77.6 mmHg (p=0.001) and Cr decreased from 1.51 to 1.44 mg/dl (p=0.03) by chewing route too. However HDL, LDL and cyclosporine serum levels had no significant differences by both of swallowing and chewing routes. We conclude that undamaged garlic (swallowed) had no lowering effect on lipid level of serum. But Crushed garlic (chewed) reduces cholesterol, triglyceride, MDA and blood pressure. Additionally creatinine reduced without notable decrease in cyclosporine serum levels may be due to cyclosporine nephrotoxicity ameliorating effect of garlic.
==== Body
Introduction
Cardiovascular disease (CVD) is the main cause of morbidity and mortality in renal transplant recipients [1] and it accounts for about 40% of deaths in this group of patients [2]. Several risk factors for CVD, such as lipid abnormalities and insulin resistance, may partly explain the accelerated development of atherosclerosis following renal transplantation [1]. In atherogenesis process, macrophages oxidate LDL and produce foam cells by OX-LDL that are characteristic for atherosclerosis [3]. Lipoprotein abnormalities are common in renal disease [4] these are reported in 50–80% of renal transplant recipients [5].
Pathogenesis of hyperlipidemia in Renal transplant recipients (RTR) is not fully understood, but dosage of steroid, consumption of cyclosporine, anti-hypertensive medication, rising of serum creatinine, proteinuria and diabetes mellitus have been considered [2].
There is increased degree of oxidative stress in RTR [6]. In these patients following hyperlipidemia and lipoprotein abnormalities, free radicals produced. Also immunosuppressive drugs can induce photosensitization reactions. These reactions lead to production of free radicals and aggravate lipid peroxidation [7].
The medical treatment of hyperlipidemia in transplant recipients is problematic [2]. Bile binding resins interfere with cyclosporine absorption and also lead to hyper triglyceridemia. Nicotinic acid causes hyperglycemia and rising of uric acid. Fibric acids cause myopathy, dyspepsia, gall bladder stones and rises of creatinine. Statins can induce hepatotoxicity, myopathy rhabdomyolysis (Especially if accompanied with cyclosporine) [2,4,5]. So alternative methods are deserved importance.
Garlic has been used for centuries as an herbal medicine in treating abscesses, cough, poisoning, parasites, worms, digestive and circulatory problems, snake bites [8] hemorrhoids, abdominal pain, loss of appetite and pneumonia [9]. Epidemiologic studies suggest that consumption of garlic may protect against carcinogenesis. In particular, the development of gastric and colorectal cancers seems to be prevented by alliums consumption [10,11]. Also garlic was known as an effective material in decreasing of blood pressure [12] and cholesterol [13,14] also can inhibit LDL oxidation [15-17], platelet aggregation and adhesion [18,19] and can increase Nitric oxide production [20]. Because of these beneficial effects of garlic, we decided to study the effect of garlic on lipid profile, lipid peroxidation and cyclosporine serum level and because of its irritant odor in chewing we chose two routes of swallowing and chewing for comparing of their efficacy.
Method and materials
50 renal transplant recipients with stable renal function (based on serum Cr<1.8 mg/dl) were selected randomly (hyperlipidemic and normolipidemic) (Table 1)(see additional file 1, Table1, word). Patients had been transplanted more than 1 year and were treating with triple immunosuppressive regimen including of: cyclosporine, prednisolone and azathioprine / or mycophenolate mofetil. Some of patients were under hypolipemic agents and antihypertensive drugs. Drug regimen did not change in 2 months period before and during the study.
We divided patients into 2 groups randomly (A, B). Group A patients took one clove of raw aged garlic (1 gr) by swallowing and group B patients by chewing the same amount. At the start of study we checked their dietary regimen including of: intake of calorie, total fat (TF), saturated fatty acids (SFA), monounsaturated fatty acids (MUFA), polyunsaturated fatty acids (PUFA) and cholesterol. We checked also some clinical and Para clinical parameters including of: blood pressure, weight, triglyceride (TG), cholesterol (Chol), low density lipoprotein (LDL), high density lipoprotein (HDL), serum malondialdehyde (MDA), blood urea nitrogen (BUN), creatinine (Cr) and cyclosporine serum level. Patients took garlic for 2 months then we checked above parameters again. After one month wash-out period the patients took garlic by the other route (swallowing changed to chewing and vice versa) with the same condition (as a cross over design).
Lipid profiles (TG, Chol, LDL, HDL), BUN and Cr were measured by standard enzymatic method, MDA determined by colorimetric method using thiobarbituric acid reactions [21] and cyclosporine serum level measured by RIA.
Data were analyzed by paired sample t test and Non parametric 2 related sample test using SPSS11.5 program. A difference was considered statistically significant when the P value was <0.05.
Results
In this study 2 patient (group A) because of heart burn, 1 patient (group A) because of bloating, 3 patients because of change in drug regimen were excluded (group B) 44 patients continued the study (22 patients in each group).
In patients who swallowed garlic, weight, intake of calorie, SFA, MUFA, PUFA had no significant differences but intake of TF and Chol were increased during the study as compared with the pre garlic period. In patients who chewed garlic, weight, intake of calorie did not change however intake of TF, SFA, MUFA, PUFA and Chol increased during the study (table 2) (see additional file 2, Table 2, word).
Comparison between results of Chewing and Swallowing of garlic indicate that there is significant differences in diastolic blood pressure (P = 0.016), triglyceride (P = 0.008) and cholesterol (P = 0.04) but not in systolic blood pressure (P = 0.187), HDL (P = 0.925), LDL (P = 0.354), MDA (P = 0.587), BUN (P = 0.657), Cr (P = 0.119) and cyclosporine serum level (P = 0.155). Other data are provided in table3 (see additional file 3, Table3, word).
Conclusion
Several factors after transplantation produce hyperlipidemia include: weight gain and increase body fat mass due to appetite improvement [22].
Lopes et al reported that moderate energy restriction of about 30% and reducing fat in diet decreased cholesterol and LDL [23] however in our study dietary intake not only did not decreased but it increased significantly.
Adler and holub showed that LDL were reduced (14.2%) and total cholesterol were significantly lower (11.5%) with taking 900 mg garlic/day for 12 weeks in hypercholesterolemic men [24].
Tohidi and Rahbani showed that taking 1200 mg garlic powder for 4 weeks reduced total cholesterol (9%), triglyceride (11%), LDL (15%), systolic blood pressure (3%) and diastolic blood pressure (2%) [15].
Steiner et al with giving 7.2 gr aged garlic extract (AGE) for 4 weeks indicated reduction in cholesterol (6.1%), LDL (4%), systolic and diastolic blood pressure (5.5%) [14].
Issacsohn et al reported that taking 900 mg garlic powder for 12 week did not change in lipid profile [8].
Mader showed that taking 800 mg garlic powder for 4 months reduced cholesterol (9%) and triglyceride (15%) [25].
Lash et al reported that taking garlic tablets at a dose of 680 mg two times a day for 6, 12 weeks decreased LDL (6%) and total cholesterol (4%) significantly in hypercholesterolemic renal transplant patients [2].
Brinker mentioned that Effectiveness might be decreased by garlic's ability to induce metabolism and decrease levels of drugs like cyclosporine which are substrates of cytochrome P450 3A4. It can potentially cause transplant rejection [26].
Blech et al showed progression of atherosclerosis is relevant with oxidative stress and indicated with measurement of MDA [27].
MDA is important marker of lipid peroxidation [28] and reported that serum MDA decreased by antioxidants consumption [29].
In this study after chewing of garlic, systolic and diastolic blood pressure decreased 5%, 8%, like as Steiner and Tohidi studies. Also cholesterol and triglyceride reduced 4%, 10%, according to Mader and Tohidi studies but less than these studies. Decreased Cholesterol in this study was like as Lash study however he did not report triglyceride reduction after garlic consumption. HDL did not change in all studies. Cyclosporine serum level decreased not significantly in our study and as we known it is the first study which demonstrated the effect of garlic on cyclosporine level in RTR.
Ingestion of garlic by chewing (or crushed garlic) can reduce cholesterol, triglyceride, MDA, systolic and diastolic blood pressure even in the presence of increasing fat intake. But undamaged garlic (swallowed) had no significant effect on serum lipids (TG, Chol, LDL and HDL), diastolic blood pressure, and BUN, Cr and cyclosporine serum level. Our hypothesis is that it is because of inability of Alliin to convert to Allicin. So the specific garlic odor is a hallmark of releasing of Allicin.
Acute cyclosporine nephrotoxicity is predominantly functional in that it produces no particular histological features. It is probably the result of renal arteriolar constriction, a feature that has been demonstrated in animals. In chronic cyclosporine nephrotoxicity the principal injury is to the small arterioles where there is vacuolation of smooth muscle and endothelial cells [30].
Reducing creatinine without notable decrease in cyclosporine serum level by chewing of garlic may be cyclosporine nephrotoxicity protecting effect of garlic as its effect on gentamycin nephrotoxicity due to its antioxidant effect [31] or calcium channel blockers like effects [32] or nitric oxide increasing vasodilatations [20].
Additional studies are necessary in order to investigate effect of garlic on cyclosporine serum levels, nephrotoxicity and serum creatinine.
In conclusion we found that crushed garlic (chewed) reduced total cholesterol, triglyceride, MDA (lipid peroxidation) and blood pressure that have important role in cardiovascular disease. Therefore garlic consumption can prevent this disease.
Supplementary Material
Additional File 1
Table 1
Click here for file
Additional File 2
Table 2
Click here for file
Additional File 3
Table 3
Click here for file
==== Refs
Emilio R Domingo Gonzalez L Juan Carlos R Gema Fernandez F Debora I Julio Gonzalez C Apolipoprotein C-III and E Polymorphisms and Cardiovascular Syndrome, Hyperlipidemia, and Insulin Resistance in Renal Transplantation Am J Transplantation 2002 2 343 348 10.1034/j.1600-6143.2002.20409.x
Lash JP Cardoso LR Mesler PM Walczak DA Pollak R The Effect of Garlic on Hypercholesterolemia in Renal Transplant Patients Transp proce 1998 30 189 191 10.1016/S0041-1345(97)01227-X
Porek C Crawford D Yagi K Molecular mechanisms of oxidant induced transformation Active oxygen's, lipid peroxides and antioxidants 1993 1 Scientific societies press, Japan 189 194
McKay DB Milford EL Tolkoff NE Rubin Brenner BM, Levine SA Clinical Aspects of Renal Transplantation The Kidney 2000 2 6 WB Saunders, USA 2584
Braun WE Johnson RJ, Feehally J The Medical Management of Renal Transplant Recipient Comprehensive Clinical Nephrology 2004 2 Mosby, Spain 1109 10
Simic-Ogrizovic S Simic T Reljic Z Blagojevic R Radivojevic D Lezaic V Markers of Oxidative Stress after renal transplantation Transpl Int 1998 11 125 129 10.1007/s001470050443
Jhan K Belign M McClean F A Pharmacist's guide to cholesterol issues NARD Journal 1991 37 45
Isaacsohn JL Moser M Stein EA Dudley K Davey JA Liskov E Black HR Garlic Powder and Plasma Lipids and Lipoproteins Arch Intern Med 1998 158 1189 1194 9625398 10.1001/archinte.158.11.1189
Aouadi R Aouidet A Elkadhi A Ben Rayana C Jaafoura H Tritar B Nagati K Effect of Fresh Garlic (Allium Sativum) on Lipid Metabolism in Male Rates Nutrition Res 2000 20 273 280 10.1016/S0271-5317(99)00159-1
Agarwal KC Therapeutic actions of garlic constituents Med Res Rev 1996 16 111 24 Review 8788216 10.1002/(SICI)1098-1128(199601)16:1<111::AID-MED4>3.0.CO;2-5
Dorant E van den Brandt PA Goldbohm RA A prospective cohort study on the relationship between onion and leek consumption, garlic supplement use and the risk of colorectal carcinoma in The Netherlands Carcinogenesis 1996 17 477 84 8631133
Silagy CA Neil HAW A Meta – analysis of the effect of garlic on blood pressure J Hypertens 1994 12 463 468 8064171
Stevinson C Pittler H Edzard E Garlic for Treating Hypercholesterolemia Ann Intern Med 2000 133 420 429 10975959
Steiner M Khan AH Holbert D Lin RI A double-blind crossover study in moderately hypercholesterolemic men that that compared the effect of aged garlic extract and placebo administration on blood lipids Am J Clin Nutr 1996 64 866 870 8942410
Tohidi m Rahbani m Evaluation of the effect of garlic powder on blood pressure, serum lipids and lipoproteins Pharmacy journal of Tabriz Univ Med Sci 2000 4 16 20
Mikhin VP Gormnatskii NI The effect of Garlic preparation alisat on the level of Lipid peroxidation Products. Activity of various antioxidant enzymes and blood levels of lipoproteins in patients with atherosclerosis Ball-EK sp-Biol-med 1996 1229 502 504
Heiner K Thomas S Effect of a Garlic oil preparation on serum lipoproteins and cholesterol metabolism JAMA 1998 279 1900 1904 9634262 10.1001/jama.279.23.1900
Rahman K Billington D Dietary Supplementation with Aged Garlic Extract Inhibits ADP-Induced Platelet Aggregation in Humans J Nutr 2000 130 2662 2665 11053504
Steiner M Li W Aged Garlic Extract, a Modulator of Cardiovascular Risk Factors: A Dose- Finding Study on the Effects of AGE on Platelet Functions J Nutr 2001 131 980S 984S 11238801
Morihara N Sumioka I Moriguchi T Uda N Kyo E Aged Garlic Extract enhances production of nitric oxide Life Sci 2002 71 509 517 12052435 10.1016/S0024-3205(02)01706-X
Yagi K Lipid peroxides and related radicals in clinical medicine Free Radicals in Diagnostic Medicine 1994 1 Plenum Press, New York 1 15
Appel G Lipid abnormalities in renal disease Kidney Int 1991 39 169 2002630
Lopes IM Martin M Errasti P Martinez JA Benefits of a dietary intervention on weight loss, body composition, and lipid profile after renal transplantation Nutrition 1999 15 7 10 9918055 10.1016/S0899-9007(98)00137-3
Adler AJ Holub BJ Effect of garlic and fish-oil supplementation on serum lipid and lipoprotein concentrations in hypercholesterolemic men Am J Clin Nutr 1997 65 445 450 9022529
Mader FH Treatment of hyperlipidemia with garlic-powder tablets: evidence from the German Association of General Practitioner's multicentric placebo-controlled double-blind study Arzneimittel Forschung Drug Re 1990 40 3 8
Brinker F Herb Contraindications and Drug interactions 1998 2 Sandy (OR), Eclectic Med Publications 105
Belch JJ Mackay IR Hill A Jening SP Oxidative stress is present in atherosclerotic peripheral arterial disease and further increased by diabetes mellitus Int-Angiol 1995 14 385 8708432
Raharjo S Sofos JN Schmitt GR Solid phase acid extraction improves thiobarbituric acid methods to determine lipid oxidation J Food Sci 1993 58 921 932
Eander RC Tocopherol influences in vivo induce of lipid peroxidation in post menopausal woman given fish oil J Nutr 1996 126 643 652 8598549
Salaman J Feduska N Fabre JW Williams JD, Asscher AW, Moffat DB, Sanders E Renal transplantation clinical atlas of the kidney 1991 1 Lippincott, Singapore 14.27 14.28
Perdaza Chaverri J Maldonado PD Medina-Campos ON Olivares-Corichi IM Granados-Silvestre MLA Hernandez-Pando R Ibura-Rubio ME Garlic ameliorates Gentamycin Nephrotoxicity: Relation to antioxidant enzymes Free Radical Biology and Medicine 2000 29 602 611 11033412 10.1016/S0891-5849(00)00354-3
Mehrens T Thiele S Suwelack B Kempkes M Hohage H The beneficial effects of calcium channel blockers on long-term kidney transplant survival are independent of blood-pressure reduction Clin Transplantation 2000 14 257 261 10.1034/j.1399-0012.2000.140313.x
| 15943877 | PMC1173136 | CC BY | 2021-01-04 16:39:19 | no | Lipids Health Dis. 2005 May 19; 4:11 | utf-8 | Lipids Health Dis | 2,005 | 10.1186/1476-511X-4-11 | oa_comm |
==== Front
Microb Cell FactMicrobial Cell Factories1475-2859BioMed Central London 1475-2859-4-151592151810.1186/1475-2859-4-15ResearchIndustrial-scale production and purification of a heterologous protein in Lactococcus lactis using the nisin-controlled gene expression system NICE: The case of lysostaphin Mierau Igor [email protected] Peter [email protected] Swam Iris [email protected] Barry [email protected] Esther [email protected] James [email protected] Eddy J [email protected] NIZO food research, P.O. Box 20, 6710 BA Ede, The Netherlands2 Biosynexus, Inc., 9119 Gaither Road, Gaithersburg, MD 20877, USA2005 27 5 2005 4 15 15 18 4 2005 27 5 2005 Copyright © 2005 Mierau et al; licensee BioMed Central Ltd.2005Mierau et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The NIsin-Controlled gene Expression system NICE of Lactococcus lactis is one of the most widespread used expression systems of Gram-positive bacteria. It is used in more than 100 laboratories for laboratory-scale gene expression experiments. However, L. lactis is also a micro-organism with a large biotechnological potential. Therefore, the aim of this study was to test whether protein production in L. lactis using the NICE system can also effectively be performed at the industrial-scale of fermentation.
Results
Lysostaphin, an antibacterial protein (mainly against Staphylococcus aureus) from S. simulans biovar. Staphylolyticus, was used as a model system. Food-grade lysostaphin expression constructs in L. lactis were grown at 1L-, 300-L and 3000-L scale and induced with nisin for lysostaphin production. The induction process was equally effective at all scales and yields of about 100 mg/L were obtained. Up-scaling was easy and required no specific effort. Furthermore, we describe a simple and effective way of downstream processing to obtain a highly purified lysostaphin, which has been used for clinical phase I trials.
Conclusion
This is the first example that shows that nisin-regulated gene expression in L. lactis can be used at industrial scale to produce large amounts of a target protein, such as lysostaphin. Downstream processing was simple and in a few steps produced a highly purified and active enzyme.
==== Body
Background
Lactococcus lactis is a Gram-positive bacterium that is widely used in food production such as cheese and butter manufacturing [1]. In the last two decades the physiology and genetics of this bacterium have been thoroughly characterized [2]. At present several genomes are either completely sequenced or close to completion [3,4]. Furthermore, L. lactis is easily genetically accessible and a wide variety of genetic tools have been developed [5]. Because of the genetic accessibility and the ease of its handling, a variety of new applications have been developed. Examples are the expression of cytokines and bacterial or viral antigens [6,7], enzymes [8], membrane proteins [9] and metabolic transformations [10]. These studies show that L. lactis is a suitable host for applications beyond its traditional use in food fermentations.
One of the crucial developments has been the construction of a food grade [11] and regulated gene expression system based on the regulation mechanism of the nisin A operon of L. lactis. In this operon the gene product nisin (a 34 amino acid bacteriocin) activates its own transcription at ng/ml amounts [12]. The elements of this regulatory system have been isolated and inserted in a suitable host strain, constituting the powerful regulated gene expression system NICE (NIsin-Controlled gene Expression system) [5,13]. The NICE system is widely used on laboratory scale for research and for over-expression of genes of interest [8,9,14]. However, experience in large scale application and fermentation of the NICE system is very limited.
Lysostaphin is a 25-kD antibacterial protein, produced by Staphylococcus simulans biovar. Staphylolyticus, that can hydrolyze the Gly-Gly bonds in the cell wall of the pathogens S. aureus and S. epidermidis and thus lyse these bacteria [15,16]. Lysostaphin has been proven to be an effective agent against the widespread hospital infectious agent S. aureus [17-19].
In the present study we describe for the first time a large-scale (3000 L) regulated gene expression process for the production of a heterologous protein – lysostaphin – in L. lactis using the NICE expression system. Furthermore, we also describe the purification of lysostaphin from the 3000-L fermentation batches resulting in a 90% pure pharmacological intermediate that has been further purified, formulated and used in clinical phase I studies.
Results
Construction of a lysostaphin expression system in L. lactis
The coding sequence of the lysostaphin gene of S. simulans biovar. Staphylolyticus, lacking its first two alanine residues [15], was cloned after PCR amplification into the NICE vector pNZ8148, resulting in plasmid pNZ1709 (see Table 1). The obtained construct was verified by nucleotide sequencing. Subsequently, the chloramphenicol-resistance cassette was exchanged for the lacF gene of L. lactis [11], leading to pNZ1710 (Fig. 1A). In this plasmid expression of the lysostaphin gene lss is under control of the nisin-inducible nisA promoter. Furthermore, this plasmid is selected by a food-grade mechanism, i.e. growth on lactose, and does not contain an antibiotic-resistance gene or its remnants. Lysostaphin, a 25-kD antibacterial protein, was produced in the cytoplasm of the cells. Figure 1B shows an SDS-PAGE image of the intracellular soluble protein fractions of NZ3900 (pNZ1710) before and after induction with nisin. After induction with nisin, lysostaphin is accumulated in the cell to about 10% of the soluble protein fraction, as estimated from SDS-PAGE (Figure 1B). Lysostaphin was isolated, purified (see below) and its N-terminal amino acid sequence was determined to be Thr-His-Glu-His-Ser-Ala [15]. This indicates that the correct protein was produced, that no significant intracellular degradation occurred and that the N-terminal formyl-methionine residue was removed.
Table 1 Strains and Plasmids
Bacterial strain/plasmid Properties Reference
Bacteria
Lactococcus lactis subsp. cremoris NZ3900 Integration of nisRnisK in the chromosome; integration of the lac operon in the in the chromosome and deletion of lacF resulting in a lactose- negative host strain that can be complemented by lacF [25]
Plasmids
pNZ8148 PnisA, CmR; replicon of rolling circle plasmid pSH71, basic NICE vector, derivative of pNZ8048 [12] [26]
pNZ1709 Lysostaphin gene under control of PnisA, CmR; derivative of pNZ8148 This work
pNZ1710 Lysostaphin gene under control of PnisA, lacF; expression vector for food- grade expression, selection on growth with lactose; derivative of pNZ1709 This work [11]
Figure 1 Plasmid bearing the lysostaphin gene and over expression of this gene. A, plasmid construct for nisin-controlled lysostaphin production. PnisA, nisin-controlled promoter; matLss-2A, coding sequence for mature lysostaphin lacking the first 2 alanine residues; Term., transcription terminator; repC and repA, replication genes; lacF, food-grade lactose selection marker. B, SDS-PAGE showing the intracellular production of lysostaphin upon induction with nisin. 1, molecular weight marker; 2, cell extract without nisin-induction; 3, cell extract with induction with 10 ng/mL nisin; Lss, lysostaphin
Development of a fermentation medium and an induction scheme for lysostaphin production
In the development of human and animal pharmaceuticals it is important that the product is guaranteed BSE (Bovine Spongiform Encephalomyelitis) agent-free. All commercially available pre-formulated media for lactococci [e.g. M17 [20] contain components of animal origin such as meat extract and are possible sources for the BSE agent. Therefore, a new medium based on hydrolysed plant protein and yeast extract was developed. The key components of that medium are an entirely plant-based peptone and a yeast extract that is clear in a solution of at least 1%. The peptone chosen, was made from soy protein digested with the plant derived proteinase papain. Additionally, the medium contained 5% lactose (certified BSE free) to allow unlimited growth under pH-regulated growth conditions. Finally, Mg2+ and Mn2+ were added as known growth enhancers for lactic acid bacteria [20]. For details of the composition and sterilization of the medium see Methods.
Induction of lysostaphin production was carried out at an optical density at 600 nm of about 1 (light path 1 cm) (mid exponential growth phase) (0.3 g/L cell dry weight [21]) by adding nisin (Figure 2). After induction, lysostaphin production proceeded for 6 – 8 hours. Figure 2 shows that upon induction, growth of the culture is severely inhibited. This is likely the result of lysostaphin accumulation in the cell that appears to have growth inhibiting properties (viable plate counts drop within 20 min after induction 3–4 orders in magnitude).
Figure 2 Growth characteristics of 1-L culture. Culture at 1-L scale of L. lactis NZ3900 containing the lysostaphin expression plasmid pNZ1710 with and without induction by 10 ng/mL nisin.
Scale-up of lysostaphin production to 300 L and 3000 L
Lysostaphin production was scaled up from 1-L scale to 300-L scale and eventually 3000-L fermentations. Growth and induction conditions found at laboratory scale were directly transferred to the 300-L and 3000-L scale: Induction at an optical density OD600 = 1 with 10 ng/ml nisin. Details of media preparation for the larger scales are described in Methods. Figure 3 shows the growth characteristics of induced cultures of L. lactis carrying the plasmid pNZ1710 at all three scales. All three cultures behaved very similarly, despite the difference in scale of more than 3 orders of magnitude. The lysostaphin yields of the 1-L and 3000-L fermentations after induction were approximately 100 mg/L. Four consecutive 3000-L production runs were carried out to produce raw material for further down stream processing. The growth characteristics of the four fermentation runs were virtually identical. In each run approximately 300 g lysostaphin (100 mg/L) were produced and subsequently used as starting material for purification.
Figure 3 Comparison of growth characteristics of 1-L, 300-L and 3000-L cultures. Culture of L. lactis NZ3900 containing the lysostaphin expression plasmid pNZ1710 at 1-L, 320-L and 3000-L scale with induction by 10 ng/mL nisin. As comparison, growth of an uninduced culture at 1-L scale is shown.
Downstream processing of the 3000-L lysostaphin production batch
A downstream processing protocol was designed for the preparation of a pharmaceutical intermediate that could be used for further purification and formulation to carry it into clinical phase I trials.
The basic operations were concentration and washing of the cells, destruction of the cells by homogenization, removal of the cell debris, and capturing of the lysostaphin by chromatography.
The fermenter content was concentrated about 20-fold by filtration (tangential flow) over a 0.8 μm ceramic membrane (3.8 m2) and then 200% diafiltrated to remove most of the residual medium components. The retentate was subjected to continuous homogenization at 1400 bar, 80 L/h. The homogenization procedure was repeated three times to ensure complete release of the intracellular lysostaphin produced.
The cell debris was then separated from the intracellular fraction by filtration (tangential flow) over a 0.8 μm ceramic membrane (3.8 m2). The homogenate was first concentrated to 100 L and then 300 % diafiltrated to wash out any lysostaphin associated with the cell debris. The resulting lysostaphin-containing filtrate of about 400 L was stored frozen before loading onto the chromatography column.
Capturing lysostaphin
A cation-exchange chromatography capture step was selected based on the relatively alkaline isoelectric point (pH 9.5) of lysostaphin [16]. Because of superior performance, the strong exchanger SP-Sepharose FF was chosen over the weak exchanger CM-Sepharose FF. Optimum lysostaphin binding was found at pH 7.5 in phosphate buffer. Lysostaphin was eluted using a NaCl step gradient at 0.5 M NaCl with the same pH and phosphate buffer concentrations (Methods) as used for loading (Figure 4A and 4B show an elution profile and SDS-PAGE analysis). Since maximum binding of lysostaphin was hindered by unknown components in the cell extract, the flow-through was re-fed to the column twice to capture more than 90% of lysostaphin from the cell extract. Finally, the eluate was diluted and all captured lysostaphin was applied to the column at once for an additional recapture step (Methods). The resulting material was ca. 90% pure lysostaphin as determined by SDS-PAGE analysis (Methods). The lysostaphin production yield in the fermentation was about 100 mg/l. Therefore about 300 g lysostaphin had been produced in each 3000-L fermentation run. The mean total yield of the downstream process was about 120 g, resulting in 40% recovery of the originally produced lysostaphin.
Figure 4 Purification of the overproduced lysostaphin. A, Typical chromatogram of a lysostaphin capture step. NaCl concentration (brown line) and absorption at 280 nm (blue line) are indicated. Fractions that were analysed by SDS-PAGE are indicated by grey bars underneath. The lysostaphin fraction (F6) is indicated with a black bar. B, SDS-PAGE analysis of the different fractions of the capture chromatography. 1, molecular weight marker; 2, cell extract before loading; 3, flow-through fraction (F3); 4, 5 and 6, fractions F4, F5 and F7; 7, lysostaphin fraction F6; 8, 9 and 10, fraction F6 diluted 1:2, 1:4 and 1:8.
Discussion
Nisin-regulated gene expression in Lactococcus lactis has been shown to be an effective and multifunctional tool [2,5,7,9]. L. lactis has numerous characteristics that make it an interesting host for industrial-scale heterologous protein production: it is 100% food grade, including plasmid selection systems, it produces no endotoxins or other toxic substances, it produces no inclusion bodies and no spores, it is grown in simple non-aerated, stirred fermentations and produces no extracellular proteinases. With these characteristics, it can be used for food applications, as a source of enzymes, for metabolic transformations and for the production of biologicals. The fact that L. lactis is food grade also means that cellular components or enzymes can be used with only partial purification or directly in the crude cell extract without further purification. Despite 10 years of laboratory use of the NICE system, we are not aware of any reports that describe the step to large-scale industrial application of this tool. This paper describes the successful development and scale-up of a process for the production of a heterologous protein using lysostaphin as an example. Lysostaphin is an antibacterial protein from S. simulans biovar. Staphylolyticus, that has potential in topical and systemic applications for the treatment of Staphylococcus aureus infections [17-19]. Lysostaphin is a clear example for the benefits of a regulated expression system, since constitutive intracellular lysostaphin expression leads to rapid cell death. Another advantage of regulated gene expression is that the cells can be pre-grown to a certain cell density before the energetically costly production of a foreign protein is switched on.
Furthermore, a food-grade plasmid selection system based on lactose consumption has been used, making the use, detection and tracing of antibiotics unnecessary. This food-grade construct allowed the production of a pharmaceutical intermediate in a cheaper food-grade production plant rather than in a cGMP facility. Similarly, such a process could be used to make other industrially interesting intermediates such as enzymes for food and (bio)chemical applications, probiotic preparations, etc.
The process described in this paper was first developed at 1-L scale and subsequently transferred to the 300-L and 3000-L scale. This scale-up was without problems, without specific calculations and without changes in the available equipment. L. lactis is a simple fermentative, oxygen-tolerating bacterium converting lactose or glucose into lactic acid. Therefore, no oxygen transfer is needed during the fermentation. The only condition that needs to be met is appropriate mixing of the whole culture to ensure evenly distributed nutrients and effective distribution of nisin for the induction of gene expression. One significant difference exists in the addition of nisin to the culture at different scales. At laboratory scale this addition takes a few seconds, while at 3000-L scale it takes 2 – 5 min. Despite the difference in addition times, no adverse effects on the induction process and lysostaphin production were observed.
We carried out four consecutive fermentations at the 3000-L scale. The growth of the culture in all four fermentation runs was nearly identical, as was the final yield of lysostaphin after purification (about 120 g per batch), indicating the biological and technical robustness of the process.
The down-stream processing shows that L. lactis can effectively be separated from the fermentation medium and that the cell content can be released by high-pressure continuous homogenization. In the present process three passages for complete destruction of the cells are used. Preliminary results show that two passages may be enough. The cell debris can simply be removed by a second filtration step in the same equipment that is used for the cell separation. After separation of the cell content from the debris, the product can be used as crude extract with enriched enzyme activity or different routes can be followed to further purify or enrich the active component: ultrafiltration, selective precipitation and chromatography. In the present process, lysostaphin was captured on SP-Sepharose FF in two consecutive steps. We found that an as yet unknown component in the cell extract hampered binding of lysostaphin to the resin. This obstacle was overcome by repeated loading of the flow-through to the same column. The solution of captured lysostaphin apparently lost the interfering component and could be completely bound to the resin in one recapture run. Further research is needed to identify the interfering compound and to design a process in which lysostaphin can be captured in a single run.
While 100 mg/L yield is relatively low for bulk productions, it may be acceptable for specialized productions. Careful optimization of growth and induction can lead to considerable yield increases. This was also demonstrated for lysostaphin were in a separate set of experiments lysostaphin production was optimized and a yield of 300 mg/L was reached [22]. One of the challenges is the development of a fermentation process in which this fermentative organism can be grown to higher cell densities.
The downstream process as a whole consists only of a few steps and is thereby simple, fast and highly reproducible. After further optimization it could easily be carried out within 48 h after the fermentation run, limiting the whole process to one week from seed culture to product.
Conclusion
The present publication describes for the first time that it is possible to use the lactic acid bacterium L. lactis for industrial scale heterologous protein production. Furthermore, we demonstrate that the widely used nisin-controlled gene expression system NICE is fully operative at this scale. This opens the way for a food grade alternative expression system to the commonly used host E. coli. Other advantages are that L. lactis does not produce endotoxins or inclusion bodies, and does not produce spores and extracellular proteinases. This opens up a wide range of pharmaceutical, cosmetics, biochemical, food, and feed applications.
Methods
Bacterial strains and plasmids
Table 1 shows the strain and the plasmids used in this study. The bacteria were maintained as frozen stock at -80°C.
Growth media and cultivation conditions
For the genetic construction work the bacteria were routinely grown in M17 medium [20] fortified with 1% glucose or 0.5% lactose and 5 μg/ml chloramphenicol for selection on chloramphenicol resistance.
For 1-L, 300-L and 3000-L fermentations the following medium was used: 5% lactose (pharmaceutical-grade lactose Lactochem 207, Borculo DOMO Ingredients, Zwolle, The Netherlands), 1.5% peptone from soy (VWR-Merck, product number 111932, Amsterdam, The Netherlands), 1% yeast extract (BioSpringer, product number 1105C0/180, Maisons Alfort, France), 1 mM MgSO4 (VWR-Merck), 0.1 mM MnSO4 (VWR-Merck). For the 1-L scale all ingredients were dissolved in water and the medium was sterilized for 20 min at 110°C. For the 300- and 3000-L scale all ingredients were dissolved in water and subsequently stream-sterilized (Crepaco, Bryan, Texas, U.S.A.) for 20 s at 140°C (batch-wise sterilization will cause chemical reactions in the medium that will inhibit growth of the bacteria). Bacteria were inoculated at 1 % and grown at 30°C. For 300- and 3000-L fermentations inoculum was prepared as follows. 2 mL frozen stocks were removed from -80°C freezer, inoculated into 100 and 300 ml fermentation medium, respectively, and grown under acidifying conditions as standing culture for 16 h. For the 300-L fermentation, 30 ml of this culture were inoculated into 3 L fermentation medium and cultivated under acidifying conditions as standing culture for 16 h. For the 3000-L fermentation, 300 mL were inoculated into 30 L fermentation medium in a 75-L fermentor and cultivated for 16 h with a stirring speed of 50 rpm under acidifying conditions.
Nisin induction
0.04% nisin powder (Sigma-Aldrich Chemie, Zwijndrecht, The Netherlands) was dissolved in 0.05% acetic acid and precipitated proteins were removed by centrifugation. Cells were grown to an optical density at 600 nm of 1 (light path 1 cm) (= 0.3 g/L cell dry weight [21]) at which point 10 ng/ml of nisin (final concentration) was added. Following the induction, the culture was incubated for 6 – 8 h before harvest. The 3000-L fermentations were terminated by rapid cooling to 4°C and stored for about 12 h at this temperature before further processing.
Molecular techniques
Standard genetic techniques were carried out according to Sambrook et al. [23]. SDS-PAGE was performed according to Laemmli [24]. N-terminal amino acid sequencing was outsourced (Protein Sequencing Laboratories, University of Leiden, Netherlands).
Detection and quantification of lysostaphin
Lysostaphin production was routinely monitored using SDS-PAGE. The activity of the enzyme was determined using the Staphylococcus carnosus cell wall degradation assay [15]. Lysostaphin was quantified using capillary electrophoresis as described in Mierau et al. [22]. Purity of lysostaphin was determined from scanning and analysis of SDS-PAGE gels with a PowerLookIII scanner (UMAX Systems GmbH, Düsseldorf, Gemany) and the ImageMaster software (GE Healthcare, formerly Amersham Biosciences, Rosendaal, The Netherlands).
Equipment for fermentation, filtration and homogenization
For 1-L, 30-L, 300-L, and 3000-L fermentations, stirred, temperature- and pH-regulated tank fermentors were used (1 L: Applicon fermentor with Biocontroller ADI 1030, Applicon, Frederiksberg, Denmark; 30L and 300 L: Chemap fermentors, Chemap AG, Volketswil, Switzerland; 3000 L: custom made fermentor). Stirring was carried out with a propeller blade stirrer at approximately 50 rpm, to ensure proper mixing of the base and of nisin. During the fermentation, temperature and base consumption were recorded to monitor the process. To determine the cell density of the cultures and to find the time point for induction, samples were taken at regular intervals and the OD at 600 nm (light path 1 cm) was determined.
For microfiltration, a one-stage filtration installation was used with a Ceraver ceramic membrane of 0.8 μm pore size and 3.8 m2 surface (Membralox, Pall, East Hills, New York, U.S.A.). The flow rate was approximately 330 L/h.
Concentrated cell suspensions were disintegrated using a continuous homogenization process at 1400 bar with a flow rate of about 80 L/h (Homogenizer 10.51 VH, APV, Hendrik Ido Ambacht, The Netherlands). To prevent overheating, the pressure-reduction nozzle was cooled with ice water.
Chromatography
Large-scale chromatography was carried out with a Bioprocessor (max. flow rate of 120 L/h) and a BPG300 column (Amersham Biosciences, Roosendaal, The Netherlands). For the separation 14.1 L Sepharose Fast Flow resin (17-0792-04, Amersham Biosciences) was used, resulting in a bed height of 20 cm. The sample was prepared for chromatography by adjusting the pH to 7.2 with 0.4 M NaH2PO4 pH 7.5. The following buffers and cleaning solutions were used: Equilibration buffer, 50 mM NaH2PO4 pH 7.5; Elution buffer, 50 mM NaH2PO4 + 0.5 M NaCl pH 7.5; Regeneration buffer, 50 mM NaH2PO4 + 1 M NaCl pH 7.5; Cleaning solution, 1 M NaOH. A standard chromatography run was carried out as follows. The column was equilibrated with 2 volumes equilibration buffer at 120 L/h, the pH-adjusted sample was loaded at 100 L/h, the loaded column was washed with 3 volumes equilibration buffer at 120 L/h, lysostaphin was eluted with 2 column volumes elution buffer at 120 L/h and the column was cleaned with 1 column volume cleaning solution at 50 L/h. Finally, the column was regenerated with 2 column volumes regeneration buffer at 120 L/h. This process was repeated for all subsequent runs. Since the whole sample of about 400 L could not be loaded at one time (Results), it was divided into four portions of approximately 100 L each.
For recapture, the following buffers and solutions were used: Equilibration buffer, 12.5 mM NaH2PO4 + 75 mM NaCl pH 7.0; Elution buffer, 25 mM NaH2PO4 + 0.25 M NaCl pH 7.0, Cleaning solution and regeneration buffer were as mentioned above. For recapture, the eluate of the capture step was desalted using diafiltration and the pH was adjusted to pH 7.0. The column was equilibrated with 2 column volumes of equilibration buffer at 120 L/h, the sample was loaded at 100 L/h, the column was washed with 3 volumes equilibration buffer at 120 L/h and the product was eluted with 2 column volumes elution buffer. Cleaning and regeneration was done as described above.
Authors' contributions
IM and JM set up and supervised the project. IM was in charge of the genetic work. ES and BB worked out and were in charge of the fermentations. PL set up and was in charge of the filtration processes. IvS and EF set up and carried out the chromatography steps.
Acknowledgements
We thank Nico Vaandrager, Johan Klok and Anne Wiersma for excellent technical assistance. Furthermore, we thank Andy Lees for critical reading of the manuscript.
==== Refs
Leroy F Devuyst L Lactic acid bacteria as functional starter cultures for the food fermentation industry Trends in Food Science and Technology 2004 15 67 78 10.1016/j.tifs.2003.09.004
Konings WN Kok J Kuipers OP Poolman B Lactic acid bacteria: the bugs of the new millennium Current Opinion in Microbiology 2000 3 276 282 10851157 10.1016/S1369-5274(00)00089-8
Bolotin A Wincker P Mauger S Jaillon O Malarme K Weissenbach J Ehrlich SD Sorokin A The complete genome sequence of the lactic acid bacterium Lactococcus lactis ssp. lactis IL1403 Genome Res 2001 11 731 753 11337471 10.1101/gr.GR-1697R
Bolotin A Ehrlich SD Sorokin A Studies of genomes of dairy bacteria Lactococcus lactis Sci Aliment 2002 22 45 53
de Vos WM Safe and sustainable systems for food-grade fermentations by genetically modified lactic acid bacteria International Dairy Journal 1999 9 3 10 10.1016/S0958-6946(99)00038-2
Steidler L In situ delivery of cytokines by genetically engineered Lactococcus lactis Antonie Van Leeuwenhoek 2002 82 323 331 12369199 10.1023/A:1020656220815
Nouaille S Ribeiro LA Miyoshi A Pontes D Le Loir Y Oliveira SC Langella P Azevedo V Heterologous protein production and delivery systems for Lactococcus lactis Genet Mol Res 2003 2 102 111 12917806
Blatny JM Ertesvåg H Nes IF Valla S Heterologous gene expression in Lactococcus lactis; expression of the Azotobacter vinelandii algE6 gene product displaying mannuronan C-5 epimerase activity FEMS Microbiology Letters 2003 227 229 235 14592713 10.1016/S0378-1097(03)00685-2
Kunji ER Slotboom DJ Poolman B Lactococcus lactis as host for overproduction of functional membrane proteins Biochimica et Biophysica Acta 2003 1610 97 108 12586384
Hols P Kleerebezem M Schanck AN Ferain T Hugenholtz J Delcour J de Vos WM Conversion of Lactococcus lactis from homolactic to homoalanine fermentation through metabolic engineering Nature Biotechnology 1999 17 588 592 10385325 10.1038/9902
Platteeuw C van Alen-Boerrigter I van Schalkwijk S de Vos WM Food-grade cloning and expression system for Lactococcus lactis Appl Environ Microbiol 1996 62 1008 1013 8975595
Kuipers OP de Ruyter PGGA Kleerebezem M de Vos WM Quorum sensing-controlled gene expression in lactic acid bacteria Journal of Biotechnology 1998 64 15 21 10.1016/S0168-1656(98)00100-X
de Ruyter PG Kuipers OP Beerthuyzen MM Alen-Boerrigter I de Vos WM Functional analysis of promoters in the nisin gene cluster of Lactococcus lactis J Bacteriol 1996 178 3434 3439 8655538
Hickey RM Ross RP Hill C Controlled autolysis and enzyme release in a recombinant lactococcal strain expressing the metalloendopeptidase enterolysin A Appl Environ Microbiol 2004 70 1744 1748 15006800 10.1128/AEM.70.3.1744-1748.2004
Thumm G Götz F Studies on prolysostaphin processing and characterization of the lysostaphin immunity factor (Lif) of Staphylococcus simulans biovar staphylolyticus Molecular Microbiology 1997 23 1251 1265 9106216 10.1046/j.1365-2958.1997.2911657.x
Trayer HR Buckley CE Molecular properties of lysostaphin, a bacteriolytic agent specific for Staphylococcus aureus Journal of Biological Chemistry 1970 245 4842 4846 5456157
Kokai-Kun JF Walsh SM Chanturiya T Mond JJ Lysostaphin cream eradicates Staphylococcus aureus nasal colonization in a cotton rat model Antimicrob Agents Chemother 2003 47 1589 1597 12709327 10.1128/AAC.47.5.1589-1597.2003
Wu JA Kusuma C Mond JJ Kokai-Kun JF Lysostaphin disrupts Staphylococcus aureus and Staphylococcus epidermidis biofilms on artificial surfaces Antimicrob Agents Chemother 2003 47 3407 3414 14576095 10.1128/AAC.47.11.3407-3414.2003
von Eiff C Kokai-Kun JF Becker K Peters G In vitro activity of recombinant lysostaphin against Staphylococcus aureus isolates from anterior nares and blood Antimicrob Agents Chemother 2003 47 3613 3615 14576128 10.1128/AAC.47.11.3613-3615.2003
Terzaghi BE Sandine WE Improved medium for lactic streptococci and their bacteriophages Applied Microbiology 1975 29 807 813 16350018
Pedersen MB Koebmann BJ Jensen PR Nilsson D Increasing acidification of nonreplicating Lactococcus lactis delta-thyA mutants by incorporating ATPase activity Appl Environ Microbiol 2002 68 5249 5257 12406711 10.1128/AEM.68.6.3010-3023.2002
Mierau I Olieman C Mond J Smid EJ Optimization of the Lactococcus lactis nisin-controlled gene expression system NICE for industrial applications Microb Cell Fact 2005 4 16 15921537 10.1186/1475-2859-4-16
Sambrook J Fritsch EF Maniatis T Molecular cloning: A laboratory manual, 2nd ed. 1989 Cold Spring Harbour, Cold Spring Harbour Laboratory Press
Laemmli UK Cleavage of structural proteins during the assembly of the head of bacteriophage T4 Nature 1970 227 680 685 5432063
de Ruyter PG Kuipers OP de Vos WM Controlled gene expression systems for Lactococcus lactis with the food- grade inducer nisin Appl Environ Microbiol 1996 62 3662 3667 8837421
de Vos WM Gene cloning and expression in lactic streptococci FEMS Microbiology Letters 1987 46 281 295 10.1016/0378-1097(87)90113-3
| 15921518 | PMC1173137 | CC BY | 2021-01-04 16:24:35 | no | Microb Cell Fact. 2005 May 27; 4:15 | utf-8 | Microb Cell Fact | 2,005 | 10.1186/1475-2859-4-15 | oa_comm |
==== Front
Mol CancerMolecular Cancer1476-4598BioMed Central London 1476-4598-4-211596975010.1186/1476-4598-4-21ResearchRegulation of pancreatic cancer cell migration and invasion by RhoC GTPase and Caveolin-1 Lin Min [email protected] Melinda M [email protected] Sofia D [email protected] Madanamohan [email protected] Golen Kenneth L [email protected] Department of Internal Medicine, Division of Hematology/Oncology, The University of Michigan Comprehensive Cancer Center, Ann Arbor, Michigan 48109, USA2 Department of Cell and Molecular Physiology, The University of North Carolina, Chapel Hill, North Carolina 27599, USA3 Department of Neuroscience, The University of Michigan Medical School, Ann Arbor, Michigan 48109, USA2005 21 6 2005 4 21 21 29 3 2005 21 6 2005 Copyright © 2005 Lin et al; licensee BioMed Central Ltd.2005Lin et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
In the current study we investigated the role of caveolin-1 (cav-1) in pancreatic adenocarcinoma (PC) cell migration and invasion; initial steps in metastasis. Cav-1 is the major structural protein in caveolae; small Ω-shaped invaginations within the plasma membrane. Caveolae are involved in signal transduction, wherein cav-1 acts as a scaffolding protein to organize multiple molecular complexes regulating a variety of cellular events. Recent evidence suggests a role for cav-1 in promoting cancer cell migration, invasion and metastasis; however, the molecular mechanisms have not been described. The small monomeric GTPases are among several molecules which associate with cav-1. Classically, the Rho GTPases control actin cytoskeletal reorganization during cell migration and invasion. RhoC GTPase is overexpressed in aggressive cancers that metastasize and is the predominant GTPase in PC. Like several GTPases, RhoC contains a putative cav-1 binding motif.
Results
Analysis of 10 PC cell lines revealed high levels of cav-1 expression in lines derived from primary tumors and low expression in those derived from metastases. Comparison of the BxPC-3 (derived from a primary tumor) and HPAF-II (derived from a metastasis) demonstrates a reciprocal relationship between cav-1 expression and p42/p44 Erk activation with PC cell migration, invasion, RhoC GTPase and p38 MAPK activation. Furthermore, inhibition of RhoC or p38 activity in HPAF-II cells leads to partial restoration of cav-1 expression.
Conclusion
Cav-1 expression inhibits RhoC GTPase activation and subsequent activation of the p38 MAPK pathway in primary PC cells thus restricting migration and invasion. In contrast, loss of cav-1 expression leads to RhoC-mediated migration and invasion in metastatic PC cells.
Pancreatic cancerRhoC GTPasecaveolin-1cell migrationmetastasisMAPK
==== Body
Background
Caveolin-1 (cav-1) is the major structural component of small Ω-shaped plasma membrane invaginations called caveolae [1]. Caveolae regulate plasma membrane signal transduction, with cav-1 acting as a scaffolding molecule to sequester and organize multi-molecular signaling complexes [2,3]. Many proteins which regulate multiple cellular activities such as growth and survival contain a putative cav-1 binding domain [2-4]. Recent evidence suggests a crucial role for cav-1 in regulating cellular migration and metastasis[5-7]. In a tumor progression model of breast cancer, loss of cav-1 corresponded to increased metastasis, while ectopic expression of cav-1 inhibited metastasis[8]. Furthermore, disruption of the Cav-1 gene in transgenic mice promotes mammary tumorigenesis and increased formation of metastases[9]. Conversely, in esophageal squamous cell carcinoma, lung adenocarcinoma, prostate, colon and clear cell renal cancers, high levels of cav-1 protein is associated with increased metastatic potential[10-12]. The molecular mechanism(s) of how cav-1 regulates tumor cell migration and metastasis has not been thoroughly explored.
Recent immunohistochemical studies have implicated increased cav-1 expression as a poor prognostic factor for pancreatic adenocarcinoma (PC) [13]. In the present study, we set out to determine whether cav-1 played a role in PC cell migration and invasion; initial steps in the metastatic cascade. Additionally, we attempted to identify the molecules regulated by cav-1 that are involved in PC cell migration and invasion.
Numerous molecules have been identified which interact with cav-1 [14-18]. Among these are the small monomeric GTPases Ras and RhoA [3,4]. The Ras Homology or Rho-subfamily of GTPases are typically involved in actin cytoskeleton rearrangement during cellular migration (reviewed in [19]). RhoC GTPase is a member of the Rho-subfamily that is associated with aggressive and highly metastatic tumors including PC [20-28].
Several studies have implicated RhoC as the predominant Rho GTPase in PC tumors and its expression is associated with metastasis and decreased survival (6 month versus 12 month for patients whose tumor expressed low or no RhoC) [29]. In a study aimed at characterizing genes involved in PC, laser capture microdissection (LCM) was used to compare normal pancreatic ductal cells with pancreatic cancers by cDNA microarray analysis [30]. RhoC was overexpressed and found in primary tumors that were locally invasive and in tumors from a variety of metastatic sites [30]. Another cDNA microarray study compared LCM isolated samples from 10 PC tumor specimens with 5 chronic pancreatitis and 5 normal pancreas specimens [31]. In PC tumors, RhoC was overexpressed 3- to 6-fold compared to normal and 2- to 4-fold compared with chronic pancreatitis (Craig Logsdon, personal communication) [31]. In each of these studies RhoA, Rac1 and Cdc42 were not significantly altered nor were they associated with aggressive or metastatic disease.
Similar to Ras, Rho GTPases are activated by a complex network of regulatory proteins and exists in the cell in an inactive GDP-bound and active GTP-bound state [32,33]. Unlike Ras, no activating mutations have been identified for Rho proteins in human cancers [32,33]. Therefore, increased Rho activity appears to be due to aberrant expression and/or dysregulation of regulatory proteins[32,33]. Like several Rho-subfamily members, RhoC contains a putative cav-1 binding domain. In the present study we demonstrate that cav-1 expression in PC cell lines regulates RhoC activation and cellular migration and invasion through the mitogen activated protein kinase (MAPK) pathway.
Results
Caveolin-1 Expression in Pancreatic Cancer Cell Lines
Ten pancreatic adenocarcinoma (PC) cell lines were obtained from ATCC and compared with immortalized human pancreatic ductal epithelial (HPDE) cells for caveolin-1 (cav-1) expression. Figure 1A are the results of immunoblot analysis for total cellular cav-1 protein expression. Cav-1 expression was variable among the cell lines, with high cav-1 levels detected in the HPDE and PC cells derived from primary tumors (BxPC-3 and MiaPaCa-2). Comparatively, cav-1 levels in immortalized HPDE cells were lower than the BxPC-3 and MiaPaCa-2 cell lines. Expression was low or absent in cell lines derived from metastatic tumors or ascites fluid (HPAF-II, Capan-1, SW1990, SU86.86, Capan-2 and AsPc-1). The exception to this is the CFPAC-1 cell line, which was derived from a liver metastasis in a patient that had chronic cystic fibrosis. Interestingly, the Panc-1 cell line, which was derived from an invasive intraductal extension of a primary tumor, had an intermediate expression level. These results suggest that high cav-1 expression may be associated with primary tumors, while loss of cav-1 may be associated with metastases.
Figure 1 Caveolin-1 expression in human pancreatic adenocarcinoma cell lines. Panel A is a comparison of caveolin-1 protein expression in human ductal pancreatic epithelial (HPDE) and 10 pancreatic cancer cell lines. Aliquots of 50 μg total protein were probed by immunoblotting with a monoclonal antibody specific for caveolin-1. Both the α and β isoforms of caveolin-1 were detected. Immunoblot analysis of β-actin served as a loading control. B. Comparison of BxPC-3 and HPAF-II migration in a colloidal gold random migration assay. PC cells were seeded onto coverslips coated with colloidal gold, stimulated by the addition of 10% FBS and allowed to migrate for 16 h at 37°C. Phagokinetic tracks were photographed and areas measured. Invasion was measured by a Matrigel invasion assay using 10% serum as a chemoattractant. C. HPAF-II cells were transiently transfected with GFP-cav-1 or GFP only and their migratory capabilities measured in a blue-fluorescent bead migration assay in a manner identical to what was described for the colloidal gold assay. A representative Western blot demonstrating that ectopic cav-1 protein levels in the HPAF-II cells were similar to the BxPC-3 cell line. Phagokinetic tracks from GFP-expressing cells were imaged and areas measured. GFP-cav-1 cells were significantly less migratory (*p = 0.0032) and less invasive (*p = 0.002) than the controls.
To determine whether cav-1 expression plays a central role in cell migration and invasion we chose the BxPC-3 and HPAF-II cell lines to represent tumor cells derived from primary tumors and metastases, respectively. Morphologically, both of the cell lines are similar and have a well-differentiated, epithelial appearance. We first tested the cells in a colloidal gold random motility assay to assess basal, non-ECM mediated migratory capabilities of cells. Figure 1B shows that after 16 h the BxPC-3 cell line was essentially non-migratory with an average phagokinetic track area of 357 ± 107 square pixels in contrast to the HPAF-II cell line which was highly migratory with an average track area of 1567 ± 227 square pixels. Similarly, when tested for their ability to invade through a Matrigel coated filter in response to a serum chemoattractant, the HPAF-II cells were 10-fold more invasive than the BxPC-3 cells.
Next we sought to establish whether loss of cav-1 was responsible for the increased migratory and invasive capabilities of the HPAF-II cells. To accomplish this we transiently transfected the HPAF-II cells with a GFP-tagged cav-1 expression vector. Using a variation of the colloidal gold assay, we compared untransfected, GFP-vector control and GFP-cav-1 transfected HPAF-II cells in a blue fluorescent bead random migration assay. This assay is performed the same as the colloidal gold assay, but allows us to fluorescently identify GFP-transfected cells. Figure 1C demonstrates that ectopic re-expression of cav-1 to levels similar to that of the BxPC-3 cells significantly decreases HPAF-II migration (890 ± 67 square pixels; p = 0.0032) compared with the untransfected or GFP-vector control HPAF-II cells (1630 ± 183 and 1435 ± 235 square pixels, respectively).
Similarly in a Matrigel invasion assay, re-expression of cav-1 in the HPAF-II cells decreased the invasiveness of the cells nearly 2-fold compared with the controls. The results from the motility and invasion experiments suggest that HPAF-II migration and invasion is mediated by cav-1.
RhoC GTPase Induces PC Cell Migration
Due to their role in cellular migration, invasion and metastasis the Rho GTPases were logical molecular candidates to interact with cav-1 to mediate cell migration and invasion. RhoC GTPase is prevalent in metastatic tumors, particularly in PC [20-28]. Hence, we chose to study RhoC GTPase in relationship to cav-1. As shown in Figure 2A RhoC GTPase is expressed on the protein level to varying degrees in the panel of 10 pancreatic cancer cell lines that were analyzed for cav-1 expression. However, more important than expression is the activation state of the GTPase. Figure 2B is a comparison of RhoC expression and activation in the BxPC-3 and HPAF-II cell lines. RT-PCR and immunoblot analysis confirm that RhoC is highly expressed on the mRNA and protein levels both in the BxPC-3 and HPAF-II cells. To determine the relative levels of active RhoC in these cell lines, a GST fusion protein of the Rho-binding domain of the downstream Rho effector protein, rhotekin, was used to selectively pull out GTP-bound RhoC. Although levels of total GDP/GTP-bound RhoC were similar for both cell lines, levels of active RhoC was considerably higher in the HPAF-II cell line.
Figure 2 RhoC GTPase expression in BxPC-3 and HPAF-II cell lines. A. Total GDP/GTP-bound RhoC was measured in a panel of 10 PC cell lines. Protein (30 μg) was harvested from actively growing PC cell lines. B. Expression of RhoC in BXPC-3 and HPAF-II cells was measured by RT-PCR and immunoblot analysis using RhoC-specific primers and antibody, respectively. To determine the amount of active, GTP-bound RhoC a Rho activation assay was performed. GST-rhotekin is used to pull down RhoC-GTP from total cell lysates followed by immunoblotting with a RhoC-specific antibody. C. The RhoC activation assay was performed on the GFP-cav-1 HPAF-II transfectants 24 h after transfection. Differences in active and total RhoC levels were measured by densitometry and are represented.
To determine if cav-1 expression affected RhoC activation, levels of GTP-bound RhoC was measured in the GFP-cav-1 and control HPAF-II cells (Figure 2C). Compared with the controls, transient ectopic re-expression of cav-1 decreased levels of active RhoC by 6-fold without effecting total RhoC protein levels, suggesting regulation of RhoC activation by cav-1.
Next, to directly implicate RhoC in PC cell migration and invasion, we generated stable HPAF-II dominant negative RhoC (dnRhoC) transfectants. For clarity and simplicity the results shown are from a polyclonal population which is representative of three individual clones that were tested. Figure 3A shows a 57% decrease of active RhoC in the HPAF-II/dnRhoC transfectants compared with the untransfected and vector transfected controls. As a positive control the HPAF-II cells were treated with C3 exotransferase. C3 exotransferase is a toxin derived from Clostridium botulinum and is effective at inhibiting RhoA, -B and -C activity with virtually no effect on Rac1 or Cdc42 [34,35]. C3 exotransferase treatment reduced active RhoC levels by 85%. As shown, both C3 exotransferase treatment and expression of dnRhoC did not significantly alter total levels of RhoC protein.
Figure 3 Establishment of stable, dominant negative RhoC HPAF-II transfectants. A. Results of a RhoC activation assay and total RhoC Western blot comparing wildtype, C3 exotransferase treated, vector control and dominant negative RhoC (dnRhoC) transfected HPAF-II cells. B. Comparison of total (dark grey) and active (light grey) RhoA levels utilizing a RhoA-specific antibody. C. Results of the effect of dnRhoC transfection and C3 treatment on HPAF-II cells in a colloidal gold migration assay. Inhibition of RhoC lead to a significant decrease in migration (*p = 0.001) and invasion (*p = 0.023).
Dominant negative Rho GTPases work by entering into a non-productive interaction with Rho Guanine Exchange Factors (RhoGEFs), the proteins which catalyze the exchange of GDP for GTP. Due to the close homology between RhoC and RhoA (91% on the protein level), the possibility exists that both GTPases can be activated by the same RhoGEFs in vivo. As shown in Figure 3B, RhoA activity was not significantly altered by expression of dnRhoC. Again, as a positive control HPAF-II cells were treated with C3 exotransferase; this reduced RhoA activity an average of 63%. Therefore, RhoC activity was specifically decreased in the HPAF-II cells expressing dominant negative RhoC.
As shown in Figure 3C, inhibition of RhoC activity, either by dnRhoC or by treatment with C3 exotransferase, significantly reduced HPAF-II cell migration by nearly 4-fold (p = 0.001) and invasion by 3-fold (p = 0.023), suggesting a key role for RhoC GTPase mediating HPAF-II cell migration and invasion.
Association of Cav-1 and RhoC GTPase in PC cells
Next, we considered the possibility that RhoC activity and PC migration and invasion are regulated by cav-1 through physical interaction of the GTPase with the scaffolding protein. Proteins that associate with cav-1 contain the canonical cav-1 binding domain, ΦXΦXXXXΦ or ΦXXXXΦXXΦ (where Φ= Trp, Phe or Tyr) [36,37]. A review of the RhoC protein sequence revealed a putative cav-1 binding sequence at amino acid residues 35–43 (YVPTVFENY).
Figure 4A are the results of immunoprecipitation assays using either a polyclonal or monoclonal antibody specific for cav-1 followed by immunoblotting for RhoC. Cav-1 and RhoC proteins co-immunoprecipitated in the BxPC-3 cells but not in the HPAF-II cell line. Unexpectedly, the association between cav-1 and RhoC was restored in the HPAF-II/dnRhoC cell line suggesting that inhibition of RhoC activity leads to re-expression of cav-1 protein.
Figure 4 Interaction of cav-1 and RhoC in BxPC-3 and HPAF-II PC cell lines. A. Aliquots of 500 μg total protein from BxPC-3, HPAF-II and HPAF-II/dnRhoC transfectants were immunoprecipitated with either a monoclonal or polyclonal antibody to cav-1. Proteins were separated by SDS-PAGE, transferred to nitrocellulose and probed with either a RhoC-specific or cav-1 antibody. Normal IgG negative controls are also shown. B. Immunoblot analysis of 50 μg of total protein for total cav-1 expression using a cav-1 specific monoclonal antibody and actin was measured as a loading control. Densitometry was performed on each blot and expression levels were normalized to the actin control. Expression levels are represented as arbitrary units (AU).
The results of a cav-1 immunoblot for total cellular protein are shown in Figure 4B. Expression of cav-1 was increased 3.1-fold in the HPAF-II/dnRhoC cells compared with the parental HPAF-II and vector control cells. Actin was used as a loading control. Together with the immunoprecipitation data, these data suggest that cav-1 expression is in a reciprocal relationship with RhoC activation. High cav-1 protein expression leads to inhibition of RhoC activation while inhibition of RhoC activity leads to partial re-expression of cav-1. Furthermore, interaction of cav-1 and RhoC may result in decreased RhoC activation, limiting cell migration and invasion.
PC Cell Migration Involves the Mitogen Activated Protein Kinase (MAPK) Pathway
Previous studies demonstrated that activation of p42/p44 extracellular regulated kinase (Erk) decreased cav-1 protein levels in constitutively-active Ras transformed NIH3T3 cells [38]. In the same set of studies it was shown that inhibition of oncogenic Ras-induced Erk activation with PD98059 increased cav-1 expression 5-fold [38]. That same group also demonstrated that ectopic re-expression of cav-1 decreased Erk activation in Ras-transformed CHO cells [39]. Our laboratory has previously demonstrated that RhoC can mediate inflammatory breast cancer cell migration and invasion through co-activation of the p42/p44 Erk and p38 arms of the MAPK pathway [40]. With these studies in mind we next examined whether RhoC can activate p42/p44 Erk in PC, subsequently decreasing cav-1 expression and leading to increased cellular migration. Also, we examined the potential involvement of p38 MAPK in this process.
Figure 5A is a comparison of active (phospho-) and total levels of p42/p44 Erk and p38 MAPK in BxPC-3, HPAF-II and HPAF-II/dnRhoC cells that were serum starved for 16 h and stimulated with 10% serum alone or after pre-treatment with C3 exotransferase. Although active phospho-Erk levels were stimulated in both cell lines, active p42/p44 Erk was considerably higher in the BxPC-3 cell line compared with the HPAF-II cell line. Pretreatment with C3 exotransferase slightly decreased active Erk in serum stimulated cell lines. Similar results were previously demonstrated and are unexpected since the HPAF-II cell line harbors an activating G12D K-Ras mutation, while the BxPC-3 cell line has a wildtype K-Ras [41,42]. In the HPAF-II/dnRhoC cells, the overall phospho-p42/p44 levels were higher compared with the parental HPAF-II cell line suggesting that inhibition of RhoC leads to increased Erk activation.
Figure 5 Analysis of the mitogen activated protein kinase pathway (MAPK) in the BxPC-3 and HPAF-II cells. A. Actively growing PC cells were serum starved for 24 h and left untreated or pretreated for 1 h with 5 μg C3 exotransferase, stimulated by the addition of 10% serum and proteins harvested 15 min later. Activated p42/p44 Erk and p38 MAPK were detected by antibodies specific for the phosphorylated forms of each of those proteins. Immunoblots were stripped and reprobed with antibodies specific for total forms of each protein. B. Comparison of active p42/p44 Erk and p38 MAPK in GFP-cav-1 and GFP-vector transient transfectants at 24 h after transfection. C. Results of a colloidal gold migration assay after 30 min pre-treatment of cells with 30 μM of either PD98059 or SB220025. Phagokinetic tracks were imaged at 6 h after stimulation. Results are given for DMSO control (dark grey), PD98059 treated (medium grey) and SB220025 treated (light grey). SB220025 treatment of the HPAF-II cells led to a significant decrease (*p = 0.001) in migration and invasion. Levels of cav-1 protein were increased, while active RhoC was decreased in SB220025 treated HPAF-II cells.
Converse to what was observed for p42/p44 Erk the levels of phospho-p38 MAPK were low in the BxPC-3 and HPAF-II/dnRhoC cells and higher in the HPAF-II cells. C3 treatment decreased p38 activity in the BxPC-3 and HPAF-II cell lines. The levels of active p38 in the HPAF-II/dnRhoC and C3 treated HPAF-II cells were comparable; demonstrating an approximate 52% decrease in activity compared with the serum stimulated HPAF-II cells. Taken together, an inverse relationship between p42/p44 Erk and p38 MAPK signaling is suggested in the PC cells.
Levels of active Erk and p38 MAPK were assessed in the HPAF-II/GFP-cav-1 transfectants. As shown in Figure 5B, levels of active phospho-p42/p44 Erk increased in the GFP-cav-1 transfectants compared with the untransfected and control GFP-vector transfectants. Conversely, levels of phosphorylated p38 decreased in the cav-1 transfectants, mirroring what is observed in the HPAF-II/dnRhoC cells.
To tease out the individual roles of the p42/p44 Erk and p38 MAPK pathways in PC cellular migration, the BxPC-3 and HPAF-II cells were treated with the pharmacologic inhibitors PD98059 (to inhibit MEK1 and subsequently Erk) or SB220025 (to inhibit p38) and tested in migration and invasion assays (Figure 5C). SB220025 treatment significantly reduced HPAF-II migration and invasion (p = 0.001). PD98059 treatment had no effect on the cells ability to move in either assay suggesting that migration and invasion occurs through signaling of the p38 MAPK pathway.
Changes in cav-1 expression due to inhibitor treatment are also shown in Figure 5C. Cav-1 expression was slightly less after PD98059 treatment in both PC cell lines. Treatment with SB220025 increased cav-1 expression dramatically in the HPAF-II and slightly in the BxPC-3 cells, implying a reciprocal relationship between cav-1 expression and p38 activation. Furthermore, RhoC activity was appreciably reduced in the SB220025 treated HPAF-II cells strengthening the notion that cav-1 is in a reciprocal relationship with RhoC and p38 activity.
Methyl-β-cyclodextrin increases PC cell motility
Lastly, we treated the BxPC-3 cell line with methyl-β-cyclodextrin (MβCD), which sequesters cholesterol, disrupts caveolae and mis-localizes cav-1 in the cell [43]. As shown in Figure 6A, MβCD treatment significantly increased the area of BxPC-3 migration (p = 0.0001). Increased cellular migration was accompanied by a significant increase in active GTP-bound RhoC (Figure 6B; p = 0.0001). Consistent with previous observations, Figure 6C demonstrates that levels of phospho-p42/p44 Erk in the MβCD treated cells decreased while levels of phospho-p38 MAPK increased. The changes in MAPK proteins approached but did not achieve statistical significance. These data further suggest that activation of RhoC GTPase is regulated by cav-1 and that RhoC signals through the p38 arm of the MAPK pathway to induce cellular migration.
Figure 6 Effect of disruption of caveolae and mis-localization of caveolin-1 after treatment of BxPC-3 cells with MβCD. A. Results of a colloidal gold random motility assay performed after a 1 h pre-treatment of BxPC-3 cells with 5 mM MβCD in growth medium. Phagokinetic tracks were measured 16 h after stimulation. MβCD treated BxPC-3 cells were significantly more migratory than untreated cells (* p = 0.0001) B. Total (light grey) and active GTP-bound (dark grey) levels of RhoC were measured in MβCD treated BxPC-3 cells 3 h after treatment. A statistical difference in RhoC activation was reached with a p value of 0.0001. C. Immunoblot analysis of basal levels of active (light grey) and total (dark grey) p42/p44 Erk and p38 MAPK in MβCD treated BxPC-3 cells 3 h after treatment.
Discussion
Both loss and overexpression cav-1 has been described in tumor progression, sometimes within the same tumor type [44-47]. Loss and overexpression of cav-1 has also been associated with the progression to a metastatic phenotype within different cancers [47,8]. However, the molecular mechanism(s) by which cav-1 confers metastatic ability has not been thoroughly explored. Here we present evidence for the involvement of RhoC GTPase and p38 MAPK in the migratory and invasive phenotype after loss of cav-1 expression in PC cells.
Immunohistochemical staining of primary human PC tumors demonstrated that high cav-1 expression correlated with large PC tumor size and a poor prognosis [13]. In our current study we found that PC cell lines derived from primary tumors have high cav-1 expression while those derived from distant metastases have significantly reduced or lost expression. Two recent studies have demonstrated a biphasic expression of cav-1 in primary vs. metastatic tumors [48,49]. Both in renal cell and oral carcinomas, cav-1 is overexpressed in primary tumors but low or absent in distant metastases. It is possible that biphasic expression of cav-1; meaning, overexpression in primary tumors and low or absent expression in metastases occurs in pancreatic cancer.
Many proteins including the majority of the Rho GTPases have the ability to associate with cav-1 [3,4,14-18]. RhoC has been shown to be the predominant Rho protein in PC and is associated with particularly aggressive disease [21-29]. Given the role of RhoC in cellular migration and its prevalence in metastatic tumors, it was a logical choice to examine in relationship to cav-1.
One effect that high cav-1 expression could have in PC cells is regulating RhoC activity, thus limiting cell migration and promoting growth. This could potentially explain why PC cells harboring activating K-Ras mutations have lower activated phospho-p42/p44 Erk levels than expected [41,42]. Our data suggest that expression of high cav-1 levels in primary PC tumor cells favor the p42/p44 Erk pathway which is associated with cell growth and survival (reviewed in [50]). When cav-1 expression is diminished or lost, the p38 MAPK pathway is favored and cell migration occurs. Data in the current study also indicate that the p38 MAPK pathway leads to RhoC-mediated PC cell migration and invasion, while the Erk pathway does not. This is contrary to what we had demonstrated for RhoC-mediated migration and invasion in inflammatory breast cancer [40].
Experiments in NIH3T3 and CHO cells suggest a reciprocal and reversible relationship between p42/p44 Erk activation and cav-1 expression [38,39]. Our data suggests that this is not the case in PC cells. High cav-1 expression is associated with higher levels of active p42/p44 irregardless of K-Ras mutational status. The same studies also suggested a role for active protein kinase A (PKA) in reversibly suppressing cav-1 expression [38,39]. Our early studies suggested that this is also not the case in PC cells (data not shown). Instead our data suggest that suppression of RhoC or p38 MAPK activity restores cav-1 expression and suppresses migration and invasion.
This study is the first to draw a functional link between RhoC and cav-1 and to demonstrate involvement of the MAPK signal transduction pathway. This study also presents a fresh link to the regulation of RhoC-mediated migration and invasion, giving a functional role for RhoC in PC tumor cells.
Conclusion
This study suggests that cav-1 expression suppresses RhoC GTPase activation and RhoC-mediated cellular migration and invasion. Further, expression of cav-1 favors activation of the p42/p44 Erk pathway, perhaps promoting growth and survival. Loss of cav-1 expression leads to increased RhoC activation favoring the p38 MAPK pathway and cellular migration and invasion. Inhibition of RhoC or p38 activation results in decreased cellular movement and partial re-expression cav-1 protein. These results may imply a biphasic expression of cav-1 during PC tumor progression promoting growth and survival in the primary tumor cells and motility in the metastatic cells. Lastly, this study suggests a functional role for RhoC GTPase in PC migration and invasion.
Methods
Cell Culture
All pancreatic cancer (PC) cell lines were obtained from American Type Culture Collection (ATCC; Rockville, MD) and grown in their required growth medium per ATCC description. Specifically, BxPC-3 cells were grown in 90% RPMI 1640 (Bio Whittaker, Walkersville, MD), 10% FBS (Biosource International, Rockville, MD) and HPAF-II cells in 90% EMEM (Bio Whittaker), 10% FBS. Human pancreatic ductal epithelial (HPDE) cells (a gift from Dr. Ming Sound Tsao) were grown in keratinocyte serum-free (KSF) medium with 0.2 ng/ml EGF and 30 μg/ml bovine pituitary extract (InVitrogen Gibco, Carlsbad, CA) [51,52]. In the inhibitor studies, cells were pre-incubated for 0.5 h with 30 μM PD98059 or SB220025, (Calbiochem, San Diego, CA) or as a control, with DMSO carrier and stimulated with normal growth medium. C3 exotransferase (5 μg) was introduced into cells as previously described using a lipid transfer mediated method [53,40] and treated for 1 h before analysis. BxPC-3 cells were treated with 5 mM MβCD in medium for 2 h.
Plasmids and Transfection
Wild type human cav-1 was cloned from immortalized HPDE cells by RT-PCR. Total RNA was isolated from HPDE cells by Trizol (InVitrogen) and cDNA synthesized with the Wizard AMV-reverse transcriptase kit (Promega, Madison, WI). A 10 μl aliquot of cDNA was amplified by PCR, ligated into the pGEM-T Easy vector (Promega), and sequenced by the University of Michigan DNA sequencing core. The cav-1 sequence was confirmed using the NCBI BLAST database. The internal stop codon was removed using QuikChange (Stratagene, La Jolla, CA) and the insert transferred to pcDNA3.1/CT-GFP (InVitrogen). Cav-1 constructs were transiently transfected into PC cells using FuGene6 (Roche, Indianapolis, IN) and assayed 24 h later. Empty vector was used as a transfection controls. Dominant negative RhoC (dnRhoC) in pcDNA3.1 was obtained from Guthrie cDNA resource center. The dnRhoC construct or vector control was introduced into HPAF-II cells using FuGene6 and stable transfectants established by growing cells continuously in 250 μg/ml neomycin. Individual clones as well as a polyclonal population were chosen and tested in these experiments. Expression of the transgene was confirmed by RT-PCR for sequences unique to the plasmid. The results of the polyclonal population are reported because they are representative of the data obtained from all the clones.
RT-PCR Analysis
RT-PCR of RhoC expression in PC cell lines was performed as previously described [54] using primers specific for RhoC GTPase and GAPDH control. Triplicate analysis of RT-PCR was performed.
Immunoprecipitation/Western Blot Analysis
Protein was harvested from cells using RIPA buffer as previously described [54]. Lysates were separated by SDS-PAGE on a 4–15% gel, transferred to nitrocellulose, blocked and probed with a polyclonal antibody for RhoC [55], monoclonal antibodies for cav-1, (Pharmingen, San Diego, CA), RhoA (Cytoskelton Inc., Denver, CO.) or total and phospho- p42/p44 MAPK and p38 MAPK (Cell Signaling Technologies, Beverly, MA). After incubation with a goat anti-rabbit-HRP or goat anti-mouse-HRP antibody (Santa Cruz Biotechnology, Santa Cruz, CA), immunoblots were developed with Lumiglo (Cell Signaling Technologies), exposed to Hyperfilm (Amersham, Piscataway, NJ) and images recorded on an Alpha Image 950 documentation system (Alpha Innotech, San Leandro, CA). Densitometry was performed using ImageJ software (version 1.30; available from the NIH at ). Immunoprecipitation experiments were preformed by pre-incubating 1 mg of protein lysate with normal isotype control for 1 h followed by an overnight incubation with a polyclonal or monoclonal antibody for cav-1 (Pharmingen) and Protein A/G agarose PLUS (Santa Cruz) at 4°C. Precipitates were repeatedly pelleted by brief centrifugation followed by 4 washings with ice-cold PBS. Proteins were separated by SDS-PAGE on a 15% gel, transferred to nitrocellulose, blocked and probed with a polyclonal antibody to RhoC and exposed as described above.
Rho Activation Assay
Rho activation assay reagents were supplied to us by John Collard, Ph.D. at the Netherlands Cancer Institute [56,57] and used to determine activation of RhoC in PC cells. Cells, 5–10 × 106 cells/assay, were grown in fresh serum-containing medium for 16 h and lysed with ice-cold GST-FISH buffer (10% Glycerol, 50 mM Tris pH 7.4, 100 mM NaCl, 1% NP-40 and 2 mM MgCl2). Protein lysates, 900 μl of 1 μg/μl, were precleared with glutathione sepharose for 1 h at room temperature. The cleared lysates were incubated with GST-rhotekin-sepharose conjugate at 4°C for 30 min. The sepharose conjugate was collected by centrifugation, resuspended in Laemelli buffer, separated by SDS-PAGE, transferred to nitrocellulose and probed with a RhoC antibody developed by our laboratory [55]. Densitometry was performed as described above. The results of triplicate assays were combined and described as percent of Rho activation detected in the wild-type cell line.
Random Migration and Invasion Assays
Colloidal gold random migration assays were performed as previously described [54]. Random migration assays were also performed using the blue fluorescent bead motility HitKit (Cellomics, Pittsburg, PA). Cells, 200/well in serum free medium (SFM), were seeded onto a lawn of blue fluorescent beads in a BSA coated 96-well plate (BD Biosciences, San Diego, CA). Cells were stimulated 1 h later with 10% FBS, incubated 6 h at 37°C, and fixed. Cells were visualized by fluorescent microscopy, digitally recorded and the areas of phagokinetic tracks measured using the ImageJ software. Migration assays were performed at least three separate times and at least 250 cell tracks were measured per condition in each analysis.
The Matrigel invasion assays were performed as previously described using pre-coated Transwell filters with 8 μ pores [58,54]. Cells were harvested and resuspended in serum-free medium containing 0.1% BSA at a concentration of 3.75 × 105 cells/ml, and 0.5 ml was added to the top chambers and placed for 24 h at 37°C in a 10% CO2 incubator. The cell suspension was aspirated from the top chamber of all three assays. Excess Matrigel was removed from the invasion assay filter using a cotton swab. The filters were then cut away from the Transwell assembly, fixed top side down with methanol to a glass microscope slide, stained with H&E, and the entire surface of the filters counted. The number of cells that moved in serum-free (i.e. no chemoattractant) controls was considered background and was subtracted from the number of cells counted in the samples containing chemoattractant.
Statistical Analysis
All experiments and assays were performed a minimum of three separate times. Statistical analysis was performed by the University of Michigan Biostatistics Core using the Wilcoxon rank-sum analysis and Students t-test.
Abbreviations
PC, pancreatic cancer; cav-1, caveolin-1; MAPK, mitogen activated protein kinase; SFM, serum-free medium; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; HPDE, human pancreatic ductal epithelial cells; FBS, fetal bovine serum; RT-PCR, reverse transcriptase polymerase chain reaction; GFP, green fluorescent protein; MβCD, methyl β cyclodextrin.
Authors' contributions
ML carried out the majority of Western blots, motility and Rho activation assays. MMD carried out RT-PCR and Western blot analysis. SDM contributed to initial experimental design and molecular analysis. MB performed caveolin-1 immunoprecipitation experiments and aided in experimental design. KvG conceived the project, designed the experiments, performed molecular analysis and significantly contributed to the writing of the manuscript.
Acknowledgements
Supported in part by the Thomas and Suzanne McPhee Pancreatic Research Fund and the University of Michigan Comprehensive Cancer Center support grant 5 P30 CA46592 (K.L.v.G.) and by a NIH grant R01CA77612 (S.D.M.). We would like to thank Kent Griffith, M.S., M.P.H. for performing statistical analysis, Drs. Craig Logsdon, Diane Simeone and Cynthia M. van Golen for insightful discussions and Ms. Robyn Blanzy-Hodges for help in preparing the manuscript. Technical assistance in preliminary experiments was provided by Dr. LiWei Bao.
==== Refs
Fujimoto T Kogo H Nomura R Une T Isoforms of caveolin-1 and caveolar structure J Cell Sci 2000 113 Pt 19 3509 3517 10984441
Okamoto T Schlegel A Scherer PE Lisanti MP Caveolins, a family of scaffolding proteins for organizing "preassembled signaling complexes" at the plasma membrane J Biol Chem 1998 273 5419 5422 9488658 10.1074/jbc.273.10.5419
Song KS Li S Okamoto T Quilliam LA Sargiacomo M Lisanti MP Co-purification and direct interaction of Ras with caveolin, an integral membrane protein of caveolae microdomains. Detergent-free purification of caveolae microdomains J Biol Chem 1996 271 9690 9697 8621645 10.1074/jbc.271.16.9690
Gingras D Gauthier F Lamy S Desrosiers RR Beliveau R Localization of RhoA GTPase to endothelial caveolae-enriched membrane domains Biochem Biophys Res Commun 1998 247 888 893 9647788 10.1006/bbrc.1998.8885
Zhang W Razani B Altschuler Y Bouzahzah B Mostov KE Pestell RG Lisanti MP Caveolin-1 inhibits epidermal growth factor-stimulated lamellipod extension and cell migration in metastatic mammary adenocarcinoma cells (MTLn3). Transformation suppressor effects of adenovirus-mediated gene delivery of caveolin-1 J Biol Chem 2000 275 20717 20725 10748172 10.1074/jbc.M909895199
Fiucci G Ravid D Reich R Liscovitch M Caveolin-1 inhibits anchorage-independent growth, anoikis and invasiveness in MCF-7 human breast cancer cells Oncogene 2002 21 2365 2375 11948420 10.1038/sj.onc.1205300
Lu Z Ghosh S Wang Z Hunter T Downregulation of caveolin-1 function by EGF leads to the loss of E-cadherin, increased transcriptional activity of beta-catenin, and enhanced tumor cell invasion Cancer Cell 2003 4 499 515 14706341 10.1016/S1535-6108(03)00304-0
Sloan EK Stanley KL Anderson RL Caveolin-1 inhibits breast cancer growth and metastasis Oncogene 2004 23 7893 7897 15334058 10.1038/sj.onc.1208062
Williams TM Medina F Badano I Hazan RB Hutchinson J Muller WJ Chopra NG Scherer PE Pestell RG Lisanti MP Caveolin-1 gene disruption promotes mammary tumorigenesis and dramatically enhances lung metastasis in vivo. Role of Cav-1 in cell invasiveness and matrix metalloproteinase (MMP-2/9) secretion J Biol Chem 2004 279 51630 51646 15355971 10.1074/jbc.M409214200
Ho CC Huang PH Huang HY Chen YH Yang PC Hsu SM Up-regulated caveolin-1 accentuates the metastasis capability of lung adenocarcinoma by inducing filopodia formation Am J Pathol 2002 161 1647 1656 12414512
Horiguchi A Asano T Asakuma J Asano T Sumitomo M Hayakawa M Impact of caveolin-1 expression on clinicopathological parameters in renal cell carcinoma J Urol 2004 172 718 722 15247769 10.1097/01.ju.0000130943.23317.08
Bender FC Reymond MA Bron C Quest AF Caveolin-1 levels are down-regulated in human colon tumors, and ectopic expression of caveolin-1 in colon carcinoma cell lines reduces cell tumorigenicity Cancer Res 2000 60 5870 5878 11059785
Suzuoki M Miyamoto M Kato K Hiraoka K Oshikiri T Nakakubo Y Fukunaga A Shichinohe T Shinohara T Itoh T Kondo S Katoh H Impact of caveolin-1 expression on prognosis of pancreatic ductal adenocarcinoma Br J Cancer 2002 87 1140 1144 12402154 10.1038/sj.bjc.6600619
Wu C Butz S Ying Y Anderson RG Tyrosine kinase receptors concentrated in caveolae-like domains from neuronal plasma membrane J Biol Chem 1997 272 3554 3559 9013605 10.1074/jbc.272.6.3554
Gleizes PE Noaillac-Depeyre J Dupont MA Gas N Basic fibroblast growth factor (FGF-2) is addressed to caveolae after binding to the plasma membrane of BHK cells Eur J Cell Biol 1996 71 144 153 8905291
Bilderback TR Grigsby RJ Dobrowsky RT Association of p75(NTR) with caveolin and localization of neurotrophin-induced sphingomyelin hydrolysis to caveolae J Biol Chem 1997 272 10922 10927 9099750 10.1074/jbc.272.16.10922
Li S Couet J Lisanti MP Src tyrosine kinases, Galpha subunits, and H-Ras share a common membrane-anchored scaffolding protein, caveolin. Caveolin binding negatively regulates the auto-activation of Src tyrosine kinases J Biol Chem 1996 271 29182 29190 8910575 10.1074/jbc.271.46.29182
Li S Okamoto T Chun M Sargiacomo M Casanova JE Hansen SH Nishimoto I Lisanti MP Evidence for a regulated interaction between heterotrimeric G proteins and caveolin J Biol Chem 1995 270 15693 15701 7797570 10.1074/jbc.270.26.15693
Raftopoulou M Hall A Cell migration: Rho GTPases lead the way Dev Biol 2004 265 23 32 14697350 10.1016/j.ydbio.2003.06.003
del Peso L Hernandez-Alcoceba R Embade N Carnero A Esteve P Paje C Lacal JC Rho proteins induce metastatic properties in vivo Oncogene 1997 15 3047 3057 9444953 10.1038/sj.onc.1201499
Clark EA Golub TR Lander ES Hynes RO Genomic analysis of metastasis reveals an essential role for RhoC Nature 2000 406 532 535 10952316 10.1038/35020106
Carr KM Bittner M Trent JM Gene-expression profiling in human cutaneous melanoma Oncogene 2003 22 3076 3080 12789283 10.1038/sj.onc.1206448
Marionnet C Lalou C Mollier K Chazal M Delestaing G Compan D Verola O Vilmer C Cuminet J Dubertret L Basset-Seguin N Differential molecular profiling between skin carcinomas reveals four newly reported genes potentially implicated in squamous cell carcinoma development Oncogene 2003 22 3500 3505 12776202 10.1038/sj.onc.1206571
Horiuchi A Imai T Wang C Ohira S Feng Y Nikaido T Konishi I Up-regulation of small GTPases, RhoA and RhoC, is associated with tumor progression in ovarian carcinoma Lab Invest 2003 83 861 870 12808121
Kamai T Tsujii T Arai K Takagi K Asami H Ito Y Oshima H Significant association of Rho/ROCK pathway with invasion and metastasis of bladder cancer Clin Cancer Res 2003 9 2632 2641 12855641
Shinto E Tsuda H Matsubara O Mochizuki H [Significance of RhoC expression in terms of invasion and metastasis of colorectal cancer] Nippon Rinsho 2003 61 Suppl 7 215 219 14574884
Wang W Yang LY Yang ZL Huang GW Lu WQ Expression and significance of RhoC gene in hepatocellular carcinoma World J Gastroenterol 2003 9 1950 1953 12970882
Kondo T Sentani K Oue N Yoshida K Nakayama H Yasui W Expression of RHOC is associated with metastasis of gastric carcinomas Pathobiology 2004 71 19 25 14555841 10.1159/000072958
Suwa H Ohshio G Imamura T Watanabe G Arii S Imamura M Narumiya S Hiai H Fukumoto M Overexpression of the rhoC gene correlates with progression of ductal adenocarcinoma of the pancreas Br J Cancer 1998 77 147 152 9459160
Crnogorac-Jurcevic T Efthimiou E Nielsen T Loader J Terris B Stamp G Baron A Scarpa A Lemoine NR Expression profiling of microdissected pancreatic adenocarcinomas Oncogene 2002 21 4587 4594 12085237 10.1038/sj.onc.1205570
Logsdon CD Simeone DM Binkley C Arumugam T Greenson JK Giordano TJ Misek DE Kuick R Hanash S Molecular profiling of pancreatic adenocarcinoma and chronic pancreatitis identifies multiple genes differentially regulated in pancreatic cancer Cancer Res 2003 63 2649 2657 12750293
van Golen KL Inflammatory Breast Cancer: Relationship between growth factor signaling and motility in aggressive cancers Breast Cancer Research 2003 5 174 12793901 10.1186/bcr598
Lin M van Golen KL Rho-regulatory proteins in breast cancer cell motility and invasion Breast Cancer Res Treat 2003
Aktories K Hall A Botulinum ADP-ribosyltransferase C3: a new tool to study low molecular weight GTP-binding proteins Trends Pharmacol Sci 1989 10 415 418 2515641 10.1016/0165-6147(89)90191-0
Aktories K Braun U Rosener S Just I Hall A The rho gene product expressed in E. coli is a substrate of botulinum ADP-ribosyltransferase C3 Biochem Biophys Res Commun 1989 158 209 213 2492192 10.1016/S0006-291X(89)80199-8
Couet J Belanger MM Roussel E Drolet MC Cell biology of caveolae and caveolin Adv Drug Deliv Rev 2001 49 223 235 11551396 10.1016/S0169-409X(01)00139-9
Couet J Li S Okamoto T Ikezu T Lisanti MP Identification of peptide and protein ligands for the caveolin-scaffolding domain. Implications for the interaction of caveolin with caveolae-associated proteins J Biol Chem 1997 272 6525 6533 9045678 10.1074/jbc.272.48.30429
Engelman JA Zhang XL Razani B Pestell RG Lisanti MP p42/44 MAP kinase-dependent and -independent signaling pathways regulate caveolin-1 gene expression. Activation of Ras-MAP kinase and protein kinase a signaling cascades transcriptionally down-regulates caveolin-1 promoter activity J Biol Chem 1999 274 32333 32341 10542274 10.1074/jbc.274.45.32333
Engelman JA Chu C Lin A Jo H Ikezu T Okamoto T Kohtz DS Lisanti MP Caveolin-mediated regulation of signaling along the p42/44 MAP kinase cascade in vivo. A role for the caveolin-scaffolding domain FEBS Lett 1998 428 205 211 9654135 10.1016/S0014-5793(98)00470-0
van Golen KL Bao LW Pan Q Miller FR Wu ZF Merajver SD Mitogen activated protein kinase pathway is involved in RhoC GTPase induced motility, invasion and angiogenesis in inflammatory breast cancer Clin Exp Metastasis 2002 19 301 311 12090470 10.1023/A:1015518114931
Yip-Schneider MT Lin A Marshall MS Pancreatic tumor cells with mutant K-ras suppress ERK activity by MEK-dependent induction of MAP kinase phosphatase-2 Biochem Biophys Res Commun 2001 280 992 997 11162624 10.1006/bbrc.2001.4243
Berger DH Jardines LA Chang H Ruggeri B Activation of Raf-1 in human pancreatic adenocarcinoma J Surg Res 1997 69 199 204 9202670 10.1006/jsre.1997.5022
Ilangumaran S Hoessli DC Effects of cholesterol depletion by cyclodextrin on the sphingolipid microdomains of the plasma membrane Biochem J 1998 335 ( Pt 2) 433 440 9761744
Koleske AJ Baltimore D Lisanti MP Reduction of caveolin and caveolae in oncogenically transformed cells Proc Natl Acad Sci U S A 1995 92 1381 1385 7877987
Engelman JA Wykoff CC Yasuhara S Song KS Okamoto T Lisanti MP Recombinant expression of caveolin-1 in oncogenically transformed cells abrogates anchorage-independent growth J Biol Chem 1997 272 16374 16381 9195944 10.1074/jbc.272.26.16374
Razani B Schlegel A Liu J Lisanti MP Caveolin-1, a putative tumour suppressor gene Biochem Soc Trans 2001 29 494 499 11498016 10.1042/BST0290494
Li L Yang G Ebara S Satoh T Nasu Y Timme TL Ren C Wang J Tahir SA Thompson TC Caveolin-1 mediates testosterone-stimulated survival/clonal growth and promotes metastatic activities in prostate cancer cells Cancer Res 2001 61 4386 4392 11389065
Hung KF Lin SC Liu CJ Chang CS Chang KW Kao SY The biphasic differential expression of the cellular membrane protein, caveolin-1, in oral carcinogenesis J Oral Pathol Med 2003 32 461 467 12901727 10.1034/j.1600-0714.2003.00185.x
Bockhorn M Roberge S Sousa C Jain RK Munn LL Differential gene expression in metastasizing cells shed from kidney tumors Cancer Res 2004 64 2469 2473 15059900
Chang F Steelman LS Shelton JG Lee JT Navolanic PM Blalock WL Franklin R McCubrey JA Regulation of cell cycle progression and apoptosis by the Ras/Raf/MEK/ERK pathway (Review) Int J Oncol 2003 22 469 480 12579299
Ouyang H Mou L Luk C Liu N Karaskova J Squire J Tsao MS Immortal human pancreatic duct epithelial cell lines with near normal genotype and phenotype Am J Pathol 2000 157 1623 1631 11073822
Liu N Furukawa T Kobari M Tsao MS Comparative phenotypic studies of duct epithelial cell lines derived from normal human pancreas and pancreatic carcinoma Am J Pathol 1998 153 263 269 9665487
van Golen KL Wu ZF Qiao XT Bao L Merajver SD RhoC GTPase overexpression modulates induction of angiogenic factors in breast cells Neoplasia 2000 2 418 425 11191108 10.1038/sj.neo.7900115
van Golen KL Wu ZF Qiao XT Bao LW Merajver SD RhoC GTPase, a novel transforming oncogene for human mammary epithelial cells that partially recapitulates the inflammatory breast cancer phenotype Cancer Res 2000 60 5832 5838 11059780
Kleer CG van Golen KL Zhang Y Wu ZF Rubin MA Merajver SD Characterization of RhoC expression in benign and malignant breast disease: a potential new marker for small breast carcinomas with metastatic ability Am J Pathol 2002 160 579 584 11839578
Evers EE van der Kammen RA ten Klooster JP Collard JG Rho-like GTPases in tumor cell invasion Methods Enzymol 2000 325 403 415 11036622
Diekmann D Hall A In vitro binding assay for interactions of Rho and Rac with GTPase- activating proteins and effectors Methods Enzymol 1995 256 207 215 7476434
Hendrix MJ Seftor EA Seftor RE Fidler IJ A simple quantitative assay for studying the invasive potential of high and low human metastatic variants Cancer Lett 1987 38 137 147 3690504 10.1016/0304-3835(87)90209-6
| 15969750 | PMC1173138 | CC BY | 2021-01-04 16:36:35 | no | Mol Cancer. 2005 Jun 21; 4:21 | utf-8 | Mol Cancer | 2,005 | 10.1186/1476-4598-4-21 | oa_comm |
==== Front
World J Surg OncolWorld Journal of Surgical Oncology1477-7819BioMed Central London 1477-7819-3-321592979910.1186/1477-7819-3-32Case ReportLong-term survival after an aggressive surgical resection and chemotherapy for stage IV pulmonary giant cell carcinoma Shoji Fumihiro [email protected] Riichiroh [email protected] Tatsuro [email protected] Jiro [email protected] Tomomi [email protected] Hiroshi [email protected] Yukito [email protected] Department of Thoracic Oncology, Kyushu Cancer Center, 3-1-1, Notame, Minami-ku, Fukuoka 811-1395, Japan2005 2 6 2005 3 32 32 25 1 2005 2 6 2005 Copyright © 2005 Shoji et al; licensee BioMed Central Ltd.2005Shoji et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Pulmonary giant cell carcinoma is one of the rare histological subtypes with pleomorphic, sarcomatoid or sarcomatous elements. The prognosis of patients with this tumor tends to be poor, because surgery, irradiation and chemotherapy are not usually effective.
Case presentation
We herein report a patient with pulmonary giant cell carcinoma with stage IV disease in whom aggressive multi-modality therapy resulted in a long-term survival. A 51-year-old male underwent an emergent operation with a partial resection of small intestinal metastases due to bleeding from the tumor. The patient also underwent a left pneumonectomy due to hemothorax as a result of the rapid growth of the primary tumor. Thereafter, two different regimens of chemotherapy and a partial resection for other site of small intestinal metastases and a splenectomy for splenic metastases were performed. The patient is presently doing well without any evidence of recurrence for 3 years after the initial operation.
Conclusion
This is a first report of a rare case with stage IV pulmonary giant cell carcinoma who has survived long-term after undergoing aggressive surgical treatment and chemotherapy.
==== Body
Background
The recent World Health Organization (WHO) classification of lung tumors has unified the heterogeneous group of non-small cell lung carcinomas that contains sarcoma or sarcoma-like components under the designation of "carcinomas with pleomorphic, sarcomatoid or sarcomatous elements" [1]. This group includes different entities, such as pleomorphic carcinoma, spindle cell carcinoma, giant cell carcinoma, carcinosarcoma and pulmonary blastoma. In general, these tumors are rare, comprising approximately from 0.1–0.4% of all lung malignancies [2-4]. The patients with these tumors tend to demonstrate a despondent clinical course and the prognosis for them is gernally poor [5], because surgery, irradiation and chemotherapy are ineffective. We experienced a pulmonary giant cell carcinoma patient with stage IV disease in whom aggressive multi-modality therapy consisting of surgical resections for the primary lesion and multi-organ metastases and also chemotherapy which together resulted in a long-term survival.
Case presentation
A 51-year-old male was admitted in June 2001, due to hemosputum, cough, hemo-stool and an abnormal shadow on a chest roentgenogram. Laboratory results showed severe anemia with hemoglobin of 4.0 g/dl (13.6 < normal range < 16.8 g/dl) and hematocrit of 16.0 % (40 < normal range < 48 %). The patient's chest X-ray demonstrated a huge mass lesion in the left upper lung field (Figure 1). Computed tomography (CT) of the chest showed a mass shadow, measuring 7.0 × 7.0 cm in size in the left upper lobe (S1+2) without any invasion of the surrounding tissue such as the vessels, plexus or thoracic wall and with no mediastinal lymph node swelling. Abdominal CT revealed a huge mass, measuring 12.7 × 7.5 cm in size in the small intestine. Prior to performing any treatment for the presumed lung cancer, we tried to stop the continuous bleeding from tumor in the small intestine. As a result, we performed an emergency operation. The tumor was observed in the jejunum at a location about 30 cm from the ligament of Treitz on the anal side and a 25 cm length of the jejunum, including the tumor, was thus resected. Six days later, the patient experienced sudden chest pain, dyspnoea and hemoptysis. The patient's chest X ray showed the left lung mass shadow to have rapidly increased in size, while the broncho-fiberscopy findings showed bleeding from the left upper bronchus and an obstruction of the left lower bronchus due to coagulation. Hemothorax due to a rupture of the lung induced by the rapid growth of the tumor was found after an emergency thoracotmy. The tumor was so large that it was difficult to approach the interlobular pulmonary artery. Therefore, a left pneumonectomy with mediastinal lymph nodal dissection was performed. Thereafter, intraoperative intrapleural hypotonic cisplatin treatment [6] was performed because some tumor cells were suspected to exist in the pleural cavity due to the rupture of the tumor. A histological examination revealed pure giant cell carcinoma containing no sarcomatoid component, similar to that found in the small intestine (Figure 2). As a result, we diagnosed the patient to have stage IV disease (pathological stage T2N0M1) according to the TNM classification [1]. The patient had an uneventful recovery without any complications. However, about 4 months after the first operation, the patient was diagnosed to have a recurrence at another site in the small intestine and spleen by abdominal CT. The patient received 2 cycles of chemotherapy (cisplatin 40 mg/m2 + gemcitabine 800 mg/m2+ vinorelbine 20 mg/m2), at days 1 and 8, and thereafter every 4 weeks). The splenic metastases increased in size while the size of the tumor in the small intestine decreased. At this time, no recurrence site except for those in the small intestine and spleen were found, therefore, to avoid the risk of bleeding either from tumors in the small intestine or a rupture of spleen in the future, surgical treatment consisting of a partial resection of the small intestine and a splenectomy was performed. The intestinal tumor was found in the jejunum at a location about 10 cm from the ligament of Treitz on the anal side and a total 20 cm length of the jejunum, including the tumor, was resected. A pathological examination revealed a proliferation of pure giant cell carcinoma with extensive necrosis both in the small intestine and the spleen, thus suggesting the chemotherapy to be effective in the both organs. Thereafter, the patient received 2 additional cycles of this triplet chemotherapy. The patient experienced neither any hematological nor severe non-hematological adverse events. About 6 months later, metastases in multiple abdominal lymph nodes were found (Figure 3A). The patient was started on chemotherapy (carboplatin AUC = 2 + paclitaxel 60 mg/m2, on days 1 and 8, and thereafter every 3 weeks). After receiving a total of 10 cycles of chemotherapy on an outpatient basis, abdominal CT showed the chemotherapeutic effect to be a complete response (Figure 3B), without any severe hematological or non-hematological adverse events. At present, the patient has survived for about 3-years since the first operation and a complete response has been maintained for 15 months.
Figure 1 Posterior-anterior view of a chest X-ray film demonstrated a huge mass shadow in the left upper lung field.
Figure 2 Pathological findings of the left lung. The section consists of a diffuse proliferation of atypical, giant and bizarre cells (arrowhead). No sarcomatoid component is seen.
Figure 3 Computerised tomographic scan before and after treatment. A) Abdominal CT showed multiple lymph node swelling, suggesting the presence of metastases (arrowhead). B) Abdominal CT showed the lymph nodes metastases to have completely disappeared.
Discussion
According to the treatment guidelines for unresectable non-small cell lung cancer of American Society of Clinical Oncology (ASCO)[7], chemotherapy prolongs survival and is the most appropriate treatment for stage IV patients with a good performance status. Although both resections of primary lung cancer and either brain or adrenal metastases are occasionally recommended in highly selected patients, a surgical resection of other metastasized sites is hardly ever performed. Therefore, the present patient is an extremely rare case because he underwent an emergency surgical resection of small intestinal metastases and a primary tumor due to bleeding from both tumors, as well as a surgical resection of other metastases in the small intestine and spleen in order to avoid a risk of bleeding from the recurrent site in the future.
Fishback et al, reported the overall survival of total 78 patients with pleomorphic (spindle/ giant cell) carcinoma (stage I-IV), among whom 57 patients received a surgical resection, to be poor with a median survival time of 10 months and a survival rate of 10% at 5 years [8]. According to Chang et al, the mean survival time of resected pleomorphic carcinoma patients was 5 months while the median survival time of pleomorphic carcinoma patients treated with concurrent or sequential chemo-radiotherapy was 2.7 months [9]. To our knowledge, a case of a long-term survivor with stage IV pleomorphic (spindle/ giant cell) carcinoma has never been previously reported. The tumor histology of the present case was very rare, pure giant cell carcinoma, which belongs to the category of carcinomas with pleomorphic, sarcomatoid or sarcomatous elements according to new WHO classification, and the prognosis is estimated to be poor. Although pleomorphic carcinoma has been reported to usually be resistant to chemotherapy, we first chose chemotherapy including cisplatin, gemcitabine and vinorelbine, which has been shown to demonstrate the highest response rate in advanced non-small cell lung cancer based on our experience. In our prior phase II trial using this combination chemotherapy in 79 advanced non-small cell lung cancer patients, the response rate was 56% and the 1-year survival rate was 75% while the toxicity levels were acceptable [10]. After recurrence, we chose chemotherapy with carboplatin and paclitaxel, which is most frequently used for the treatment of advanced non-small cell lung cancer. Since the standard treatment method using carboplatin and paclitaxel in Japan is the administration of AUC of 6 and 200 mg/m2, respectively once every 3 weeks [11], the administered regimen (carboplatin AUC = 2 and paclitaxel 60 mg/m2, on days 1 and 8, and thereafter every 3 weeks) in this patient was unusual and the dose intensity was relatively small. However, this regimen nevertheless effectively treated his disease and he was also able to work normally during the treatment process. At present, the patient has survived for 3 years since the first operation and has remained healthy without any signs of recurrence for 15 months after the last treatment.
Conclusion
This is a first report of a rare case with stage IV pulmonary giant cell carcinoma who has survived long-term after undergoing aggressive surgical treatment and chemotherapy.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
FS: Conceived the study, participated in its design and coordination and drafted the manuscript.
RM and TO: carried out the literature search and helped in drafting the manuscript
JI, TN and HW: participated in the study design and helped with preparation of the manuscript
YI: Shaped the idea for the manuscript, coordinated the study and edited the manuscript.
All authors conceived of the study, and participated in its design and coordination. All authors read and approved the final manuscript.
Acknowledgements
We thank Mr. Brian Quinn for critical comments on the manuscript.
Written consent was obtained from the patient for the publication of this case.
==== Refs
Travis WD Colby TV Corrin B Shimosato Y Brambilla E World Health Organization International Histological Classification of Tumors. Histological typing of Lung and Pleural Tumors 1999 3 Berlin: Springer Verlag
Travis WD Travis LB Devesa SS Lung cancer Cancer 1995 75 191 202 8000996
Brambilla E Travis WD Colby TV Corrin B Shimosato Y The new World Health Organization classification of lung tumors Eur Respir J 2001 18 1059 1068 11829087 10.1183/09031936.01.00275301
Travis WD Pathology of lung cancer Clin Chest Med 2002 23 65 81 11901921 10.1016/S0272-5231(03)00061-3
Colby TV Koss MN Travis WD Rosai J, Sobin LH Tumors of the lower respiratory tract Atlas of Tumor Pathology 1995 Washington DC: Armed Forces Institute of Pathology 259 275
Ichinose Y Tsuchiya R Koike T Yasumitsu T Nakamura K Tada H Yoshimura H Mitsudomi T Nakagawa K Yokoi K Kato H A prematurely terminated phase III trial of intraoperative intrapleural hypotonic cisplatin treatment in patients with resected non-small cell lung cancer with positive pleural lavage cytology: the incidence of carcinomatous pleuritis after surgical intervention J Thorac Cardiovasc Surg 2002 123 695 699 11986597 10.1067/mtc.2002.120712
Pfister DG Johnson DH Azzoli CG Sause W Smith TJ Baker S JrOlak J Stover D Strawn JR Turrisi AT Somerfield MR American society of clinical oncology treatment of unresectable non-small cell lung cancer guideline: Update 2003 J Clin Oncol 2004 22 330 353 14691125 10.1200/JCO.2004.09.053
Fishback NF Travis WD Moran CA Guinee DG JrMcCarthy WF Koss MN Pleomorphic (Spindle/Giant cell) Carcinoma of the Lung Cancer 1994 73 2936 2945 8199991
Chang YL Lee YC Shih JY Wu CT Pulmonary pleomorphic (spindle) cell carcinoma: peculiar clinicopathologic manifestations different from ordinary non-small cell carcinoma Lung Cancer 2001 34 91 97 11557118 10.1016/S0169-5002(01)00224-0
Ichinose Y Itoh K Imamura F Araki J Matsumoto M Yamamoto H Miura T Yoshino I Triplet combination chemotherapy of cisplatin (P), gemcitabine (G) and vinorelbine (V) with lower doses maintains a sufficient effect while avoiding toxicity in non-small cell lung cancer (NSCLC) patients J Clin Oncol 2004 22 7129
Kubota K Nishiwaki Y Ohashi Y Saijo N Ohe Y Tamura T Negoro S Ariyoshi Y Nakagawa K Fukuoka M The-four-arm-cooperative-study (FACS) for advanced non-small cell lung cancer (NSCLC) J Clin Oncol 2004 22 7006
| 15929799 | PMC1173139 | CC BY | 2021-01-04 16:39:03 | no | World J Surg Oncol. 2005 Jun 2; 3:32 | utf-8 | World J Surg Oncol | 2,005 | 10.1186/1477-7819-3-32 | oa_comm |
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1600001710.1371/journal.pbio.0030218Community PageEcologyScience PolicyNoneProtecting Science from Abuse Requires a Broader Form of Outreach Community PageChan Kai M. A [email protected] Paul A. T Porder Stephen 7 2005 12 7 2005 12 7 2005 3 7 e218Copyright: © 2005 Chan et al.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.Students and postdocs at Stanford University have formed an organization dedicated to promoting the use of sound science in policymaking: scienceinpolicy.org.
==== Body
The politicization of science may be as old as science itself. Famously, Galileo's championing of the theory that the Earth revolves around the sun met with staunch political opposition as a perceived challenge to the authority of the Catholic church. In the Soviet Union, Lysenko rejected the widely held chromosomal theory of inheritance in favor of a theory of environmental influences that aligned more closely with the philosophical underpinnings of communism. More recently, the assertion of the South African president Thabo Mbeki that AIDS is not caused by HIV flew in the face of decades of research and threatened to undermine proper treatment of the disease. In the view of many, science under the current United States administration is also under threat. Underlying many government policies that depend upon or affect science is a pattern by which evidence—which decision makers could use to craft well informed policies—is changed into a subjective tool for political or ideological goals [1–8]. The resulting perceptions of government hostility toward science threaten to drive frustrated federal scientists from agencies and to undermine the already flagging public respect for science. Ultimately, blame for the tension between science and politics does not lie with politicians alone. The minor role that scientists have played in the public arena allows such tensions to persist, and to grow elsewhere.
To confront this issue, we joined with other graduate students and postdocs at Stanford University to initiate scienceinpolicy.org—a grassroots organization dedicated to promoting the use of sound science in US policymaking. Assessing the major environmental issues in the public eye, we found and categorized a widespread pattern of manipulation and suppression of environmental science affecting issues as diverse as climate change, forestry policy, endangered species protection, clean water, and air pollution on our website (www.scienceinpolicy.org). For example, when the US Environmental Protection Agency sought to warn the public of health threats from air pollution following September 11, 2001, the White House edited press releases to significantly change the meaning from one of warning to placation. Because we researched and cited both news reports and the primary literature, and distributed our analysis for friendly peer review, our analysis attained a level of scientific credibility beyond that of regular media.
Professor Eric Weischaus (Princeton Professor of Molecular Biology, Nobel laureate), Dr. Diana Zuckerman (President of the National Research Council for Women and Families) and two student organizers participate in a public forum at Princeton University on scientific integrity in policymaking
(Photo: Princeton Environmental Action)
We alerted our colleagues to the information we had gathered by circulating emails, thereby 1) receiving additional informal peer review to ensure that our assessment and criticisms were fair and accurate, 2) empowering ourselves for additional outreach efforts such as writing op-eds and letters to the editor, appearing on radio programs, etc., and 3) establishing connections with other groups working to protect science. For example, when the Union of Concerned Scientists (Cambridge, Massachusetts, United States of America) spearheaded a more professional campaign with Nobel laureates and other prominent scientists, we seized the opportunity to act as a grassroots arm of their effort. We proposed a series of public forums on scientific integrity in policymaking at university campuses across the country, and initiated and promoted several of these events through our distribution list of almost 2,000 environmental scientists. At Princeton University (Princeton, New Jersey, United States of America), hundreds of students and community members packed the auditorium to overflowing.
Despite the favorable response, it was a challenge to marshal sufficient time and effort to execute an effective campaign. Impeding our progress was the very structure and culture of the academic community, in which research is valued above all else, and energy expended towards any other end is energy that could have been spent attempting to advance one's career in an extremely competitive job market. Despite apparently universal agreement on the need for scientist outreach to benefit both society and science (e.g., [9,10]), outreach is not encouraged institutionally, and may be actively deterred [11]. Outreach only rarely benefits young scientists striving for jobs and promotions. If adamant researchers persist with civic engagement, they frequently find themselves unprepared by a graduate training that emphasizes research skills almost exclusively.
Calling for greater scientist engagement in society is appropriate but cannot succeed without the institutional changes necessary to promote such behavior. Such changes will only occur when scientists change their own institutions. Reward structures, hiring and promotion decisions, and the structure and content of graduate training are all amenable to change by motivated students, postdocs, professors, and institutional officials. Since a healthy relationship between science and society may depend upon a marked reform of scientific training and oversight, scienceinpolicy.org is expanding its focus to encourage institutional changes that promote researchers reaching out to the media, the public, and policy makers.
To advance our vision of a vibrant academy that makes important contributions to a thriving society, we advocate a variety of concrete changes to academic culture and institutions. Graduate courses on effective outreach and communication with the media would empower interested students with needed abilities. New fellowship programs and awards for engagement like the Aldo Leopold Leadership Program (Stanford, California, United States of America) would provide training, camaraderie, and motivation for communicating with policy makers and the public. Institutional acceptance and/or support for an explicit time tithe for outreach would assuage the feeling that engaging with society means neglecting academic duty. New funding for outreach to, and partnerships with, governmental institutions, NGOs, etc, would provide both enticement and capital. Explicit consideration of outreach in promotion decisions—as exists in at least one institution—would reduce the opportunity costs associated with outreach and affirm the value of these activities [12]. We need to discuss and evaluate the numerous ways to appropriately encourage civic engagement, starting now.
One need only to look to the broader public for evidence that science is facing a crisis in the US. The public continues to debate the teaching of evolution in public schools. Meanwhile, the popularity of Michael Crichton's novel State of Fear—in which scientists are complicit in environmentalists' plot to manipulate the public by vastly exaggerating the threats of climate change—promotes and amplifies the already debunked claims of climate-change naysayers. But such public responses are only symptoms of a broader disconnect between science and society, a result of our academic isolation in the Ivory Tower.
Citation: Chan KMA, Higgins PAT, Porder S (2005) Protecting science from abuse requires a broader form of outreach. PLoS Biol 3(7): e218.
Kai M. A. Chan is with the Center for Conservation Biology at the Department of Biological Sciences at Stanford University, Stanford, California, United States of America. Paul A. T. Higgins is with the Department of Environmental Science Policy and Management at the University of California, Berkeley, California, United States of America. Stephen Porder is with the Department of Biological Sciences at Stanford University.
==== Refs
References
Union of Concerned Scientists Scientific integrity in policy making: Further investigation of the Bush administration's misuse of science 2004 Cambridge (Massachusetts) Union of Concerned Scientists 34
Union of Concerned Scientists Scientific integrity in policy making: An investigation into the Bush administration's misuse of science 2004 Cambridge (Massachusetts) Union of Concerned Scientists 49
United States House of Representatives Committee on Government Reform—Minority Staff Politics and science in the Bush administration 2003 (District of Columbia) Washington 39
Waxman HA Politics and science: Investigating the state of science under the Bush administration 2005 Available: http://democrats.reform.house.gov/features/politics_and_science/index.htm . Accessed 12 May 2005
Chan KMA Porder S Higgins PAT Kramer SB Concern is more than just ‘ruffled feathers’: If a government abuses science to justify its policies, scientists have a duty to speak out Nature 2004 428 255
Porder S Chan KMA Higgins PAT Scientists must conquer reluctance to speak out: When science is under political assault, keeping a dignified silence is counterproductive Nature 2004 431 1036
Higgins PAT The Bush administration and climate change Science 2004 306 2041 15609424
Association of Reproductive Health Professionals Preserving core values in science 2005 Washington (District of Columbia) Association of Reproductive Health Professionals
Marincola E Research advocacy: Why every scientist should participate PLoS Biol 2003 1 e71 10.1371/journal.pbio.0000071 14691543
Reid WV Bridging the science–policy divide PLoS Biol 2004 2 e27 10.1371/journal.pbio.0020027 14966530
Diamond JM Lessons from environmental collapses of past societies 2004 Washington (District of Columbia) National Council for Science and the Environment 29
Schlesinger WH Scientists must boldly address important environmental issues The Herald Sun (Durham) 2003 13 Sect A
| 16000017 | PMC1174820 | CC BY | 2021-01-05 08:21:25 | no | PLoS Biol. 2005 Jul 12; 3(7):e218 | utf-8 | PLoS Biol | 2,005 | 10.1371/journal.pbio.0030218 | oa_comm |
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1600002010.1371/journal.pbio.0030234FeatureDevelopmentScience PolicyHomo (Human)Follow the Money—The Politics of Embryonic Stem Cell Research FeatureRusso Eugene 7 2005 12 7 2005 12 7 2005 3 7 e234Copyright: © 2005 Eugene Russo.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.Gene Russo examines the broader implications of Proposition 71 - a California initiative to fund and promote research into human embryonic stem cells.
==== Body
German embryonic stem cell scientist Oliver Brustle faces major challenges in his lab on a daily basis that have little to do with science. His country's policy on human embryonic stem cells (hESCs) is among the most restrictive in Europe. Collaborations with other countries are difficult. And he's overwhelmed by the time and paperwork necessary to comply with regulations and to apply for permits.
Brustle, an investigator at the University of Bonn (Bonn, Germany), often ponders the possibility of taking his work elsewhere. “I think about it every day,” he says. “Sometimes [the regulations] get to the point where they're suffocating.” Many of his German colleagues have similar sentiments, he says. In Germany, scientists can only work on hESC lines derived in labs outside the country, and even then, only if the line was derived prior to 2002. Federal law prohibits the derivation of new hESC lines.
Yet Brustle is optimistic that hESC technology, flush with research dollars in many countries and garnering interest worldwide, is destined for greatness. He takes solace in the fact that his long-time scientific pursuit has potential for success on several different fronts. Indeed, though some hESC scientists like Brustle are frustrated, others are thriving.
The most recent front is California, where Proposition 71 [1], passed last November, sent science policy shock waves across the United States and around the world (Box 1). Promising an impressive US$3 billion in funding for hESC research over the next ten years, “Prop 71” could further complicate an already complex landscape of laws and funding sources (assuming supporters can prevail over legislation that has stalled the measure). Other states in the US are struggling to catch up, setting the stage for a patchwork of policies and funding efforts.
Box 1. Proposition 71
California's Proposition 71 calls for US$3 billion of stem cell research funding over 10 years. The money will be raised by general fund bonds. The measure includes the establishment of the California Institute for Regenerative Medicine (http://www.cirm.ca.gov/) to regulate and oversee stem cell research, a mission that entails managing ethical research practices and awarding grants through peer review (peer reviewers will be based outside the state of California). The Institute's 29-member Independent Citizens Oversight Committee consists of representatives from California universities, nonprofit institutions, patient advocacy groups, and the biotechnology industry. Grant reviews will be by out-of-state scientists to prevent conflicts of interest. Unlike the policy dictated by President Bush in August 2001, Prop 71 permits the funding of the derivation of new human embryonic stem cell lines, including via SCNT research (therapeutic cloning). Currently, 78 hESC lines around the world are eligible for federal US funding, but only 22 have been developed into distribution-quality cell lines. With interest, Prop 71 is projected to cost the state approximately US$6 billion if paid over a 30-year period. Lawsuits attempting to stop Prop 71 are currently pending, according to a spokesperson for the California Institute for Regenerative Medicine. It's not yet clear, she said, for how long the litigation will stall the sale of bonds and the awarding of research dollars.
Countries from Europe to Asia also have a patchwork of different policies (Figure 1), as economic and scientific interests clash with religious and cultural mores in a battle over research that has yet to help a single patient. And while many non-Californian stem cell researchers are cheering the Prop 71 windfall, they're wary of brain drain to California in a field that already has a shortage of talent. Some supporters also worry that a mosaic of state laws will make collaborations more difficult and future therapies hard to administer. In short, Prop 71 is likely to help shape the future of stem cell research in the US and around the world.
Figure 1 Stem Cell Policy Map
Countries colored in brown have a permissive or flexible policy on human embryonic stem cell research. All have banned human reproductive cloning. These countries represent about 3.4 billion people, more than half the world's population. “Permissive” (countries in dark brown) means that various embryonic stem cell derivation techniques are permitted, including SCNT. “Flexible” (countries in light brown) means that stem cells may be derived from human embryos donated by fertility clinics only, excluding SCNT. Countries in yellow have either a restrictive policy or no established policy.
(Image: William Hoffman, MBBNet)
An Intriguing Proposition
Stanford University (Stanford, California, United States of America) stem cell scientist Irv Weissman, who played a key role in the multimillion dollar effort that recruited researchers, advocates, and Hollywood actors to support Prop 71, attributes the effort's inception to a 2002 US National Academy of Sciences report on cloning. It contended that therapeutic cloning, or somatic cell nuclear transfer (SCNT), could be used to generate disease-specific stem cell lines; the research had more future applications than just cell and organ transplantation (Box 2). Scientists might be able to actually revert diseased cells to their primordial form and then monitor them to see how and why abnormalities develop. That notion sparked interest from many people unhappy with President Bush's policy, which dictates that no federal funding may be used to work on hESC lines derived after August 2001. (On May 24 the US House of Representatives passed a bill that would make more cell lines available for federal funding. Bush has vowed to veto the bill.) Also of concern: stem cell lines approved by the National Institutes of Health (NIH) (http://stemcells.nih.gov/research/registry/), which can only be grown on a layer of mouse feeder cells, would not be appropriate for clinical use since animal viruses could theoretically jump to humans. According to some, the president's policy was simply too restrictive.
Box 2. Sources of hESCs
Researchers primarily obtain human embryonic stem cells from frozen embryos donated by IVF programs. Cell lines may be grown by isolating hESCs from the inner cell mass of a human blastocyst, a five-day-old embryo. These cells are cultured indefinitely with the help of fibroblast feeder layers. Several groups are now attempting to grow feeder-free cell lines, which would enable investigators to know exactly what molecular factors are contributing to hESC growth [2]. According to Austin Smith, director of the Institute for Stem Cell Research at the University of Edinburgh, if scientists can discover the minimum requirements for self-renewal, they'll be able to better control and direct robust hESC growth, which could be advantageous for both basic and clinical hESC research.
Researchers may also obtain hESCs via SCNT, as was demonstrated last year by a group in South Korea [3]. The technique involves removing the nucleus, and hence the nuclear genome, of an oocyte and replacing it with the nucleus of an adult cell. The activated egg can then form a blastocyst, which contains identical genetic material to that of the donor adult cell, and this is the process often referred to as therapeutic cloning. By using SCNT rather than a donated IVF embryo, researchers would be able to control the genotype of hESCs, which could help circumvent tissue rejection problems in future clinical applications as well as help investigators study the development of diseases via disease cell lines. In May, Hwang et al. became the first to derive patient-specific human embryonic stem cells from SCNT blastocysts [4]. Even when countries allow SCNT, lawfully securing donated oocytes can be a challenge. Last year, Hwang et al. [3] used 200 oocytes before deriving a single cell line. His more recent study [4] was much more efficient. His group's work has yet to be widely repeated. If implanted in a uterus, a blastocyst generated via nuclear transfer could theoretically develop into a human being genetically identical with the nuclear donor.
And then Hollywood got involved. Producers Janet and Jerry Zucker, who have a child with diabetes, asked Weissman to educate them about stem cells, which led to the formation of an advocacy organization called CuresNow (Los Angeles, California, United States of America). State Senator Debra Ortiz took up the cause, and suggested that California support research where the NIH could not. Soon after, the politically savvy Robert Klein, a board member of the Juvenile Diabetes Foundation, began spearheading the Prop 71 effort. (Klein has a child with diabetes as well.) Weissman and other scientists conducted dozens of talks, educating the public about stem cells and cautioning them about the potentially slow pace of hESC research. Governor Arnold Schwarzenegger decided to endorse the plan. And in November of last year, voters voiced their approval.
Though some states already had initiatives under way, Prop 71 spurred and accelerated efforts. Many states are now engaged in a race to attract stem cell research with laws and regulations that defy the Bush administration policy. In some cases, states are pushing for funding packages to offset the federal funding shortfall and insure that their most promising scientists don't head west. And in other cases, states are reacting strongly in the opposite fashion, considering major restrictions on the sort of stem cell research that scientists in their state can conduct.
“As this issue has hit the state level, there has been a noticeable change in the terms of the debate” says Daniel Perry, executive director of Coalition for the Advancement of Medical Research (Washington, District of Columbia, United States of America), which represents patient organizations, universities, scientific societies, and foundations. Before Proposition 71, moral status questions were at the fore, e.g., debates as to when life begins and how society should balance the rights of a potential life with those of a person suffering from a life-threatening disease. “Now the debate is all about economic development,” says Perry. “Are we going to be creating jobs in Delaware or Illinois or Pennsylvania or are we going to be losing them to states next door that are in the stem cell business?” Stem cell research supporters fear their states, many with a history of biotechnology and life sciences investment, will lose talent, dollars, and infrastructure to California or other states if they don't protect their own research investment.
State by State
More than 30 US states are considering some sort of legislation related to stem cell research, for or against. Particularly active, according to Perry, are Wisconsin, Illinois, New York, Delaware, Texas, Florida, Washington, and Missouri. In Illinois, the state house and senate are considering a bill that imposes a surcharge on elective cosmetic surgery as a way to raise an estimated US$1 billion for stem cell research over the next ten years. In New Jersey, with the support of Governor Richard Cody, the legislature approved US$150 million in state funding to build a Stem Cell Institute. New Jersey citizens will vote on a US$230 million bond this November that would finance stem cell research over the next seven years. Connecticut recently approved $100 million for stem cell research.
The senate and house in Massachusetts recently voted to explicitly allow all aspects of stem cell research, including work on surplus embryos and SCNT. Legislators had enough votes to override Massachusetts governor Mitt Romney's veto. “For those of us worried that everything could get shut down, this is a big relief,” says Leonard Zon, a Howard Hughes Medical Institute (Chevy Chase, Maryland, United States of America) investigator and the director of the stem cell research program at Children's Hospital Boston (Boston, Massachusetts, United States of America). Zon, who works on deriving blood stem cells from embryonic stem cells, spent considerable time helping to support the legislation. Massachusetts has some high-profile stem cell researchers, and many feared that California would be too tantalizing to resist. According to a spokesperson for Romney, the governor supports hESC research but not SCNT. He contends that only a few Massachusetts companies are actually focused on stem cell research, and hence he's not concerned about scientists migrating west en masse.
In Maryland, a bill to fund hESC research was narrowly defeated in April of this year after months of political wrangling by a divided legislature. Like Governor Romney, Republican governor Robert Ehrlich supports, according to a spokesperson, embryonic stem cell research and President Bush's policy. But he has taken no official position on SCNT, nor on the recently defeated bill (which never actually got to his desk). Using money garnered from cigarette company restitution, the bill would have funneled US$25 million to embryonic stem cell research on discarded embryos, but would not have supported SCNT. Many in Maryland are keen to protect the state's life sciences investment. Already, though, California schools are making inquiries. “We certainly are having our share of people who are being recruited away at all levels,” says John Gearhart, a professor of biochemistry and molecular biology at Johns Hopkins University (Baltimore, Maryland, United States of America) and one of the first scientists to derive pluripotent human stem cells in 1998 (Figure 2). Gearhart is encouraged that the bill came within one vote of passing, and hopes that supporters will have better luck in the next legislative session. But he notes that time is running out; the Maryland legislature doesn't meet again until January of 2006, and any program that's approved would require a couple of years to ramp up.
Figure 2 Embryonic Stem Cells
(A) shows hESCs.
(B) shows neurons derived from hESCs.
(Images: Nissim Benvenisty)
Other states, actually acting prior to Prop 71 based on concerns about researchers doing any sort of cloning, therapeutic or otherwise, have enacted outright restrictions, mostly with regard to SCNT. Arkansas, North and South Dakota, Iowa, and Michigan all have laws specifically restricting aspects of hESC research. A Missouri state senate bill that would have criminalized SCNT recently stalled. In some states, including Texas, Illinois, and New York, lawmakers have proposed bills both for and against stem cell research.
Best of a Bad Regulatory Situation
Despite the positive activity in many places, researchers like Gearhart and Weissman admit that it's only the best of a bad situation. A blanket federal funding policy would be much preferable. “I like to say that the glass is half full,” says Carl Gulbrandsen, director of the nonprofit WiCell Research Institute, which supports human embryonic stem cell research at the University of Wisconsin-Madison (Madison, Wisconsin, United States of America). Gulbrandsen distributes hESC lines derived by Wisconsin stem cell researcher Jamie Thomson, another pioneer of hESC research. It's encouraging, notes Gulbrandsen, that a large state like California, already suffering from severe deficit, has supported stem cell research. And Prop 71 has raised the value of the research in terms of licensing opportunities, with companies, both biotechs and pharmas, showing increased interest.
But Prop 71 also promises to increase the cost of doing research by making the attraction and retention of faculty more expensive as California and potentially other states offer millions in grants and salaries to stem cell researchers. “It's going to be very difficult,” says Gulbrandsen. “We're going to have to open the pocketbook and spend a lot of money to make sure the research goes forward.” In general, the state-by-state funding model “is a very poor public policy,” he adds. “If you end up with a patchwork of regulations, you're going to both decrease the movability of scientists and you're going to discourage collaborations.” Gulbrandsen suggests that intellectual property could be an issue as well, with states passing their own regulations, controlling their own intellectual property, and hence potentially creating turf wars among themselves and conflicts with federal policy. He says, “It's not a good road to walk down.”
Johns Hopkins president William Brody, though he supported funding efforts in Maryland, calls the growing influence of state legislators on the lab “scientific tourism.” “It's just going to create havoc about what you can do where, and I don't think in the end it's good science policy or good health care policy,” he says, noting that many states have actually restricted the research, in effect putting in place harsher regulations than those of the federal government. And regulations in other types of research could follow, Brody contends.
The quality of the research is at issue as well. Brody worries that smaller states won't be able to replicate rigorous grant peer review. Weissman is most concerned that the considerable monies available could lead to the wasting of funding on sub-par research. He advocates carrying the money forward until it can be well spent, perhaps on expensive clinical trials that are years away.
The Proposition Heard around the World
Meanwhile, researchers overseas are taking notice of Prop 71, also with a mix of support and trepidation. “There's been a little bit of complacency in Europe because of the Bush decree,” says Austin Smith, director of the Institute for Stem Cell Research at the University of Edinburgh (Edinburgh, United Kingdom), and coordinator of EuroStemCell (Edinburgh, United Kingdom), an international consortium for stem cell research funded by the European Union. “European institutions have got to get their act together, because otherwise all our best people will go to California or other states setting up programs.” Alan Trounson, director of Monash Institute of Reproduction and Development in Clayton, Australia, has little doubt that California's stem cell funding riches will drain away talent from down under. “It will certainly be a very attractive place for young people and mid-career people to go,” he says, noting that Prop 71 should help give young stem cell investigators confidence that their field has a secure future.
Laws regulating hESC research in Europe and around the world vary considerably. Countries like the United Kingdom, Sweden, and Singapore are among the most liberal. In the UK, researchers are allowed to use embryonic stem cells from discarded embryos, and they're allowed to create embryos for the purposes of research. The approval process, however, can be laborious, according to Smith, though it has improved considerably in the last couple of years. Researchers must get approval from their local hospital ethical board, then they must apply for a license from the Human Fertilisation and Embryology Authority. HFEA inspects the quality of the science, the researcher's justification for using human embryos, whether scientists are complying with the law, and whether they're following correct patient consent procedures. Once a license is granted, HFEA conducts yearly inspections to insure that every embryo is accounted for.
In the beginning of April 2005, Sweden's government specifically approved of producing embryonic stem cell lines using SCNT. Using hESCs from leftover embryos had already been allowed. Protections in place are akin to those of the UK. “I'm very happy with the Swedish law,” says Outi Hovatta, a stem cell researcher at the Karolinska Institutet in Stockholm, Sweden. “There are clear regulations. We know what we can do, and we know what we can't do.”
After a year of public education outreach that included discussions with bioethicists, scientists, and religious leaders, Singapore, having already invested substantial federal funds into the life sciences, established specific hESC regulations.
The rules allow research on surplus in vitro fertilization (IVF) embryos using federal funding, and allow SCNT on a case by case basis. They also prohibit reproductive cloning; anyone caught attempting reproductive cloning receives a S$100,000 fine and ten years in jail. “I've never seen so much support from a government when it comes to stem cell biology,” Ariff Bongso, a professor of obstetrics and gynecology at the National University of Singapore (Figure 3). Bongso was the first person to isolate embryonic stem cells from human embryos in 1994. To derive hESCs, researchers must go to the hospital institutional review board, then to the government ministry of health, then to a government advisory bioethical council. Stem cell excitement has reached a fever pitch in Singapore, according to Bongso. “Everybody's talking about stem cells,” he says.
Figure 3 Ariff Bongso
Ariff Bongso is a professor of obstetrics and gynecology and a pioneer of human embryonic stem cell research at the National University of Singapore.
Countries like Israel and Australia have somewhat liberal laws, thought not quite as permissive. In Australia, laws used to be state by state, and cell lines could only be harvested from IVF embryos frozen prior to 2002. But a federal law now permits research on any frozen embryo, with the usual consent and ethical review. For now, SCNT has not been endorsed. In Israel, researchers can use embryonic stem cells from frozen embryos, but egg donation is not allowed, making SCNT all but impossible. Nissim Benvenisty, a stem cell researcher at the Hebrew University of Jerusalem (Jerusalem, Israel), says that many families are happy to donate discarded embryos that have been diagnosed with genetic diseases so that researchers can study diseased cell development. Countries like Switzerland, Spain, and France allow stem cell research with some restrictions.
Austria and Germany are much more restrictive. Brustle attributes his country's regulations to religious views, a legacy of eugenics, and negative attitudes towards new technologies. Brustle finds it difficult to conduct collaborations and he finds funding set-up somewhat nonsensical. For example: The European Commission, which will fund established stem cell lines, has a more liberal policy than Germany. Germany's European Commission stem cell research contribution ends up being funneled to countries with more liberal policies. Germany, then, indirectly pays for research in other countries that it actually deems illegal.
European collaborative efforts, though, continue to pick up steam. EuroStemCell, organized via the European Union, has 27 different member labs throughout Europe, with interests in adult and embryonic stem cells, clinical and basic research. According to Smith, the organization's charge is the exchange of materials and people. Smith would like to build up a cadre of European researchers trained in stem cell biology. “We have to collaborate in Europe to be competitive with the US,” he says. “It's not so clear that that's an imperative if you're sitting in California with all that money.”
Citation: Russo E (2005) Follow the money—The politics of embryonic stem cell research. PLoS Biol 3(7): e234.
Eugene Russo is a freelance science writer based in Takoma Park, Maryland, United States of America. E-mail: [email protected]
Abbreviations
hESChuman embryonic stem cell
HFEAHuman Fertilisation and Embryology Authority
IVFin vitro fertilization
NIHNational Institutes of Health
SCNTsomatic cell nuclear transfer
==== Refs
References
California Secretary of State Proposition 71: Stem cell research Funding. Bonds. Initiative constitutional amendment and statute 2004 Sacramento (California) California Secretary of State Available: http://www.ss.ca.gov/elections/bp_nov04/prop_71_entire.pdf . Accessed 16 May 2005
Xu RH Peck RM Lee DS Feng X Ludwig T Basic FGF and suppression of BMP signaling sustain undifferentiated proliferation of human ES cells Nat Methods 2005 17 185 190
Hwang WS Ryu YJ Park JH Park ES Lee EG Evidence of a pluripotent human embryonic stem cell line derived from a cloned blastocyst Science 2004 303 1669 1674 14963337
Hwang WS Roh SI Lee BC Kang SK Kwon DK Patient-specific embryonic stem cells derived from human SCNT blastocysts Science 2005 In press
| 16000020 | PMC1174821 | CC BY | 2021-01-05 08:21:24 | no | PLoS Biol. 2005 Jul 12; 3(7):e234 | utf-8 | PLoS Biol | 2,005 | 10.1371/journal.pbio.0030234 | oa_comm |
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1600002110.1371/journal.pbio.0030245EssayDevelopmentEvolutionGenetics/Genomics/Gene TherapyHomo (human)PrimatesArthropodsTeleost FishesEvolution at Two Levels: On Genes and Form EssayCarroll Sean B 7 2005 12 7 2005 12 7 2005 3 7 e245Copyright: © 2005 Sean B. Carroll.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.Emerging knowledge about organismal evolution suggests that changes in the regulation of gene expression have played a major role - a thesis proposed 30 years ago by King and Wilson.
==== Body
In their classic paper “Evolution at Two Levels in Humans and Chimpanzees,” published exactly 30 years ago, Mary-Claire King and Allan Wilson described the great similarity between many proteins of chimpanzees and humans [1]. They concluded that the small degree of molecular divergence observed could not account for the anatomical or behavioral differences between chimps and humans. Rather, they proposed that evolutionary changes in anatomy and way of life are more often based on changes in the mechanisms controlling the expression of genes than on sequence changes in proteins.
This article was a milestone in three respects. First, because it was the first comparison of a large set of proteins between closely related species, it may be considered one of the first contributions to “comparative genomics” (although no such discipline existed for another two decades). Second, because it extrapolated from molecular data to make inferences about the evolution of form, it may also be considered a pioneering study in evolutionary developmental biology. And third, its focus on the question of human evolution and human capabilities, relative to our closest living relative, marked the beginning of the quest to understand the genetic basis of the origins of human traits. Like much of Wilson and his colleagues' body of work, this contribution had a great influence on paleoanthropologists as well as molecular biologists.
The 30th anniversary of this landmark article arrives at a moment when comparative genomics, evolutionary developmental biology, and evolutionary genetics are pouring forth unprecedented amounts of new data, and the entire chimpanzee genome is available for study. It is therefore an opportune time to examine what has been and is being revealed about the relationship between evolution at the two levels of molecules and organisms, and to assess the status of King and Wilson's hypothesis concerning the predominant role of regulatory mutations in organismal evolution.
King and Wilson used the phrase “ways of life” to include both physiology and behavior (M.-C. King, personal communication) and proposed that the evolution of both anatomy and ways of life was governed by regulatory changes in the expression of genes. From the outset of this review, I make the sharp distinction between the evolution of anatomy and the evolution of physiology. Changing the size, shape, number, or color patterns of physical traits is fundamentally different from changing the chemistry of physiological processes. There is ample evidence from studies of the evolution of proteins directly involved in animal vision [2], respiration [3], digestive metabolism [4], and host defense [5] that the evolution of coding sequences plays a key role in some (but not all) important physiological differences between species. In contrast, the relative contribution of coding or regulatory sequence evolution to the evolution of anatomy stands as the more open question, and will be my primary focus.
The amount of direct evidence currently in hand is modest, and includes examples of both the evolution of coding and of non-coding, regulatory sequences contributing to morphological evolution. However, I will develop the argument, on the basis of theoretical considerations and a rapidly expanding body of empirical studies, that regulatory sequence evolution must be the major contributor to the evolution of form.
This conclusion poses particular challenges to comparative genomics. While we are often able to infer coding sequence function from primary sequences, we are generally unable to decipher functional properties from mere inspection of non-coding sequences. This has led to a bias in comparative genomics and evolutionary genetics toward the analysis and reporting of readily detectable events in coding regions, such as gene duplications and protein sequence evolution, while non-coding, regulatory sequences are often ignored. However, approximately two-thirds of all sequences under purifying selection in our genome are non-coding [6]. One consequence of the underconsideration of non-coding, regulatory sequences is unrealistic expectations about what can currently be learned about the genetic bases of morphological diversity from comparisons of genome sequences alone. The visible diversity of any group is not reflected by the most visible components of gene diversity—that is, the diversity of gene number or of coding sequences. In order to understand the evolution of anatomy, we have to study and understand regulatory sequences, as well as the proteins that connect them into the regulatory circuits that govern development. I will begin with some historical and theoretical considerations about regulatory and coding sequence evolution, then delve into the insights offered by specific experimental models of anatomical evolution, and finally, I will revisit King and Wilson's original focus and discuss how our emerging knowledge of the evolution of form bears on current efforts to understand human evolution.
A Brief History of Regulatory Thinking
Almost 50 years ago, as the first sequences of various proteins from different species were determined, the potential significance of macromolecules for understanding evolutionary processes was quickly recognized [7]. The great similarity among homologous proteins of different species was noted early [8] and raised the question to what degree such sequence changes were functionally significant [9]. With the advent of the operon model of gene regulation [10], some biologists such as Emile Zuckerkandl began to consider the possible role of “controller genes” in evolution, including in the origin of humans from ape ancestors [11]. One of the most widely noted series of theoretical contributions in this period was Roy Britten and Eric Davidson's models for gene regulation in higher organisms, which had an explicit emphasis on the importance of gene regulation in evolution [12,13].
The most influential single publication of this era, however, was Susumu Ohno's book Evolution by Gene Duplication [14]. Ohno focused on the importance of gene redundancy in allowing “forbidden” mutations to occur that could impart new functions to proteins. His opening motto, “natural selection merely modified, while redundancy created,” reflected a view of natural selection as a largely purifying, conservative process. Ohno insisted that “allelic mutations of already existing gene loci cannot account for major changes in evolution.” He proposed that the duplication of regulatory genes and their control regions must have contributed greatly to the evolution of vertebrates. But the book focused exclusively on the evolution of new proteins and did not consider the creative potential of non-coding, regulatory sequences in evolutionary diversification (see [15]).
It was against this backdrop that Allan Wilson and his colleagues began a series of investigations into the relationship between chromosomal evolution, protein evolution, and anatomical evolution in birds [16], mammals [17], frogs [18], and apes [1]. In each of four studies, the discrepancy between the evolution of proteins and the evolution of anatomy led to the conclusion that evolutionary changes in “regulatory systems” were responsible for the evolution of anatomy. Francois Jacob similarly suggested that divergence and specialization result from mutations altering “regulatory circuits” rather than chemical structures [19].
The relative contributions of different mechanisms to the evolution of anatomy depend upon both what is genetically possible, and what is permitted by natural selection. Before I delve into the data directly concerning the evolution of anatomy, and how well it fulfills King and Wilson's original expectations, it will be valuable to consider what mechanisms are available and what parameters will govern their utilization in evolution, in light of what we now understand about how genes function in multicellular organisms.
Pleiotropy and the Genetic Architecture of Multicellular Organisms
One critical parameter that affects the relative contribution of different genetic mechanisms to anatomical variation is the pleiotropy of mutations [20]. In general, it is expected that mutations with greater pleiotropic effects will have more deleterious effects on organismal fitness and will be a less common source of variation in form than mutations with less widespread effects.
Over the past 30 years, several key features of gene structure, function, and regulation in multicellular organisms have emerged that govern the pleiotropy of mutations and thus shape the capacity of species to generate anatomical variation and to evolve (see Box 1). Because of these features, mutations in different genes and different parts of genes (that is, non-coding and coding sequences) can differ dramatically in their degree of pleiotropy. For example, a mutation in the coding region of a transcription factor that functions in multiple tissues may directly affect all of the genes the protein regulates. In contrast, a mutation in a single cis-regulatory element will affect gene expression only in the domain governed by that element.
Box 1. Key Genetic Features of Multicellular Organisms
Individual regulatory proteins function in many different contexts. Signaling proteins, their receptors, signal transducers, and most transcription factors are deployed in multiple tissues, organs, or body parts. The function of each regulatory protein is context-dependent, with different genetic targets and morphogenetic outcomes in different tissues.
The expression of individual genes is regulated by multiple, modular cis-regulatory elements. The tissue-specific and temporal control of gene expression, particularly of genes encoding the regulatory proteins that shape pattern formation and cell differentiation in animals, is typically governed by arrays of discrete regulatory elements embedded in regions that flank coding regions and lie within introns [23].
Many regulatory proteins are members of large families and can overlap in function. More than 20% of human genes and a much larger fraction of plant genes belong to families [75] that are the product of the duplication and evolutionary divergence of ancestral genes.
Multiple protein forms may be encoded by single genetic loci. Through the use of alternative promoters and RNA splice sites, multiple mRNAs encoding different protein products are often produced from a single locus. Alternative protein forms (isoforms) may function in different contexts and/or possess different activities.
John Gerhart and Marc Kirschner [21,22] have discussed in depth how certain features of animal genetic regulatory systems influence “evolvability”—the capacity to generate tolerable, heritable variation. For instance, redundancy reduces constraint on change by circumventing or minimizing the potentially deleterious effects of some mutations. Compartmentation also facilitates change; by uncoupling variation in one process from variation in another, pleiotropy is decreased.
Several genetic features contribute to redundancy and compartmentation. For example, gene duplication creates initially redundant paralogs. Mutations that may have been deleterious in the ancestral gene may be tolerated and allow for the “exploration” of new variation, which can occur in coding or regulatory sequences, or both (Figure 1A). Likewise, the expanded number and diversity of cis-regulatory elements establishes compartmentation by enabling the independent control of gene transcription in different body parts (Figure 1B). The use of alternative promoters and RNA splice sites also contributes to compartmentation by enabling tissue- or cell-typespecific production of alternative forms of a protein (Figure 1C). Variation may arise either in regulatory sequences governing promoter use or splice site choice, or in coding sequences of exons. The three mechanisms gene duplication, regulatory sequence expansion and diversification, and alternative protein isoform expression accomplish essentially the same general result—they increase the sources of variation and minimize the pleiotropy associated with the evolution of coding sequences. The global question of the genetic basis of the evolution of form then boils down to the relative contribution of gene duplication, regulatory sequence evolution, and the evolution of coding sequences, over evolutionary time. I will first examine what is known about the role of regulatory sequences and then discuss the contributions of coding sequences and gene duplication to the evolution of anatomy.
Figure 1 Different Modes of Gene Evolution Increase the Diversity of Gene Function and Minimize Pleiotropy
The function of a progenitor gene with the simple structure of one cis-regulatory element (red circle) and a pair of exons (black rectangles) can be expanded and diversified in several ways.
(A) Gene duplication followed by mutations (asterisks) in either coding or regulatory sequences of the initially identical paralogs will produce genes that may be expressed in different ways or proteins with distinct functions, while the original function can be maintained.
(B) An expansion in the number of cis-regulatory elements by any of a number of means (transposition, rearrangement, duplication, point mutation) can expand the number of tissues in which the gene is active, but preserves the original function.
(C) The evolution of a new exon and splicing sites creates the potential for alternative forms of a protein to be made. Mutations in alternative exons (asterisks) need not affect the original function of the protein.
Regulatory Sequences and the Evolution of Anatomy
Over the past decade or so, comparative studies of gene expression in diverse animals and plants, across all taxonomic levels, have revealed a general association between the gain, loss, or modification of morphological traits and changes in gene regulation during development [23,24]. Changes in the expression of an individual gene may evolve through alterations in cis-regulatory sequences or in the deployment and activity of the transcription factors that control gene expression, or both.
Progress toward elucidating the mechanisms governing the evolution of specific traits and genes has required the study of models in which genetic and molecular methods enable the identification and dissection of functional differences among populations or species. The traits and species for which such detailed analysis has been possible include the trichome [25–27], bristle [28], and pigmentation patterns [29] in fruit flies; flower coloration [30], architecture [31], and branch patterns [32] in plants; and limb [33] and axial diversity in vertebrates [34].
A handful of studies have genetically demonstrated that evolution at particular loci has affected the gain [32], loss [26,27,33], or modification of morphological traits [25]. These studies—highlighted below—have firmly eliminated coding sequences as a possible cause and thereby implicated regulatory sequence evolution at loci encoding pleiotropic transcription factors. In a few cases, direct evidence of functional changes in cis-regulatory elements has been obtained [34–36].
Fruit flies display all sorts of conspicuous patterns of black pigmentation on their head, thorax, abdomen, and wings. These patterns are regulated by a variety of well-conserved signaling pathways and transcription factors that control the spatial expression of the enzymes that promote or inhibit the formation of the pigment melanin [37]. In Drosophila melanogaster and other members of the genus, structural genes, such as yellow, are regulated by an array of cis-regulatory elements that govern their expression in different body parts, such as the wing and abdomen [36] and the bristles and larval mouthparts. This modular arrangement of cis-regulatory elements had suggested that gene expression and pigment patterns evolve independently in different body parts through changes in individual cis-regulatory elements. Recent studies have demonstrated this to be exactly the case [35,36] (Figure 2A).
Figure 2 The Modular Architecture of the cis-Regulatory Regions of Pleiotropic Genes Enables the Independent Evolution of Gene Expression in Different Body Parts
(A) Expression of the yellow pigmentation gene of Drosophila is controlled by several different cis-regulatory elements (red circles). Differences in the activity of selected elements (wing and wing spot) underlie differences in pigment patterns between species (Figure based on [35].)
(B) Similarly, the expression of the Pitx1 gene of vertebrates is inferred to be controlled by multiple elements (red circles). In pelvic-reduced stickleback fish, Pitx1 expression is absent from the pelvic region. This is proposed to occur through of a selective loss of activity of the hindlimb regulatory element (cross through the red circle) (Figure based on [33].)
There are several salient general features of the evolution of pigment patterns in fruit flies. Many or all of the structural genes involved are pleiotropic; they have roles in multiple parts of the body and in other physiological processes (for example, neurotransmitter synthesis and behavior). Furthermore, they are regulated, at least in part, by widely deployed, highly conserved pleiotropic regulatory proteins, some of which are themselves regulated by deeply conserved and evolutionarily stable global regulators of body pattern formation [29]. Thus, while the coding sequences of the structural and regulatory proteins are constrained by pleiotropy, modular cis-regulatory regions enable a great diversity of patterns to arise from alterations in regulatory circuits through the evolution of novel combinations of sites for regulatory proteins in cis-regulatory elements [35]. This diversity is produced by the sort of “tinkering” with existing components envisaged by Jacob [19].
Is what is true of coloration, true of more complex traits? It is possible that because body color patterns are so critical to organismal adaptation, the genetic systems that affect them might be more flexible than those governing more complex traits such as body organization, appendage formation, and other, more slowly evolving characters. The available evidence suggests, however, that the diversification of other traits that are governed by highly pleiotropic and well-conserved proteins can also be accounted for by regulatory sequence evolution.
For example, shifts in the rostrocaudal boundaries of Hox gene expression are associated with large-scale differences in axial patterning in vertebrates, arthropods, and annelids [24]. In one case, the Hoxc8 gene of the chicken and mouse, differences in the function of one cis-regulatory element have been demonstrated to govern differences in gene expression boundaries in the neural tube and paraxial mesoderm [34].
While such differences in axial morphology are thought to evolve slowly and relatively infrequently, some features of the vertebrate skeleton, such as the pelvic skeleton of stickleback fish, evolve rapidly [38] and repeatedly [33]. Reduction of the pelvic fin, the homolog of the tetrapod hindlimb, is due to changes at the Pitx1 locus [33]. The Pitx1 protein is a pleiotropic transcription factor that affects the development of multiple tissues in fish and mice, including the hindlimb. Pelvic-reduced sticklebacks have lost Pitx1 expression in pelvic fin precursors, but possess a perfectly intact Pitx1 coding region with no sequence changes relative to populations with fully formed pelvic structures. The only explanation consistent with these observations is that regulatory mutations in a cis-element governing expression in the pelvic fin precursors has selectively abolished Pitx1 expression in this one part of the developing animal, while gene expression elsewhere is not affected (Figure 2B).
The crucial insight from the evolution of Pitx1, yellow, and Hoxc8 is that regulatory mutations provide a mechanism for change in one trait while preserving the role of pleiotropic genes in other processes. This is perhaps the most important, most fundamental insight from evolutionary developmental biology. While functional mutations in a coding region are usually poorly tolerated and eliminated by purifying selection, even complete loss-of-function mutations in regulatory elements are possible because the compartmentation created by the modularity of cis-regulatory elements limits the effects of mutations to individual body parts.
Does this mean that coding sequences cannot contribute to morphological evolution? Not at all. There are several clear examples of functional sequence changes in proteins that affect form, and I will highlight them next. The key questions to keep in mind are, how often and under what circumstances do coding sequences of regulatory molecules functionally evolve?
Coding Sequences and the Evolution of Anatomy
The body plans of arthropods and tetrapods have evolved around the use of a fairly stable complement of Hox genes in each phylum [24,39]. The stability of Hox gene number, and the conservation of Hox ortholog sequences and function, led to the initial impression that Hox proteins have not significantly diverged in function. However, it is now understood that several arthropod Hox proteins have changed in ways that are associated with shifts in form or developmental mechanisms, including the Hox3, Fushi tarazu, Ultrabithorax (Ubx), and Antennapedia [40] proteins. In the case of Hox3 and Fushi tarazu, Hox-type function has been lost in particular lineages while new functions have been gained. The Fushi tarazu protein of certain insects lost sequence motifs involved in Hox functions, and gained a motif for a new activity involved in segmentation [41,42]. Similarly, the Hox3 protein lost Hox function in insects and gained a novel dorsoventral axis patterning function. It subsequently underwent a duplication that produced two divergent genes involved in early patterning of the two body axes in one clade of flies [43–45].
In the Ubx protein, functional motifs evolved while the protein retained Hox function. Comparative and functional studies have shown that the carboxy terminus of the Ubx protein was extended in the crustacean lineage and serves as an activity-modulating domain [46]. In the insect lineage, this domain was replaced by a short glutamine/alanine-rich motif that has been well preserved throughout the course of more than 300 million years of insect evolution [47].
These arthropod Hox proteins demonstrate that some of the most conserved proteins can, under certain circumstances, evolve new and different activities. In these examples, selection against coding changes might have been relaxed because of functional redundancy among Hox paralogs. However, these events are, in the long span of the history of these lineages, rare relative to the extensive diversification of body forms. It must also be stressed that both ftz and Hox3 (and its derivatives zen and bcd) acquired entirely novel regulatory elements that governed their expression in new domains and patterns. Furthermore, Ubx regulation has been extensively diversified among arthropods [24], including within the insects [48–50]. Thus, even in the infrequent instances of overt coding sequence evolution in regulatory proteins, regulatory sequence evolution is a critical component of functional evolution, and further diversification of gene function.
Are there more common and rapid means of evolving morphological diversity via coding mutations? Definitely. One prominent example is the melanocortin-1 receptor (MC1R), which modulates the quantity and type of melanin synthesis in melanocytes. Mutations in the MC1R gene are associated with scale, fur, or plumage color variation and divergence in a wide range of species [51]. The ecological significance of alternative phenotypes suggests that the MC1R gene has evolved under natural and sexual selection. The clear-cut case of MC1R evolution raises the question, why is coding sequence evolution so prevalent in the diversification of vertebrate pigmentation, while the evolution of gene regulation plays a central role in flower and fruit fly pigmentation?
There may be particular properties of MC1R that have enabled it to play this starring role. MC1R is a member of a family of five related receptors and is the only member involved in pigment synthesis regulation [52]. Thus, the structural and regulatory diversification of this receptor family (that is, the evolution of MC1R expression in melanocytes) has produced a protein that has a much greater degree of evolutionary freedom than more pleiotropic receptors. It should be noted that MC1R coding mutations result in body-wide effects on pigmentation, and do not create or alter spots, stripes, or other patterns. The evolution of spatial patterns of pigmentation in vertebrates is still likely to involve regulatory evolution in the expression of pigmentation proteins, or regulators of receptor activity [53], via mechanisms similar to those underlying the evolution of insect color patterns.
The widespread involvement of MC1R coding variation in the visible diversity of vertebrates may then be a relatively special case, enabled by the dedication of MC1R to pigmentation and its minimal pleiotropy. It would be expected that other, more pleiotropic proteins would be constrained in their sequence variation and, hence, their contribution to morphological variation. However, it has recently been shown that morphological variation in dog breeds is associated with variation in the length of repeated amino acid sequences in the coding regions of a variety of developmentally important transcription factors [54]. These repeats are encoded by microsatellite sequences that expand or contract at very high rates, and spontaneous or induced mutations of these sites affect visible traits. The extraordinary variation in repeat lengths, and their potential effects on morphology, raises the possibility that these repeats are a source of variation in natural populations. However, this variation may have accompanying deleterious, pleiotropic effects that, while manageable under domestication, would limit its contribution to evolution under natural selection.
Gene Duplication and the Evolution of Anatomy
The history of Hox genes and the MC1R gene reflects that one condition contributing to the potential evolution of coding sequences is the generation of new genes by duplication. Ever since Ohno [14], and indeed well before [55], there has been widespread belief and expectation that gene duplication has been a major driving force in evolution. Empirical evidence suggests, however, that while gene duplication has contributed to the evolution of form, the frequency of duplication events is not at all sufficient to account for the continuous diversification of lineages. This conclusion is based primarily upon two sets of observations.
First, the estimated rate of gene duplication is about once per gene per 100 million years [56]. This figure suggests that gene duplication can contribute to genome evolution over longer spans of evolutionary time (for example, greater than 50 million years), but this rate is not sufficient to account for variation in populations (for example, quantitative trait differences) or for divergence among related species such as the 300,000 known species of beetles, or 10,000 species of birds.
Second, the relative infrequency of gene duplication is documented by the actual histories of key developmental regulatory gene families. For example, while it is very clear that during the early evolution of animals, there was an expansion in the number of Hox genes, and that during the early evolution of the vertebrates, there was an expansion in the number of Hox gene clusters, the number and diversity of Hox genes in highly diversified phyla, such as the arthropods and tetrapods, appears to have remained fairly stable for very long periods (perhaps approximately 500 million years). Other gene families, such as the Wnt family of signaling ligands, also exhibit deep ancestral complexity. Of 12 Wnt subfamilies known in vertebrates, 11 have been identified in a cnidarian [57]. Such deep ancestral complexity is much greater than would be expected under the hypothesis that diversity evolves primarily through the evolution of new genes [39,58]. Similarly, despite widespread speculation that the human genome would contain many more genes than other species, it does not, and the great majority of human genes have syntenic orthologs in the mouse [6].
Furthermore, the contribution of gene duplication to the evolution of form may be governed primarily by the divergence of the regulation of newly duplicated genes, rather than novel functions acquired by coding mutations. Both theoretical considerations and empirical data have suggested that the partitioning of the progenitor gene's functions may occur most often through regulatory mutations, or the partitioning of regulatory sequences in the original duplication event [59].
The Relative Contribution of Regulatory and Coding Sequences to Anatomical Evolution
The examples I have described demonstrate that both regulatory sequences and coding regions of the genome can and do contribute to the evolution of form. The more subjective issue is whether, from the small sample of case studies mentioned here and in the literature, one can make (and defend) statements about the relative contribution of regulatory and coding sequence evolution to the evolution of anatomy. We are, after all, in much better position now to do so than King and Wilson were 30 years ago.
While the agnostic, “wait and see” position would appear safer, that would not at all be in keeping with the bold spirit of the pioneers who first wrestled with the question. Moreover, I argue that a trend is evident, and that that trend should, of course, inform ongoing and future work. Based upon (i) empirical studies of the evolution of traits and of gene regulation in development, (ii) the rate of gene duplication and the specific histories of important developmental gene families, (iii) the fact that regulatory proteins are the most slowly evolving of all classes of proteins, and (iv) theoretical considerations concerning the pleiotropy of mutations, I argue that there is adequate basis to conclude that the evolution of anatomy occurs primarily through changes in regulatory sequences.
This conclusion comes as no surprise, given the hypotheses of King and Wilson and others framed decades ago. Indeed, most aficionados of evolutionary developmental biology would find no news here. However, I am not convinced that what we have learned about the evolution of form is being adequately considered in comparative genomics and population genetics, where the potential role of regulatory sequence evolution appears to be a secondary consideration, or ignored altogether. This neglect has fundamental bearing on the issue that first drew King and Wilson's interest—the origins of differences between chimps and humans.
Chimps and Humans Redux
The morphological differences between modern humans, human ancestors, and the great apes are the product of evolutionary changes in development. I have argued elsewhere [60] that the evolution of complex traits such as brain size, craniofacial morphology, cortical speech and language areas, hand and digit form, dentition, and body skeletal morphology must have a highly polygenic and largely regulatory basis. The great and difficult challenge, with the genome sequences of humans, chimps, and other mammals now available, is to map changes in genes to changes in traits. Many approaches are being taken, and a few intriguing associations of candidate genes and the evolution of particular traits have been discovered, such as the FOXP2 gene and the evolution of speech [61], and the MYH16 muscle-specific myosin pseudogene and the evolutionary reduction of the masticatory apparatus [62]. My concern here is not whether these specific associations did or did not play a role in human evolution; rather, my concern is the exclusive focus, by choice or by necessity, on the evolution of coding sequences in these and more genome-wide population genetic surveys of chimp–human differences [63].
There exists some disconnect between what studies in model species have underscored—the ability or sufficiency of regulatory sequences to account for the evolution of physical traits—and which models of evolution are implicitly or explicitly being tested when only coding sequence divergence is considered. Two stories concerning the FOXP2 gene illustrate the dramatically different conclusions one might draw, depending upon the methodologies and assumptions applied.
The human FOXP2 gene encodes a transcription factor, and mutations at the locus were discovered to be associated with a speech and language disorder [64]. The human FOXP2 protein differs from the gorilla and chimp protein at just two residues, raising the possibility that the two replacements that occurred in the human lineage might be significant to the evolution of speech and language. Furthermore, population genetic analysis indicates that the FOXP2 locus has undergone a selective sweep within the last 200,000 years of human evolution [61]. While it would certainly be convenient if the two changes in the FOXP2 protein were functional, the additional hypothesis must be considered that functional regulatory changes might have occurred at the FOXP2 locus. In weighing alternative hypotheses of FOXP2 or any gene's potential involvement in the evolution of form (or neural circuitry), we should ask the following questions. (i) Is the gene product used in multiple tissues? (ii) Are mutations in the coding sequence known or likely to be pleiotropic? (iii) Does the locus contain multiple cis-regulatory elements?
If the answers are yes to all of these questions, then regulatory sequence evolution is the more likely mode of evolution than coding sequence evolution. For FOXP2, this appears to be the case. FOXP2 is expressed at multiple sites, not just in the brain, but in the lungs, heart, and gut as well [64,65]. Patients with the FOXP2 mutation do have multiple neural deficits [66]. And, because FOXP2 is expressed in different organs and different regions of the brain, it is certain to possess multiple regulatory elements. Furthermore, it is an enormous, complex locus, spanning some 267 kb. Based upon a simple average base pair divergence of 1.2%, there should be over 2,000 nucleotide differences between chimps and humans in this span. Because there is much more potential for functional divergence in non-coding sequences, there is no specific reason to favor coding sequence divergence over regulatory sequence divergence at FOXP2.
The discovery of FOXP2 and its association with human speech has inspired consideration of the potential role of FOXP2 in the evolution of vocalization in other animals, and here is where strikingly different conclusions were reached depending upon the hypothesis tested and the methodology used. Song learning has evolved in three orders of birds. There are some behavioral and neural similarities between bird song and human speech in terms of their being learned at critical periods and the involvement of auditory and motor centers and specialized brain centers. A standard comparative analysis of the FOXP2 coding sequences of humans and song-learning and non-learning birds did not reveal any amino acid substitutions that were shared between song-learning birds and humans, nor any fixed differences between song-learning and non-learning birds. The study concluded there was “no evidence for its [FOXP2] role during the evolution of vocal learning in nonhuman animals” [67].
In great contrast, when FOXP2 mRNA and protein expression in the developing and adult brains of a variety of song-learners and non-learners were examined, a striking increase in FOXP2 expression was observed in Area X, a center necessary for vocal learning that is absent from non-learners [68] (Figure 3A–3C). This increase occurs in zebra finches over the developmental period when vocal learning occurs. Furthermore, in adult canaries, seasonal changes in FOXP2 expression were observed in Area X, associated with changes in the stability of the bird's song (Figure 3D–3F). Thus, remarkable changes in the regulation of FOXP2, but not the protein sequence, are correlated with the development and evolution of vocal learning in birds. These changes could arise through the evolution of FOXP2 cis-regulatory sequences, or of the regulatory or coding sequences of transcription factors that control FOXP2.
Figure 3 The Regulatory Evolution of FOXP2 and the Origins of Vocal Learning
(A–F) The patterns of FOXP2 expression in sections of bird brains are depicted The green area is the striatum. FOXP2 is upregulated in the vocalization center known as Area X (pink spots) in vocal-learning species such as the zebra finch (A) and black-capped chickadee (B) but not in non-learning species such as the ringdove (C).
(D–F) In the canary, FOXP2 expression in Area X varies over seasons; elevated expression is associated with periods during which the song is plastic (pink spot).
(Figure based on [68].)
The contrast between the negative conclusions drawn from the analysis of coding sequences and the fascinating correlation revealed by the comparative study of gene regulation in vivo highlights the general inadequacies of, and potential error in, the exclusive analysis of coding regions when considering the evolution of anatomy. But that inadequacy applies more broadly than just to the evolution of form. While standard population genetic tests have been used to search human protein sequences for statistical evidence of positive selection [63,69], several examples of positive selection on cis-regulatory sequences of physiological genes are documented [70–72]. This includes the very clear case of the erythroid-specific loss of expression of the Duffy antigen chemokine receptor in populations resistant to Plasmodium vivax malaria [73]. This loss is due to a regulatory mutation that affects an erythroid cis-regulatory sequence but has no effect on receptor expression elsewhere in the body [74].
Any statements or claims, then, about the genetic changes that “make us human” must be weighed critically in light of the power and limitations of the methodology employed, and the scope of the hypotheses being tested. While it is understandable that some biologists have reached for the “low-hanging fruit” of coding sequence changes, the task of unraveling the regulatory puzzle is yet to come.
Conclusion
The hypothesis of regulatory evolution put forward by King and Wilson 30 years ago was founded entirely on negative data, that is, the apparent insufficiency of coding sequence divergence to account for gross organismal differences. It has required several decades to obtain evidence that regulatory sequences are so often the basis for the evolution of form that, when considering the evolution of anatomy (including neural circuitry), regulatory sequence evolution should be the primary hypothesis considered. The analysis of regulatory sequence evolution poses particular challenges in that it is impossible to distinguish meaningless from functional changes by mere inspection. But, in nonhuman models where extensive experimental tools are available, there is cause for optimism that the contribution of regulatory sequences to evolution will be increasingly well understood in the near term. In order to approach the origins of human traits, much greater emphasis has to be placed on comparative studies of gene expression, regulation, and development in apes and other primates. This is precisely the requirement forecast by King and Wilson 30 years ago [1], only now we have the means to meet it.
I thank M.-C. King for correspondence regarding her 1975 paper with A. C. Wilson, L. Olds for the artwork, and A. Rokas, B. Williams, C. Hittinger, B. Hersh, P. Carroll, and S. Paddock for helpful comments. Work in my laboratory is supported by the Howard Hughes Medical Institute.
Citation: Carroll SB (2005) Evolution at two levels: On genes and form. PLoS Biol 3(7): e245.
Sean B. Carroll is at the Howard Hughes Medical Institute and Laboratory of Molecular Biology, University of Wisconsin, Madison, Wisconsin, United States of America. E-mail: [email protected]
This article is based on the Allan Wilson Memorial Lectures given at the University of California at Berkeley in October 2004.
Abbreviations
MC1Rmelanocortin-1 receptor
UbxUltrabithorax
==== Refs
References
King MC Wilson AC Evolution at two levels in humans and chimpanzees Science 1975 188 107 116 1090005
Yokoyama S Molecular evolution of color vision in vertebrates Gene 2002 300 69 78 12468088
Jessen TH Weber RE Fermi G Tame J Braunitzer G Adaptation of bird hemoglobins to high altitudes: Demonstration of molecular mechanism by protein engineering Proc Natl Acad Sci U S A 1991 88 6519 6522 1862080
Zhang J Zhang YP Rosenberg HF Adaptive evolution of a duplicated pancreatic ribonuclease gene in a leaf-eating monkey Nat Genet 2002 30 411 415 11925567
Hughes AL Natural selection and the diversification of vertebrae immune effectors Immunol Rev 2002 190 161 168 12493013
Waterston RH Lindblad-Toh K Birney E Rogers J Abril JF Initial sequencing and comparative analysis of the mouse genome Nature 2002 420 560 562
Crick FHC On protein synthesis Symp Soc Exp Biol 1958 12 138 163 13580867
Zuckerkandl E Pauling L Bryson V Vogel HJ> Evolutionary divergence and convergence in proteins Evolving genes and proteins 1965 New York Academic Press 97 166
Kimura M The rate of molecular evolution considered from the standpoint of population genetics Proc Natl Acad Sci U S A 1969 63 1182 1188
Jacob F Monod J Genetic regulatory mechanisms in the synthesis of protein J Mol Biol 1961 8 318 356
Zuckerkandl E Controller-gene diseases: The operon model as applied to beta-thalassemia, familial fetal hemoglobinemia and the normal switch from the production of fetal hemoglobin to that of adult hemoglobin J Mol Biol 1964 8 128 147 14149956
Britten RJ Davidson EH Gene regulation for higher cells: A theory Science 1969 165 349 357 5789433
Britten RJ Davidson EH Repetive and non-repetitive KNA sequences and a speculation on the origins of evolutionary novelty Quart Rev Biol 1971 46 111 138 5160087
Ohno S Evolution by gene duplication 1970 New York Springer-Verlag 160
Ohno S An argument for the genetic simplicity of man and other animals J Hum Evol 1972 1 651 662
Prager EM Wilson AC Slow evolutionary loss of the potential for intersspecific hybridization in birds Proc Natl Acad Sci U S A 1975 72 200 204 1054495
Wilson AC Maxson LR Sarich VM Two types of molecular evolution. Evidence from studies of interspecific hybridization Proc Natl Acad Sci U S A 1974 71 2843 2847 4212492
Wilson AC Sarich VM Maxson LR The importance of gene rearrangement in evolution Proc Natl Acad Sci U S A 1974 71 3028 3030 4528784
Jacob F Evolution and tinkering Science 1977 196 1161 1166 860134
Stern DL Perspective: Evolutionary developmental biology and the problem of variation Evolution 2000 54 1079 1091 11005278
Kirschner M Gerhart J Evolvability Proc Natl Acad Sci U S A 1998 95 8420 8427 9671692
Gerhart J Kirschner M Cells, embryos, and evolution 1997 Malden Blackwell Science 642
Davidson EH Genomic regulatory systems: Development and evolution 2001 San Diego Academic Press 261
Carroll SB Grenier JK Weatherbee SD From DNA to diversity: Molecular genetics and the evolution of animal design, 2nd ed 2005 Malden Blackwell Scientific 258
Stern DL A role of Ultrabithorax in morphological differences between Drosophila species Nature 1998 396 463 466 9853753
Sucena E Stern DL Divergence of larval morphology between Drosophila sechelia and its sibling species caused by cis-regulatory evolution of ovo/shaven-baby Proc Natl Acad Sci U S A 2000 97 4530 4534 10781057
Sucena E Delon I Jones I Payre F Stern DL Regulatory evolution of shavenbaby/ovo underlies multiple cases of morphological parallelism Nature 2003 424 935 938 12931187
Genissel A Pastinen T Dowell A Mackay TF Long AD No evidence for an association between common nonsynonymous polymorphisms in delta and bristle number varation in natural and laboratory populations of Drosophila melanogaster
Genetics 2004 166 291 306 15020426
Kopp A Duncan I Godt D Carroll SB Genetic control and evolution of sexually dimorphic characters in Drosophila
Nature 2000 408 533 559 11117728
Durbin ML Lundy KE Morrell PL Torres-Martinez CL Clegg MT Genes that determine flower color: The role of regulatory changes in the evolution of phenotypic adaptations Mol Phylogenet Evol 2003 29 507 518 14615189
Yoon HS Baum DA Transgenic study of parallelism in plant morphological evolution Proc Natl Acad Sci U S A 2004 101 6524 6529 15096600
Wang RL Stec A Hey J Lukens L Doebley J The limits of selection during maize domestication Nature 1999 398 236 239 10094045
Shapiro MD Marks ME Peichel CL Blackman BK Nereng KS Genetic and developmental basis of evolutionary pelvic reduction in threespine sticklebacks Nature 2004 428 717 723 15085123
Belting HG Shashikant CS Ruddle FH Modification of expression and cis-regulation of Hoxc8 in the evolution of diverged axial morphology Proc Natl Acad Sci U S A 1998 95 2355 2360 9482889
Gompel N Prud'homme B Wittkopp PJ Kassner VA Carroll SB Chance caught on the wing: Cis-regulatory evolution and the origin of pigment patterns in Drosophila
Nature 2005 433 481 487 15690032
Wittkopp PJ Vaccaro K Carroll SB Evolution of yellow gene regulation and pigmentation in Drosophila
Curr Biol 2002 12 1547 1556 12372246
Wittkopp PJ Kopp A Carroll SB Evolution in black and white: Genetic control of pigment patterns in Drosophila
Trends Genet 2003 19 495 504 12957543
Bell MA Baumgartner JV Olson EC Patterns of temporal change in single morphological characters of a Miocene stickleback fish Paleobiology 1985 11 258 271
Carroll SB Homeotic genes and the evolution of arthropods and chordates Nature 1995 376 479 485 7637779
Shiga Y Yasumoto R Yamagata H Hayashi S Evolving role of Antennapedia protein in arthropod limb patterning Development 2002 129 3555 3561 12117806
Löhr U Yussa M Pick L
Drosophila fushi tarazu : A gene on the border of homeotic function Curr Biol 2001 11 1403 1412 11566098
Alonso C Maxton-Kuechenmeister J Akam M Evolution of Ftz protein function in insects Curr Biol 2001 11 1473 1478 11566109
Telford MJ Thomas R Of mites and zen: Expression studies in a chelicerate arthropod confirm zen is a divergent Hox gene Dev Genes Evol 1998 208 591 594 9811978
Stauber M Prell A Schmidt-Ott U A single Hox3 gene with composite bicoid and zerknüllt expression characteristics in non-Cyclorrhaphan flies Proc Natl Acad Sci U S A 2002 99 274 279 11773616
Damen WGM Tautz D A Hox class 3 orthologue from the spider Cupiennius salei is expressed in a Hox-gene-like fashion Dev Genes Evol 1998 208 586 590 9811977
Ronshaugen M McGinnis N McGinnis W Hox protein mutation and macroevolution of the insect body plan Nature 2002 415 914 917 11859370
Galant R Carroll SB Evolution of a novel transcriptional repression domain in the Ultrabithorax Hox protein in insects Nature 2002 415 848 849 11859349
Mahfooz NS Li H Popadic A Differential expression patterns of the hox gene are associated with differential growth of insect hind legs Proc Natl Acad Sci U S A 2004 101 4877 4882 15041751
Palopoli MF Patel NH Evolution of the interaction between Hox genes and a downstream target Curr Biol 1998 8 587 590 9601643
Warren R Nagy L Selegue J Gates J Carroll S Evolution of homeotic gene regulation and function in flies and butterflies Nature 1994 372 458 461 7840822
Majerus MEN Mundy NI Mammalian melanism: Natural selection in black and white Trends Genet 2003 19 585 588 14585605
Catania A Gatti S Lipton JM Targeting melanocortin receptors as a novel strategy to control inflammation Pharmacol Rev 2004 56 1 29 15001661
Millar SF Miller MW Stevens ME Barsh GS Expression and transgeneic studies of the mouse agouti gene provide insight into the mechanisms by which mammalian coat color patterns are generated Development 1995 121 3223 3232 7588057
Fondon JW Garner HR Molecular origins of rapid and continuous morphological evolution Proc Natl Acad Sci U S A 2004 101 18058 18063 15596718
Taylor JS Raes J Duplication and divergence: The evolution of new genes and old ideas Annu Rev Genet 2004 38 615 643 15568988
Lynch M Conery JS The evolutionary fate and consequences of duplicate genes Science 2000 290 1151 1155 11073452
Kusserow A Pang K Sturm C Hrouda M Lentfer B Unexpected complexity of the Wnt gene family in a sea anemone Nature 2005 433 156 160 15650739
Ball EE Hayward DC Saint R Miller DJ A simple plan—Cnidarians and the origins of developmental mechanisms Nat Rev Genet 2004 5 567 577 15266339
Lynch M Katju V The altered evolutionary trajectories of gene duplicates Trends Genet 2004 20 544 549 15475113
Carroll SB Genetics and the making of Homo sapiens
Nature 2003 422 849 857 12712196
Enard W Przeworski M Fisher SE Lai CSL Kitano WV Molecular evolution of FOXP2, a gene involved in speech and language Nature 2002 428 869 872
Stedman HH Kozyak BW Nelson A Thesier DM Su LT Myosin gene mutation correlates with anatomical changes in the human lineage Nature 2004 428 415 418 15042088
Dorus S Vallender EJ Evans PD Anderson JR Gilbert SL Accelerated evolution of nervous system genes in the origin of Homo sapiens
Cell 2004 119 1027 1040 15620360
Lai CSL Fisher SE Hurst JA Vargha-Khadem F Monaco AP A forkhead-domain gene is mutated in a severe speech and language disorder Nature 2001 413 519 523 11586359
Shu W Yang H Zhang L Lu MM Morrisey EE Characterization of a new subfamily of winged-helix/forkedhead (Fox ) genes that are expressed in the lung and act as transcriptional repressors J Biol Chem 2001 276 27488 27497 11358962
Vargha-Khadem F Gadian DG Copp A Mishkin M
FOXP2 and the neuroanatomy of speech and language Nat Rev Neurosci 2005 6 131 138 15685218
Webb DM Zhang J
FoxP2 in song-learning birds and vocal-learning mammals J Hered 2005 96 1 5 15618310
Haesler S Wada K Nshdejan A Morrisey EE Lints T
FoxP2 expression in avian vocal learners and non-learners J Neurosci 2004 24 3164 3175 15056696
Clark AG Glanoski S Nielsen R Thomas PD Kejariwal A Inferring nonneutral evolution from human-chimp-mouse orthologous gene trios Science 2003 302 1876 1877
Rockman MV Hahn MW Soranzo N Goldstein DB Wray GA Positive selection on a human-specific transcription factor binding site regulating IL4 expression Curr Biol 2003 13 2118 2123 14654003
Rockman MV Hahn MW Soranzo N Loisel DA Goldstein DB Positive selection on MMP3 regulation has shaped heart disease risk Curr Biol 2004 14 1531 1539 15341739
Hahn MW Rockman MV Soranzo N Goldstein DB Wray GA Population genetic and phylogenetic evidence for positive selection on regulatory mutations at the factor VII locus in humans Genetics 2004 167 867 877 15238535
Livingstone FB The Duffy blood groups, vivax malaria, and malaria selection in human populations: A review Hum Biol 1984 56 413 425 6386656
Tournamille C Colin Y Cartron JP Le Van Kim C Disruption of a
GATA motif in the Duffy gene promoter abolishes erythroid gene expression in Duffy-negative individuals
Nat Genet 1995 10 224 228 7663520
Locton S Gaut BS Plant conserved non-coding sequences and paralogue evolution Trends Genet 2005 21 60 65 15680516
| 16000021 | PMC1174822 | CC BY | 2021-01-05 08:21:25 | no | PLoS Biol. 2005 Jul 12; 3(7):e245 | utf-8 | PLoS Biol | 2,005 | 10.1371/journal.pbio.0030245 | oa_comm |
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1600002210.1371/journal.pbio.0030246Correspondence and Other CommunicationsNeuroscienceIn VitroSchool Students as Drosophila Experimenters CorrespondenceSiyad Faiza
1
Griffiths Jodianne
1
Janjua Faira
2
Jackson Elizabeth
3
Rodrigues Ian
3
Kerr Fiona
3
Mackay Daniel
3
Lovestone Simon [email protected]
3
1John Ruskin College, Croydon, United Kingdom2Burntwood School, London, United Kingdom3King's College London, London, United Kingdom7 2005 12 7 2005 12 7 2005 3 7 e246Copyright: © 2005 Siyad et al.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.Inspired by an undergraduate science project in Drosophila genetics published in PLoS Biology, high school students investigate a model of Alzheimer's Disease.
==== Body
Students can be a valuable resource for the scientific community, as demonstrated by Chen et al. [1]. However, it is not just undergraduate students who can contribute but secondary school students also. Previously some of us have published a Drosophila model of tauopathy where we have shown that overexpression of tau results in disruption of axonal transport and an intact phenotype in both larvae and adults [2]. This phenotype is tau phosphorylation dependant and attenuated by inhibitors of Glycogen synthase kinase-3 (GSK-3). We have thought for some time that this model is ideally suited for testing compounds that might alter axonal transport or affect signalling though GSK-3 and other kinases to tau phosphorylation. However, we had other priorities in the laboratory and this work was not pursued.
The article by Chen et al. [1] coincided with planning for an open day as part of the British Association for Advancement of Science's Science Week in March 2005 and funded by the Medical Research Council as part of a Public Engagement with Science activity. As a consequence we established a collaboration between three local schools and the Institute of Psychiatry at King's College London, and students worked alongside research workers to test a series of compounds on a larval phenotype and the neuromuscular junction (NMJ) anatomy of wild-type flies (Oregon R) and transgenic flies (human 3R tau expressed in motor neurons).
We have replicated the previous observation that tau expression significantly alters motor-dependant phenotypes and in addition that the NMJ is severely disrupted. We have tested a series of compounds that alter signalling to GSK-3, and with one exception these all alter the phenotype in the predicted direction—where inhibition of GSK-3 improves the phenotype. We also tested curcumin, which has previously been shown to alter Aβ aggregation and amyloid-dependant processes [3]. We found no effect of curcumin in our tau-dependant model. Most interestingly, we found a large and very highly significant worsening of the phenotype with taxol-treated larvae, in contrast to previous studies [4].
We, the students, believe that it was worthwhile doing the experiment as it could lead to greater understanding of Alzheimer disease, and we have found that the experience has helped some of us in our commitment to pursue a career in science and medicine.
We, the scientists, have completed a pilot experiment involving multiple data points processed rapidly and using a widely available resource—students. Twelve pairs of hands allowed us to gather in a little over a day data that would have otherwise required some weeks of research time. We are now in the process of replicating and validating the student data, concentrating on those compounds that appear most effective from the student-led experiment. We are grateful to Chen et al. [1] for directing us to this valuable human resource.
Other students taking part in the study included Nadia Adesoji, Faiza Ahmad, Saira Siddiqui, Alice Cottington, Vanessa Christophers, and from Graveney School in London Alasdair MacDonald, Ashley Lawrence, Simon Perrineau, and Sidra Tulmuntaha.
Citation: Siyad F, Griffiths J, Janjua F, Jackson E, Rodrigues I, et al. (2005) School students as Drosophila experimenters. PLoS Biol 3(7): e246.
==== Refs
References
Chen J Call GB Beyer E Bui C Cespedes A Discovery-based science education: Functional genomic dissection in Drosophila by undergraduate researchers PLoS Biol 2005 3 e59 15719063
Mudher A Shepherd D Newman TA Mildren P Jukes JP GSK-3beta inhibition reverses axonal transport defects and behavioural phenotypes in Drosophila
Mol Psychiatry 2004 9 522 530 14993907
Ono K Hasegawa K Naiki H Yamada M Curcumin has potent anti-amyloidogenic effects for Alzheimer's beta-amyloid fibrils in vitro J Neurosci Res 2004 75 742 750 14994335
Zhang B Maiti A Shively S Lakhani F McDonald-Jones G Microtubule-binding drugs offset tau sequestration by stabilizing microtubules and reversing fast axonal transport deficits in a tauopathy model Proc Natl Acad Sci U S A 2005 102 227 231 15615853
| 16000022 | PMC1174823 | CC BY | 2021-01-05 08:21:24 | no | PLoS Biol. 2005 Jul 12; 3(7):e246 | utf-8 | PLoS Biol | 2,005 | 10.1371/journal.pbio.0030246 | oa_comm |
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 10.1371/journal.pbio.0030249Book Reviews/Science in the MediaNeurology/NeurosurgeryHomo (Human)Vision in Film Book Review/Science in the MediaBagchi Sanjit 7 2005 12 7 2005 12 7 2005 3 7 e249Movie Reviewed Bhansali SL, director (2005) Black [film]. Applause Entertainment Copyright: © 2005 Sanjit Bagchi.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.A new film set in India re-creates the story of Helen Keller.
==== Body
India produces the largest number of movies in the world, and the country is also home to one of the largest number of people born blind [1]. Yet it is only now that the 90-year-old Hindi film industry, based in Mumbai, has made a movie based on the life of Helen Keller, arguably the most inspirational visually challenged person in history.
Because films are supposed to be a means of transient escapism for millions of poor Indians, few movie producers dare to deviate from unrealistic and hackneyed formulae (either fairy-tale romances or action-packed good-versus-evil sagas; in both cases, at least five song-and-dance sequences complete the storyline). With a few exceptions (e.g., Koshish and Sparsh), rarely have Indian movie-makers dared to foray into more challenging territory.
However, in the past few years, a handful of young directors have begun to venture into important medical territories, such as AIDS (Phir Milenge) and autism (Koi Mil Gaya). Sanjay Leela Bhansali has been among the leaders of this nascent movement. After exploring the world of a deaf-mute couple in his debut film five years ago, Bhansali now depicts the indomitable spirit of a deaf, blind, and mute girl in Black. More audacious than before, he has added a new twist to the scaffold of the true Helen Keller story, in the form of Alzheimer's disease—a degenerative disorder affecting at least 3 million people in India, yet grossly unrecognised and under-diagnosed due to lack of awareness [2]. The film, poised to be a smash hit, may serve not only to entertain but also to educate.
Black is loosely based on The Miracle Worker, the Oscar-winning 1962 Hollywood movie about the life of Helen Keller and her teacher, Anne Sullivan. The real Helen, born in 1880 in Tuscumbia, a small rural town in northwest Alabama (United States), lost her sight and hearing to a mysterious disease (probably meningitis or scarlet fever) when she was 19 months old. The otherwise intelligent girl grew up to be a difficult child, only to be sobered by her teacher.
Sullivan began her lessons by teaching Helen to finger spell. Painstakingly, she would spell each word on the girl's wrists as D-O-L-L, C-A-K-E, W-A-T-E-R, and so on, while Helen was made to feel the objects. In the beginning, she could repeat the finger movements on her own but could not understand what they meant. The initial progress was extremely slow because Helen's parents were opposed to the strict discipline Sullivan advocated. Moreover, Sullivan found it extremely difficult to control Helen's temper tantrums and unruly manners.
To improve Helen's behaviour as well as to teach her to communicate, Sullivan moved with her into a small cottage close to the Kellers' home, but away from her parents' compassion. She did not hesitate to punish when necessary, even by refusing to “talk” through finger spelling when the girl threw tantrums. Within a few weeks, Helen's behaviour improved and a bond grew between the two. Eventually, Sullivan taught Helen to not only communicate through sign language but also read Braille books and lip-read through touch. Ultimately, Helen earned a college degree, wrote autobiographical books, including the classic The Story of My Life, and toured the world with Anne Sullivan.
In Bhansali's interpretation, Tuscumbia is replaced by Simla, the picturesque hill station in Northern India. Michelle McNally is the deaf-blind girl born to a wealthy Anglo-Indian household. Eight-year-old Ayesha Kapoor brilliantly portrays the frustrations of a child incapable of expressing herself. Unable to see, hear, or talk, Michelle throws violent tantrums, and her parents decide to put her in an asylum. This is when an eccentric teacher, Debraj (played by the Indian idol Amitabh Bachchan), enters her life, taking up the seemingly impossible challenge to illuminate Michelle's “black” existence. Bhansali not only reverses the gender of Helen Keller's real-life teacher but also portrays him as an aging alcoholic and an emotional wreck who has survived a traumatic childhood.
The subplot works well as Debraj, the protagonist, speaks against the ills of institutionalized rehabilitation of the deaf-blind. From the credit lines, it is evident that Bhansali's team has worked closely with members of the Helen Keller Institute of Mumbai for several years. The film drives home the message that even the deaf-blind can lead an independent life with proper care and support. As Debraj says, “Michelle, you've to stand up, or else they'll put a bell on your neck and you'll live like an animal.”
Debraj's strategy for rehabilitation focuses on bringing the experiences of the world to the fingertips of the deaf-blind person. It emphasizes the use of all possible non-linguistic modes of communication, such as gestures, facial expressions, and body movements. The film also portrays rehabilitation of the deaf-blind and mute girl as hard-won and awe-inspiring. It demonstrates that even without modern intervention facilities—such as computerized Braille, talking books, or text readers—that have made rehabilitation of the deaf-blind more systematic and comprehensive since the days of Helen Keller, poor Indian families can offer rich lives to children with such physical and sensory challenges. It encourages parents to create family-centred, home-based programmes, particularly for children who cannot afford to join day-care centres due to lack of suitable transportation or sufficient socio-economic status.
Even more commendable is the film's attempt to introduce Indian viewers to the vagaries of Alzheimer's disease. The audience can readily identify with an ever-enthusiastic and boisterous Debraj, who begins to suffer the abrupt memory lapses that are early signs of the disease. His eyes lose their penetrative look, and he forgets simple words and their connotations. The actor paints the pain of the mind-crippling and devastating illness with great sensitivity.
The film ends with a surrealistic note when Michelle decides to apply the same multi-sensory method on her teacher to lift him from the world of forgetfulness. However, in practical terms, it seems unlikely that rehabilitation originally meant for the deaf-blind and mute will succeed in a patient afflicted with Alzheimer's disease. In fact, the neurological, cognitive, and psychological ravages of the disease in an advanced stage are so destructive that they are completely irreversible.
Despite this unrealistic hope, the film wonderfully drives home the principal message against institution-based therapy; it thoroughly stresses home-based care and empathy. This is particularly important from the Indian viewpoint, for two reasons. First, the country doesn't have enough rehabilitation centres for Alzheimer's disease, even though the number of patients is steadily rising amidst the enormous aging population. Second, the majority of poor Indians will never be able to afford institutionalized care. Proper understanding of the disease and family-centred and home-based interventions, where possible, will be a growing need in India. Films such as Black inspire us to rise to the challenge of necessity.
Citation: Bagchi S (2005) Vision in film. PLoS Biol 3(7): e249.
Sanjit Bagchi is a medical practitioner and medical journalist based in Calcutta, India. E-mail: [email protected]
==== Refs
More Information
World Health Organization World Sight Day: 10 October 2005 Available: http://www.who.int/mediacentre/news/releases/pr79/en . Accessed 17 May 2005
HelpAge India Alzheimer's disease 2005 Available: http://www.helpageindia.org/alzheimer.php# . Accessed 18 May 2005
| 0 | PMC1174824 | CC BY | 2021-01-05 08:28:15 | no | PLoS Biol. 2005 Jul 12; 3(7):e249 | utf-8 | PLoS Biol | 2,005 | 10.1371/journal.pbio.0030249 | oa_comm |
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1600002310.1371/journal.pbio.0030255PrimerMicrobiologyMolecular Biology/Structural BiologyEubacteriaAfter 30 Years of Study, the Bacterial SOS Response Still Surprises Us PrimerMichel Bénédicte 7 2005 12 7 2005 12 7 2005 3 7 e255Copyright: © 2005 Bénédicte Michel.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.
Precise Temporal Modulation in the Response of the SOS DNA Repair Network in Individual Bacteria
The bacterial SOS response kicks in when bacteria experience DNA damage, and helps the organisms correct and survive DNA damage events. This primer provides a foundation for understanding these events.
==== Body
In order to survive in various environmental conditions, cells have a repertoire of genes that they can choose to express or silence according to their needs. Among this vast collection of genetically controlled networks, the SOS response is an inducible DNA repair system that allows bacteria to survive sudden increases in DNA damage. The importance of the SOS response is underscored by the observation that this regulatory network is widely present in bacteria, reflecting the need for all living cells to maintain the integrity of their genome.
Jump-Starting DNA Repair
The first experimental support for the existence of an inducible DNA repair network in Escherichia coli was found 30 years ago by Miroslav Radman, who introduced the term “SOS response” to describe this network [1]. Two proteins play key roles in the regulation of the SOS response: a repressor named LexA and an inducer, the RecA filament. During normal growth, the LexA repressor binds to a specific sequence—the SOS box, present in the promoter region of SOS genes—and prevents their expression. SOS genes are repressed to different degrees under normal growth conditions. This depends on the exact sequence of their SOS box (the region of a promoter that is recognized by LexA), its position in the promoter region, and the strength of the promoter.
When the cell senses the presence of an increased level of DNA damage, the LexA repressor undergoes a self-cleavage reaction and the SOS genes are de-repressed (Figure 1). A nucleoprotein complex—the RecA filament—induces the LexA cleavage reaction. RecA is a ubiquitous protein, present in nearly all bacteria and conserved in all organisms, including humans. It specifically binds single-stranded DNA (ssDNA), forming a nucleoprotein filament that has two functions [2]: the RecA filament may either invade a homologous double-stranded DNA sequence and catalyze strand exchange (the key reaction of homologous recombination), or it may promote LexA cleavage (thereby inducing the SOS response). However, RecA binding to ssDNA is also regulated. It is prevented in vivo by the ubiquitous presence of the ssDNA binding protein. Two systems allow RecA to overcome the ssDNA binding protein barrier on certain substrates: the RecFOR proteins assist RecA binding to single-strand gaps, and the RecBC proteins directly load RecA on the processed double-strand ends. Consequently, DNA-damaging agents that induce the formation of DNA single-strand gaps, such as UV light, will induce the SOS response only if the RecFOR proteins are present, whereas those that create DNA double-strand ends, such as topoisomerase poisons, will require the RecBC proteins for SOS induction [3].
Figure 1 An Oscillatory Behavior for the SOS Response
In non-induced growth conditions, the LexA repressor binds to SOS-controlled promoters, limiting or preventing their action. The basal level of expression of the genes that belong to the SOS regulon is variable. For example, in non-induced cells there are 7,500 molecules of RecA and undetectable amounts of Pol V. Upon DNA damage, RecA filaments formed at sites of damage activate the autocleavage of the LexA repressor, allowing SOS gene expression. SOS induction is reversed when damages are repaired. This is due to the disappearance of the RecA filament and allows the newly synthesized LexA molecules to bind SOS promoters. The recent work by Joel Stavans' laboratory provides evidence that, after DNA damage, individual cells oscillate between an induced and a less-induced state, and that the level of DNA damage governs the number of high-induced phases rather than their amplitude and timing [11]. Grey circles, LexA; white circles, RecA.
The SOS response has become a paradigm for the field of DNA repair. During the past 30 years, several laboratories have addressed questions concerning the function of the SOS genes and mechanisms that fine-tune their regulation. Classical techniques used to study the SOS response involved treatments of bacterial cultures by a DNA-damaging agent followed by analysis of reporter genes fused to an SOS promoter, or the direct quantification of LexA or RecA proteins by immunoblotting. More recently, microarrays were used to measure the timing and the amplitude of the induction in bacterial populations. About 40 genes were shown to be under SOS control. Most are DNA repair genes, but there are several genes that still have no known function [4,5].
SOS Genes and Their Induction Order
After UV irradiation, the amount of LexA repressor decreases nearly 10-fold in a few minutes [6]. The SOS genes, however, are not all induced at the same time and to the same level. The first genes to be induced are uvrA, uvrB, and uvrD. These proteins, together with the endonuclease UvrC, catalyze nucleotide excision repair (NER), a reaction that excises the damaged nucleotides from double-stranded DNA. As a second defense against DNA lesions, expression of recA and other homologous recombination functions increase more slowly, about 10-fold. Homologous recombination allows the repair of lesions that occur on ssDNA regions at replication forks by rendering them double-stranded (and hence a substrate for NER). The division inhibitor SfiA is also induced to give the bacterium time to complete the repairs. Finally, about 40 minutes after DNA damage (and if the damage was not fully repaired by NER and homologous recombination), the mutagenic DNA repair polymerase Pol V (encoded from umuC and umuD genes) is induced [7]. This last-ditch response also allows bacteria to render DNA lesions double-stranded—hence reparable, but at the expense of introducing errors into the genome.
Clearly, it is important for bacteria to keep all levels of the SOS response under tight control; there is no utility to the organism of using error-prone polymerases longer than absolutely necessary. Therefore, it should come as no surprise that lexA itself is an SOS gene. The constant production of LexA during the SOS process ensures that as soon as DNA repair occurs, the disappearance of the inducing signal will allow LexA to re-accumulate and repress the SOS genes. Moreover, two SOS-induced proteins, DinI and RecX, affect the stability of the RecA filament and thus may participate in the control of the SOS response [8]. Finally, in addition to DNA-damaging agents, the inactivation of certain cellular functions causes chronic SOS induction. This may be either because the gene product is involved in DNA repair and in its absence spontaneous DNA lesions persist, or because the inactivated function is essential for proper DNA duplication and the replication defect increases the amount of ssDNA [9].
Measuring the SOS Response
SOS induction was measured in bulk cultures until fluorescent microscopy techniques became available that allowed the direct measurement of gene expression in individual cells. Chronically induced cells were first used to measure SOS induction in single bacteria [10], where it was observed that the apparent homogeneity of SOS expression at the level of a population masked the occurrence of stochastic events in individuals. Indeed, the level of SOS expression in genetically identical cells grown in the same conditions was variable from cell to cell, with highly induced cells existing alongside non-induced ones. More recently, in this issue of PLoS Biology, Friedman and coworkers measured in single cells the level and kinetics of activation of SOS promoters after UV-light treatment [11]. To report promoter activity, the green fluorescent protein (GFP) gene was placed under the control of the promoters of three different SOS genes: recA, lexA, and umuCD. As expected, when the signal in a cell population was analyzed, the amount of GFP increased as a broad peak followed by a decrease as repair took place and the SOS response was shut off. Surprisingly, in individual cells, one, two, or three successive peaks of GFP expression were observed, depending on the UV dose. At UV doses lower than 10 joules (about 500 pyrimidine dimers per cell, where most should be removed by NER), one peak of GFP was observed. This was centered at 20 to 25 minutes after irradiation for the recA and the lexA promoter. Ten minutes later, as expected, the umuCD promoter was induced. At UV doses of 20 joules or higher, two to three peaks of GFP expression were observed, with the timing of the appearance of the first peak and its amplitude remaining constant. This finding changes our view of the control of the SOS response. It suggests that in each cell, the SOS response is not simply turned on to an extent that depends on the level of DNA damage and then turned off. Rather, it suggests that the SOS promoters are induced to a certain level sufficient to survive a certain dose of DNA-damaging agent, regardless of the initial amount of DNA damage. If the level of DNA damage is too high for the cells to cope with in one round of induction, a second round of induction or even a third round will follow. This interesting finding introduces a whole range of new questions, including: which factors are limiting the amplitude and controlling the timing of the peaks? The umuC and umuD genes seem to play a role in this process as their inactivation strongly perturbs the oscillatory behavior of the recA promoter. However, several models are possible as UmuC and UmuD act as a regulatory complex and as a lesion-bypass DNA polymerase [12]. Other SOS-induced proteins such as RecX and DinI that act on the RecA filament could be involved in this regulation [8]. Interestingly, digital oscillations were found also in human DNA repair governed by p53 [13], raising a parallel between the complex regulation of eukaryotic cells and the well-characterized, easily amendable SOS response of bacterial cells.
SOS and Bacterial Resistance to Antibiotics
Beyond being a model of a DNA repair regulatory network, the SOS response has played an important role in shaping the bacterial world. This is mainly because it increases mutations and genetic exchanges [14]. Pol II, Pol IV (dinB), and Pol V (umuCD) are E. coli SOS-induced DNA polymerases that are able to replicate across lesions (bypass polymerases). Among them, only Pol II is induced early and has a high fidelity on intact DNA. In UV-treated cells, these DNA polymerases can induce mutations at the site of the lesion (targeted mutations) or elsewhere (untargeted mutations); their action may be coupled to the repair of DNA double-strand breaks by homologous recombination.
The repair of double-stranded DNA breaks necessarily involves a replication re-initiation step, which can be mutagenic (Figure 2). For example, when wild-type E. coli cells are placed in the presence of a carbon source that they cannot use, some of them suffer double-strand breaks in their chromosomes that are repaired by homologous recombination, creating a substrate for Pol IV. Due to the mutagenic action of this DNA polymerase, a sub-population of cells acquires the capacity to use the carbon source and propagates [15,16]. The mutator effect of Pol III mutations, which affect the main E. coli DNA polymerase and cause chronic SOS induction, also depends in part on the action of SOS-induced polymerases, even in the absence of external damage [17].
Figure 2 A Model for SOS-Dependent Evolution to Antibiotic Resistance
Topoisomerase poisoning agents cause DNA double-strand breaks. RecBC in turn loads RecA. RecA filaments induce the SOS response and recombine. The homologous recombination reaction ends with a primer-template structure to which SOS-induced polymerases have access. DNA synthesis by these low-fidelity polymerases is accompanied by the introduction of mutations. A sub-population of mutant cells that can resist the poisoning agent invade the niche. The break-induced erroneous repair model, originally proposed for the mutagenic effects of DNA double-strand breaks in various laboratory conditions, accounts for the emergence of ciprofloxacin-resistant bacteria in a murine infection model [18]. Indented circles, RecBC; stars, SOS-induced DNA polymerases; triangle, mutation; white circles, RecA.
In the June issue of PLoS Biology, the Romesberg laboratory describes, using a murine infection model, a role for SOS induction in the appearance of E. coli mutants resistant to antibiotics [18]. The main antibiotic used is ciprofloxacin, a topoisomerase inhibitor that causes DNA double-strand breaks. The treatment of mice infections by ciprofloxacin leads to the rapid appearance of E. coli cells resistant to the antibiotic. Interestingly, when a pathogenic E. coli strain that encodes a non-cleavable LexA repressor is used, no ciprofloxacin-resistant mutant appears. This indicates that the formation of resistant cells requires SOS induction. A recombination- and SOS-dependent model is presented for the formation of ciprofloxacin-resistant mutants (Figure 2). Normally, topoisomerase poisoning causes the formation of DNA double-strand breaks, which in turn induce the SOS response and are repaired by homologous recombination. However, when the SOS response is triggered by antibiotic-induced DNA damage, the SOS-induced DNA polymerases that act at the replication forks formed by recombination generate mutants—some of which are resistant to ciprofloxacin. The SOS response is also involved, by other means, in the survival of E. coli in the presence of β-lactams [19,20]. These findings suggest that blocking SOS induction could be a general means to prevent the rapid evolution of bacteria to antibiotic resistance.
Conclusions
Work on the SOS response illustrates well the dual purpose of bacterial studies, as SOS is both a modulator of bacterial propagation during pathogenicity, and an irreplaceable source of concepts for the understanding of DNA repair regulation networks. Several important issues remain to be addressed. For example, more than a dozen SOS-induced genes encode proteins of unknown function [4]. The identification of their physiological role may reveal new levels or new means of regulation of the SOS response, links with other cellular global regulation networks, and unsuspected consequences of the SOS induction.
Citation: Michel B (2005) After 30 years of study, the bacterial SOS response still surprises us. PLoS Biol 3(7): e255.
Bénédicte Michel is a CNRS researcher at the Institut National de la Recherche Agronomique, in Jouy en Josas, France. E-mail: [email protected]
Abbreviations
GFPgreen fluorescent protein
NERnucleotide excision repair
ssDNAsingle-stranded DNA.
==== Refs
References
Radman M SOS repair hypothesis: Phenomenology of an inducible DNA repair which is accompanied by mutagenesis Basic Life Sci 1975 5A 355 367 1103845
Kuzminov A Recombinational repair of DNA damage in Escherichia coli and bacteriophage lambda Microbiol Mol Biol Rev 1999 63 751 813 10585965
McPartland A Green L Echols H Control of recA gene RNA in E. coli : Regulatory and signal genes Cell 1980 20 731 737 6998563
Fernandez De Henestrosa AR Ogi T Aoyagi S Chafin D Hayes JJ Identification of additional genes belonging to the LexA regulon in Escherichia coli
Mol Microbiol 2000 35 1560 1572 10760155
Courcelle J Khodursky A Peter B Brown PO Hanawalt PC Comparative gene expression profiles following UV exposure in wild- type and SOS-deficient Escherichia coli
Genetics 2001 158 41 64 11333217
Sassanfar M Roberts JW Nature of the SOS-inducing signal in Escherichia coli . The involvement of DNA replication J Mol Biol 1990 212 79 96 2108251
Tippin B Pham P Goodman MF Error-prone replication for better or worse Trends Microbiol 2004 12 288 295 15165607
Lusetti SL Drees JC Stohl EA Seifert HS Cox MM The DinI and RecX proteins are competing modulators of RecA function J Biol Chem 2004 279 55073 55079 15489505
O'Reilly EK Kreuzer KN Isolation of SOS constitutive mutants of Escherichia coli
J Bacteriol 2004 186 7149 7160 15489426
McCool JD Long E Petrosino JF Sandler HA Rosenberg SM Measurement of SOS expression in individual Escherichia coli K-12 cells using fluorescence microscopy Mol Microbiol 2004 53 1343 1357 15387814
Friedman N Vardi S Ronen M Alon U Stavans J Precise temporal modulation in the response of the SOS DNA repair network in individual bacteria PLoS Biol 2005 3 e238 15954802
Ferentz AE Walker GC Wagner G Converting a DNA damage checkpoint effector (UmuD2C) into a lesion bypass polymerase (UmuD′2C) Embo J 2001 20 4287 4298 11483531
Lahav G Rosenfeld N Sigal A Geva-Zatorsky N Levine AJ Dynamics of the p53-Mdm2 feedback loop in individual cells Nat Genet 2004 36 147 150 14730303
Matic I Rayssiguier C Radman M Interspecies gene exchange in bacteria: The role of SOS and mismatch repair systems in evolution of species Cell 1995 80 507 515 7859291
Foster PL Adaptive mutation in Escherichia coli
J Bacteriol 2004 186 4846 4852 15262917
McKenzie GJ Lee PL Lombardo MJ Hastings PJ Rosenberg SM SOS mutator DNA polymerase IV functions in adaptive mutation and not adaptive amplification Mol Cell 2001 7 571 579 11463382
Timms AR Bridges BA DNA polymerase V-dependent mutator activity in an SOS-induced Escherichia coli strain with a temperature-sensitive DNA polymerase III Mutat Res 2002 499 97 101 11804608
Cirz RT Chin JK Andes DR de Crécy-Lagard V Craig WA Inhibition of mutation and combating the evolution of antibiotic resistance PLoS Biol 2005 3 e176 15869329
Levin BR Microbiology. Noninherited resistance to antibiotics Science 2004 305 1578 1579 15361616
Miller C Thomsen LE Gaggero C Mosseri R Ingmer H SOS response induction by beta-lactams and bacterial defense against antibiotic lethality Science 2004 305 1629 1631 15308764
| 16000023 | PMC1174825 | CC BY | 2021-01-05 08:21:25 | no | PLoS Biol. 2005 Jul 12; 3(7):e255 | utf-8 | PLoS Biol | 2,005 | 10.1371/journal.pbio.0030255 | oa_comm |
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1600002310.1371/journal.pbio.0030255PrimerMicrobiologyMolecular Biology/Structural BiologyEubacteriaAfter 30 Years of Study, the Bacterial SOS Response Still Surprises Us PrimerMichel Bénédicte 7 2005 12 7 2005 12 7 2005 3 7 e255Copyright: © 2005 Bénédicte Michel.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.
Precise Temporal Modulation in the Response of the SOS DNA Repair Network in Individual Bacteria
The bacterial SOS response kicks in when bacteria experience DNA damage, and helps the organisms correct and survive DNA damage events. This primer provides a foundation for understanding these events.
==== Body
In order to survive in various environmental conditions, cells have a repertoire of genes that they can choose to express or silence according to their needs. Among this vast collection of genetically controlled networks, the SOS response is an inducible DNA repair system that allows bacteria to survive sudden increases in DNA damage. The importance of the SOS response is underscored by the observation that this regulatory network is widely present in bacteria, reflecting the need for all living cells to maintain the integrity of their genome.
Jump-Starting DNA Repair
The first experimental support for the existence of an inducible DNA repair network in Escherichia coli was found 30 years ago by Miroslav Radman, who introduced the term “SOS response” to describe this network [1]. Two proteins play key roles in the regulation of the SOS response: a repressor named LexA and an inducer, the RecA filament. During normal growth, the LexA repressor binds to a specific sequence—the SOS box, present in the promoter region of SOS genes—and prevents their expression. SOS genes are repressed to different degrees under normal growth conditions. This depends on the exact sequence of their SOS box (the region of a promoter that is recognized by LexA), its position in the promoter region, and the strength of the promoter.
When the cell senses the presence of an increased level of DNA damage, the LexA repressor undergoes a self-cleavage reaction and the SOS genes are de-repressed (Figure 1). A nucleoprotein complex—the RecA filament—induces the LexA cleavage reaction. RecA is a ubiquitous protein, present in nearly all bacteria and conserved in all organisms, including humans. It specifically binds single-stranded DNA (ssDNA), forming a nucleoprotein filament that has two functions [2]: the RecA filament may either invade a homologous double-stranded DNA sequence and catalyze strand exchange (the key reaction of homologous recombination), or it may promote LexA cleavage (thereby inducing the SOS response). However, RecA binding to ssDNA is also regulated. It is prevented in vivo by the ubiquitous presence of the ssDNA binding protein. Two systems allow RecA to overcome the ssDNA binding protein barrier on certain substrates: the RecFOR proteins assist RecA binding to single-strand gaps, and the RecBC proteins directly load RecA on the processed double-strand ends. Consequently, DNA-damaging agents that induce the formation of DNA single-strand gaps, such as UV light, will induce the SOS response only if the RecFOR proteins are present, whereas those that create DNA double-strand ends, such as topoisomerase poisons, will require the RecBC proteins for SOS induction [3].
Figure 1 An Oscillatory Behavior for the SOS Response
In non-induced growth conditions, the LexA repressor binds to SOS-controlled promoters, limiting or preventing their action. The basal level of expression of the genes that belong to the SOS regulon is variable. For example, in non-induced cells there are 7,500 molecules of RecA and undetectable amounts of Pol V. Upon DNA damage, RecA filaments formed at sites of damage activate the autocleavage of the LexA repressor, allowing SOS gene expression. SOS induction is reversed when damages are repaired. This is due to the disappearance of the RecA filament and allows the newly synthesized LexA molecules to bind SOS promoters. The recent work by Joel Stavans' laboratory provides evidence that, after DNA damage, individual cells oscillate between an induced and a less-induced state, and that the level of DNA damage governs the number of high-induced phases rather than their amplitude and timing [11]. Grey circles, LexA; white circles, RecA.
The SOS response has become a paradigm for the field of DNA repair. During the past 30 years, several laboratories have addressed questions concerning the function of the SOS genes and mechanisms that fine-tune their regulation. Classical techniques used to study the SOS response involved treatments of bacterial cultures by a DNA-damaging agent followed by analysis of reporter genes fused to an SOS promoter, or the direct quantification of LexA or RecA proteins by immunoblotting. More recently, microarrays were used to measure the timing and the amplitude of the induction in bacterial populations. About 40 genes were shown to be under SOS control. Most are DNA repair genes, but there are several genes that still have no known function [4,5].
SOS Genes and Their Induction Order
After UV irradiation, the amount of LexA repressor decreases nearly 10-fold in a few minutes [6]. The SOS genes, however, are not all induced at the same time and to the same level. The first genes to be induced are uvrA, uvrB, and uvrD. These proteins, together with the endonuclease UvrC, catalyze nucleotide excision repair (NER), a reaction that excises the damaged nucleotides from double-stranded DNA. As a second defense against DNA lesions, expression of recA and other homologous recombination functions increase more slowly, about 10-fold. Homologous recombination allows the repair of lesions that occur on ssDNA regions at replication forks by rendering them double-stranded (and hence a substrate for NER). The division inhibitor SfiA is also induced to give the bacterium time to complete the repairs. Finally, about 40 minutes after DNA damage (and if the damage was not fully repaired by NER and homologous recombination), the mutagenic DNA repair polymerase Pol V (encoded from umuC and umuD genes) is induced [7]. This last-ditch response also allows bacteria to render DNA lesions double-stranded—hence reparable, but at the expense of introducing errors into the genome.
Clearly, it is important for bacteria to keep all levels of the SOS response under tight control; there is no utility to the organism of using error-prone polymerases longer than absolutely necessary. Therefore, it should come as no surprise that lexA itself is an SOS gene. The constant production of LexA during the SOS process ensures that as soon as DNA repair occurs, the disappearance of the inducing signal will allow LexA to re-accumulate and repress the SOS genes. Moreover, two SOS-induced proteins, DinI and RecX, affect the stability of the RecA filament and thus may participate in the control of the SOS response [8]. Finally, in addition to DNA-damaging agents, the inactivation of certain cellular functions causes chronic SOS induction. This may be either because the gene product is involved in DNA repair and in its absence spontaneous DNA lesions persist, or because the inactivated function is essential for proper DNA duplication and the replication defect increases the amount of ssDNA [9].
Measuring the SOS Response
SOS induction was measured in bulk cultures until fluorescent microscopy techniques became available that allowed the direct measurement of gene expression in individual cells. Chronically induced cells were first used to measure SOS induction in single bacteria [10], where it was observed that the apparent homogeneity of SOS expression at the level of a population masked the occurrence of stochastic events in individuals. Indeed, the level of SOS expression in genetically identical cells grown in the same conditions was variable from cell to cell, with highly induced cells existing alongside non-induced ones. More recently, in this issue of PLoS Biology, Friedman and coworkers measured in single cells the level and kinetics of activation of SOS promoters after UV-light treatment [11]. To report promoter activity, the green fluorescent protein (GFP) gene was placed under the control of the promoters of three different SOS genes: recA, lexA, and umuCD. As expected, when the signal in a cell population was analyzed, the amount of GFP increased as a broad peak followed by a decrease as repair took place and the SOS response was shut off. Surprisingly, in individual cells, one, two, or three successive peaks of GFP expression were observed, depending on the UV dose. At UV doses lower than 10 joules (about 500 pyrimidine dimers per cell, where most should be removed by NER), one peak of GFP was observed. This was centered at 20 to 25 minutes after irradiation for the recA and the lexA promoter. Ten minutes later, as expected, the umuCD promoter was induced. At UV doses of 20 joules or higher, two to three peaks of GFP expression were observed, with the timing of the appearance of the first peak and its amplitude remaining constant. This finding changes our view of the control of the SOS response. It suggests that in each cell, the SOS response is not simply turned on to an extent that depends on the level of DNA damage and then turned off. Rather, it suggests that the SOS promoters are induced to a certain level sufficient to survive a certain dose of DNA-damaging agent, regardless of the initial amount of DNA damage. If the level of DNA damage is too high for the cells to cope with in one round of induction, a second round of induction or even a third round will follow. This interesting finding introduces a whole range of new questions, including: which factors are limiting the amplitude and controlling the timing of the peaks? The umuC and umuD genes seem to play a role in this process as their inactivation strongly perturbs the oscillatory behavior of the recA promoter. However, several models are possible as UmuC and UmuD act as a regulatory complex and as a lesion-bypass DNA polymerase [12]. Other SOS-induced proteins such as RecX and DinI that act on the RecA filament could be involved in this regulation [8]. Interestingly, digital oscillations were found also in human DNA repair governed by p53 [13], raising a parallel between the complex regulation of eukaryotic cells and the well-characterized, easily amendable SOS response of bacterial cells.
SOS and Bacterial Resistance to Antibiotics
Beyond being a model of a DNA repair regulatory network, the SOS response has played an important role in shaping the bacterial world. This is mainly because it increases mutations and genetic exchanges [14]. Pol II, Pol IV (dinB), and Pol V (umuCD) are E. coli SOS-induced DNA polymerases that are able to replicate across lesions (bypass polymerases). Among them, only Pol II is induced early and has a high fidelity on intact DNA. In UV-treated cells, these DNA polymerases can induce mutations at the site of the lesion (targeted mutations) or elsewhere (untargeted mutations); their action may be coupled to the repair of DNA double-strand breaks by homologous recombination.
The repair of double-stranded DNA breaks necessarily involves a replication re-initiation step, which can be mutagenic (Figure 2). For example, when wild-type E. coli cells are placed in the presence of a carbon source that they cannot use, some of them suffer double-strand breaks in their chromosomes that are repaired by homologous recombination, creating a substrate for Pol IV. Due to the mutagenic action of this DNA polymerase, a sub-population of cells acquires the capacity to use the carbon source and propagates [15,16]. The mutator effect of Pol III mutations, which affect the main E. coli DNA polymerase and cause chronic SOS induction, also depends in part on the action of SOS-induced polymerases, even in the absence of external damage [17].
Figure 2 A Model for SOS-Dependent Evolution to Antibiotic Resistance
Topoisomerase poisoning agents cause DNA double-strand breaks. RecBC in turn loads RecA. RecA filaments induce the SOS response and recombine. The homologous recombination reaction ends with a primer-template structure to which SOS-induced polymerases have access. DNA synthesis by these low-fidelity polymerases is accompanied by the introduction of mutations. A sub-population of mutant cells that can resist the poisoning agent invade the niche. The break-induced erroneous repair model, originally proposed for the mutagenic effects of DNA double-strand breaks in various laboratory conditions, accounts for the emergence of ciprofloxacin-resistant bacteria in a murine infection model [18]. Indented circles, RecBC; stars, SOS-induced DNA polymerases; triangle, mutation; white circles, RecA.
In the June issue of PLoS Biology, the Romesberg laboratory describes, using a murine infection model, a role for SOS induction in the appearance of E. coli mutants resistant to antibiotics [18]. The main antibiotic used is ciprofloxacin, a topoisomerase inhibitor that causes DNA double-strand breaks. The treatment of mice infections by ciprofloxacin leads to the rapid appearance of E. coli cells resistant to the antibiotic. Interestingly, when a pathogenic E. coli strain that encodes a non-cleavable LexA repressor is used, no ciprofloxacin-resistant mutant appears. This indicates that the formation of resistant cells requires SOS induction. A recombination- and SOS-dependent model is presented for the formation of ciprofloxacin-resistant mutants (Figure 2). Normally, topoisomerase poisoning causes the formation of DNA double-strand breaks, which in turn induce the SOS response and are repaired by homologous recombination. However, when the SOS response is triggered by antibiotic-induced DNA damage, the SOS-induced DNA polymerases that act at the replication forks formed by recombination generate mutants—some of which are resistant to ciprofloxacin. The SOS response is also involved, by other means, in the survival of E. coli in the presence of β-lactams [19,20]. These findings suggest that blocking SOS induction could be a general means to prevent the rapid evolution of bacteria to antibiotic resistance.
Conclusions
Work on the SOS response illustrates well the dual purpose of bacterial studies, as SOS is both a modulator of bacterial propagation during pathogenicity, and an irreplaceable source of concepts for the understanding of DNA repair regulation networks. Several important issues remain to be addressed. For example, more than a dozen SOS-induced genes encode proteins of unknown function [4]. The identification of their physiological role may reveal new levels or new means of regulation of the SOS response, links with other cellular global regulation networks, and unsuspected consequences of the SOS induction.
Citation: Michel B (2005) After 30 years of study, the bacterial SOS response still surprises us. PLoS Biol 3(7): e255.
Bénédicte Michel is a CNRS researcher at the Institut National de la Recherche Agronomique, in Jouy en Josas, France. E-mail: [email protected]
Abbreviations
GFPgreen fluorescent protein
NERnucleotide excision repair
ssDNAsingle-stranded DNA.
==== Refs
References
Radman M SOS repair hypothesis: Phenomenology of an inducible DNA repair which is accompanied by mutagenesis Basic Life Sci 1975 5A 355 367 1103845
Kuzminov A Recombinational repair of DNA damage in Escherichia coli and bacteriophage lambda Microbiol Mol Biol Rev 1999 63 751 813 10585965
McPartland A Green L Echols H Control of recA gene RNA in E. coli : Regulatory and signal genes Cell 1980 20 731 737 6998563
Fernandez De Henestrosa AR Ogi T Aoyagi S Chafin D Hayes JJ Identification of additional genes belonging to the LexA regulon in Escherichia coli
Mol Microbiol 2000 35 1560 1572 10760155
Courcelle J Khodursky A Peter B Brown PO Hanawalt PC Comparative gene expression profiles following UV exposure in wild- type and SOS-deficient Escherichia coli
Genetics 2001 158 41 64 11333217
Sassanfar M Roberts JW Nature of the SOS-inducing signal in Escherichia coli . The involvement of DNA replication J Mol Biol 1990 212 79 96 2108251
Tippin B Pham P Goodman MF Error-prone replication for better or worse Trends Microbiol 2004 12 288 295 15165607
Lusetti SL Drees JC Stohl EA Seifert HS Cox MM The DinI and RecX proteins are competing modulators of RecA function J Biol Chem 2004 279 55073 55079 15489505
O'Reilly EK Kreuzer KN Isolation of SOS constitutive mutants of Escherichia coli
J Bacteriol 2004 186 7149 7160 15489426
McCool JD Long E Petrosino JF Sandler HA Rosenberg SM Measurement of SOS expression in individual Escherichia coli K-12 cells using fluorescence microscopy Mol Microbiol 2004 53 1343 1357 15387814
Friedman N Vardi S Ronen M Alon U Stavans J Precise temporal modulation in the response of the SOS DNA repair network in individual bacteria PLoS Biol 2005 3 e238 15954802
Ferentz AE Walker GC Wagner G Converting a DNA damage checkpoint effector (UmuD2C) into a lesion bypass polymerase (UmuD′2C) Embo J 2001 20 4287 4298 11483531
Lahav G Rosenfeld N Sigal A Geva-Zatorsky N Levine AJ Dynamics of the p53-Mdm2 feedback loop in individual cells Nat Genet 2004 36 147 150 14730303
Matic I Rayssiguier C Radman M Interspecies gene exchange in bacteria: The role of SOS and mismatch repair systems in evolution of species Cell 1995 80 507 515 7859291
Foster PL Adaptive mutation in Escherichia coli
J Bacteriol 2004 186 4846 4852 15262917
McKenzie GJ Lee PL Lombardo MJ Hastings PJ Rosenberg SM SOS mutator DNA polymerase IV functions in adaptive mutation and not adaptive amplification Mol Cell 2001 7 571 579 11463382
Timms AR Bridges BA DNA polymerase V-dependent mutator activity in an SOS-induced Escherichia coli strain with a temperature-sensitive DNA polymerase III Mutat Res 2002 499 97 101 11804608
Cirz RT Chin JK Andes DR de Crécy-Lagard V Craig WA Inhibition of mutation and combating the evolution of antibiotic resistance PLoS Biol 2005 3 e176 15869329
Levin BR Microbiology. Noninherited resistance to antibiotics Science 2004 305 1578 1579 15361616
Miller C Thomsen LE Gaggero C Mosseri R Ingmer H SOS response induction by beta-lactams and bacterial defense against antibiotic lethality Science 2004 305 1629 1631 15308764
| 0 | PMC1174826 | CC BY | 2021-01-05 08:21:25 | no | PLoS Biol. 2005 Jul 12; 3(7):e264 | latin-1 | PLoS Biol | 2,005 | 10.1371/journal.pbio.0030264 | oa_comm |
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-1141589006810.1186/1471-2105-6-114Research ArticleCAGER: classification analysis of gene expression regulation using multiple information sources Ruan Jianhua [email protected] Weixiong [email protected] Department of Computer Science and Engineering, Washington University, St. Louis, MO 63130, USA2 Department of Genetics, Washington University School of Medicine, St. Louis, MO 63110, USA2005 12 5 2005 6 114 114 3 11 2004 12 5 2005 Copyright © 2005 Ruan and Zhang; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Many classification approaches have been applied to analyzing transcriptional regulation of gene expressions. These methods build models that can explain a gene's expression level from the regulatory elements (features) on its promoter sequence. Different types of features, such as experimentally verified binding motifs, motifs discovered by computer programs, or transcription factor binding data measured with Chromatin Immunoprecipitation (ChIP) assays, have been used towards this goal. Each type of features has been shown successful in modeling gene transcriptional regulation under certain conditions. However, no comparison has been made to evaluate the relative merit of these features. Furthermore, most publicly available classification tools were not designed specifically for modeling transcriptional regulation, and do not allow the user to combine different types of features.
Results
In this study, we use a specific classification method, decision trees, to model transcriptional regulation in yeast with features based on predefined motifs, automatically identified motifs, ChlP-chip data, or their combinations. We compare the accuracies and stability of these models, and analyze their capabilities in identifying functionally related genes. Furthermore, we design and implement a user-friendly web server called CAGER (Classification Analysis of Gene Expression Regulation) that integrates several software components for automated analysis of transcriptional regulation using decision trees. Finally, we use CAGER to study the transcriptional regulation of Arabidopsis genes in response to abscisic acid, and report some interesting new results.
Conclusion
Models built with ChlP-chip data suffer from low accuracies when the condition under which gene expressions are measured is significantly different from the condition under which the ChIP experiment is conducted. Models built with automatically identified motifs can sometimes discover new features, but their modeling accuracies may have been over-estimated in previous studies. Furthermore, models built with automatically identified motifs are not stable with respect to noises. A combination of ChlP-chip data and predefined motifs can substantially improve modeling accuracies, and is effective in identifying true regulons. The CAGER web server, which is freely available at , allows the user to select combinations of different feature types for building decision trees, and interact with the models graphically. We believe that it will be a useful tool to facilitate the discovery of gene transcriptional regulatory networks.
==== Body
Background
A major challenge in computational biology is to reveal the cis-regulatory logics of gene expression through analysis of high-throughput genomic data, for example, genomic sequences and microarray gene expression data. A common practice is to first identify putatively co-regulated genes by clustering gene expression patterns [1-3], and then search for common motifs from the promoter sequences of these genes [4-6]. However, motif finding methods are often sensitive to noises and usually do not consider combinatorial nature of cis-regulation. Furthermore, these methods by themselves do not reveal the actual transcription factors (TFs) that bind to particular sequence motifs.
Recently, many researchers attempted to build quantitative or qualitative models to associate a gene's expression level with regulatory motifs on its promoter sequence. Pilpel et al. [7] explicitly analyzed the combinatorial effects of motif pairs on gene expression profiles and identified many significant motif combinations. Bussemaker et al. [8] and others [9,10] modeled the expression levels of genes as a linear regression of putative binding motifs, and applied feature selection techniques to find the most significant motifs. Hu et al. [11] used decision trees to find motif combinations that best separate two sets of genes. Phuong et al. [12] applied multivariate regression trees to model the transcriptional regulation of gene expressions over several time points simultaneously. Middendorf et al. [13] used an ensemble of decision trees to model gene expression levels by combining putative binding motifs and the expression levels of putative TFs. Simonis et al. [14] combined a string-based motif finding method and linear discriminant analysis to identify motif combinations that can separate true regulons from false ones. Segal et al. [15] and Beer and Tavazoie [16] built probabilistic graphical models, e.g., Bayesian networks, to explain gene expression patterns from motifs. In these models, the predictors (features) are the matching scores of promoter sequences to putative binding motifs, and the predictions (responses) can be continuous or discrete gene expression levels or categorical cluster labels.
The features used in these models generally come from one of the following sources. First, one can use computer programs to automatically find motifs from the promoters of the genes to be modeled [10,11,14-16]. Second, predefined motifs can be obtained independently from sources such as databases of experimentally verified or putative motifs [12,13]. Third, one can enumerate all words up to a certain length as features [8,9]. In addition, TF binding data derived from Chromatin Immunoprecipitation (ChIP) assays [17] have been used as a substitution of motif scores. For example, Banerjee and Zhang [18] directly applied the method of Pilpel et al. [7] to ChIP-chip data to identify TF combinations; Gao et al. [19] replaced the variables in the linear model of Bussemaker et al. [8] with ChIP-chip data and identified significant regulators for many experimental conditions. We recently applied a decision tree method to S. cerevisiae ChIP-chip data and identified all known TFs and many interesting TF combinations for yeast cell cycle [20].
Each type of the features discussed above (motifs or binding data) has its advantages and disadvantages in modeling gene transcriptional regulations. As to our knowledge, no comparison has been made to evaluate their relative merits. Modeling accuracy is largely affected by the type of features considered in model construction. If all relevant features are included correctly, many modeling algorithms may have equally high accuracies. On the other hand, if most significant features are omitted, no model can achieve satisfactory accuracy. Furthermore, the inclusion of many irrelevant features may significantly decrease modeling accuracy. Therefore, a comparative study can identify limitations of these feature types, and provide some guidelines and justifications.
Although many classification tools are publicly available (for example, WEKA [21]), most of them are not designed specifically for modeling transcriptional regulation, and are not convenient for biological applications. The only related software is a web server called REDUCE [22], which combines linear regression and feature selection methods to identify significant motifs for specific biological events. It uses enumerated words up to a certain length as features, but does not allow other types of features such as position specific weight matrices [23] or ChIP-chip data to be used. Moreover, the linear model used in REDUCE assumes that each motif contributes linearly to gene expressions, and therefore is unable to represent complex cis-regulatory logics such as AND and OR relations [16,24].
In this study, we apply a well-studied classification method, decision tree [25-27], to model significantly up-or down-regulated genes in each of 250 microarray experiments of S. cerevisiae. The utilization of decision trees in modeling transcriptional regulation has been explored previously by others [11-13] and in our own research [20]. Here we focus on analyzing the extent to which the expression of these genes can be predicted using different features. We compare the cross-validation accuracies of the models built with different features or feature combinations. We also compare the robustness of these models by introducing noises into training data. Furthermore, we analyze the enrichment of functional categories of the genes that can pass model tests comparing to those fail the tests, and show that decision tree models can be used to detect true regulons. Finally, we present the design and implementation of a user-friendly web interface that combines multiple information sources for automated analysis of gene transcriptional regulations using decision trees. As an example, we also present a case study on the transcriptional regulation of genes in Arabidopsis thaliana in response to abscisic acid (ABA) treatment.
Results and discussion
Modeling gene transcriptional regulation with decision trees
Here we briefly introduce the modeling of gene transcriptional regulation with decision trees. For a detailed treatment of decision trees, the reader is referred to related literature [20,25-27]. Suppose that there are N genes (instances), each of which is represented by a feature vector F = f1, f2,..., fm and has a class label c, where fi is a real number and c is a category. A decision tree is built as follows. Initially, the tree has only one node, the root, which contains all the genes. Then for each node that has no child node,
1. Examine every possible binary split of the genes in the node based on each feature i, such that all genes in one subset have fi <x and those in the other subset have fi ≥ x.
2. Select the best split, and create two child nodes that contain the two subsets of genes respectively.
Steps 1 and 2 are then recursively applied to each of the child nodes until no split is possible, or until all genes in the current node have the same label. Finally, some branches of the tree may be pruned to prevent over-fitting. Nodes with or without child node are called internal nodes or leaf nodes, respectively. For examples and biological interpretations of decision trees, see Figure 4 and Figure 5.
Each entry fi in the feature vector of a gene corresponds to the matching score of the gene's promoter to the ith binding motif, or the binding data of the ith TF to the gene's promoter, depending on the type of features used. A split is equivalent to a test for a gene in the form of, for example, "is the matching score of the gene's promoter to motif A greater than x?" or "is the binding affinity of the gene's promoter to TF B greater than y?" The exact split point x or y is determined by maximizing an objective function that reflects the purity of the child nodes. Information gains and gain ratios are two frequently used objective functions [26]. Here we used information gains (see Methods).
The class label of a gene represents a property of the gene that we want to model. For example, one can cluster genes according to expression patterns, and then assign the same label to genes in the same group. Labels can also be derived from other sources such as functional annotations. In this work, we assign labels to genes according to the change of their expressions under certain conditions relative to some reference state (see Experimental setup), and focus our attention on the comparison of different features. This modeling approach is based on the assumption that co-regulated genes very often share common regulatory elements on their promoters. This approach, on the other hand, will not capture post-transcriptional modifications, and will ignore genes that share no regulatory elements with other genes. Note that the underlying assumption may not always be met, since not all co-expressed genes are co-regulated. Moreover, genes may be mislabeled due to noises in their expression data. The purpose is, therefore, to identify rules of thumb for the regulation of the majority genes, while tolerating to some extent the failure in modeling particular genes. As we will see later, decision tree models can indeed be used to detect the true regulons from putatively co-regulated genes.
Experimental setup
We collected data for 250 microarray experiments on yeast S. cerevisiae, of which 77 were for cell cycle [28,29] and 173 were for responses to various stress conditions [30]. We built decision trees for each condition with two classes of instances: positive genes that are differentially expressed (up- or down-regulated) with respect to the reference state, negative genes that are neither up- nor down-regulated. For each experiment, we selected up to 50 up- or down-regulated genes as positive instances and sampled negative instances from non-differentially expressed genes. (See next subsection for selecting candidate genes). Once the genes were selected, we modeled the regulation of up- and down-regulated genes separately. There may be common regulatory motifs shared by up- and down-regulated genes, which may not be revealed if they were modeled together.
Since most genes are not differentially expressed, the number of negative instances is far greater than the number of positive instances. Previous researches have shown that class distribution is an important factor for successful modeling [31]. In general, a skewed class distribution will lead to a degraded modeling accuracy. Therefore, for a set of n positive instances, we randomly sampled μn instances from all possible negative instances to obtain a desired class distribution. The sampling was repeated 10 times, and each set of sampled negative instances was combined with the set of positive instances to build a decision tree. The accuracy of each model was estimated with a ten-fold cross-validation and measured by the kappa statistic (see Methods). As shown in Figure 1A, the highest accuracy was achieved when μ = 3, which is consistent with the results of Simonis et al. [14]. Therefore, we used μ = 3 in all subsequent analysis.
We considered three types of features. The first type contained 356 known and putative motifs compiled by Pilpel et al. [7] (referred to as predefined motifs). Of the 356 motifs, 25 were obtained from literature, and the rest were discovered from the promoters of genes sharing similar MIPS functional categories [32]. The second type of features included motifs identified from the promoter sequences of positive genes by the AlignACE program [5] (referred to as auto motifs). The third type of features was derived from the in vivo binding data of 113 transcription factors [17] (referred to as ChIP data). For the first two types of features, each motif was represented as a position specific weight matrix. A promoter was scanned with ScanACE [5] for all motifs in the feature set, and the highest matching score for each motif was used. For the third type of features, the binding affinity of each TF (p-value < 0.001, according to [17]) to a promoter sequence was used.
In general, the inclusion of a large number of irrelevant features in training data decreases the accuracy of most classification algorithms. Therefore, feature selection methods are usually applied to reduce the number of features. Most feature selection methods can be categorized as wrappers [33] or filters [34]. A wrapper method searches for a subset of features that maximize the cross-validation accuracy of a given classification algorithm. This strategy is guaranteed to improve the classification accuracy if the same algorithm is used in feature selection and model training. However, it may over-fit to the specific classification algorithm. Furthermore, the method is computationally expensive since many iterations of the classification algorithm need to be executed. In a filter method, features are selected independently of any classification algorithm. Individual features or feature subsets are ranked according to certain scoring functions, and the top ones are selected. This approach is efficient in removing a large number of irrelevant features, but may sometimes eliminate low-ranked, nevertheless important, features. In this study, we used a filter method because of its efficiency in computation. The method ranks individual features according to their information gains and selects the top d features (see Methods and [35]). As shown in Figure 1B, the best kappa was achieved with as few as 5 features. This agrees with the fact that a transcriptional regulation only involves a few transcription factors in general. Careful inspections on individual decision trees show that with 5 features, the performance of some trees may be worse than those with more features, due to the loss of some significant features. With 10 features, the modeling accuracies were almost never worse than those with more features. Therefore we used d = 10 in all subsequent analysis. The accuracy of each classification model was estimated using a ten-fold cross-validation procedure. First, a training set was randomly divided into ten equal-sized subsets. Each subset was then used in turn as a validation set to test the accuracy of the model built with the other nine subsets. We calculated kappa [36] to measure model accuracies (see Methods). The kappa statistic, written κ, measures the agreement between the class labels and the predictions made by the classifier, corrected by the amount of agreement that may be achieved by chance. Therefore, it reflects the extent to which the differential expression between positive and negative genes can truly be explained by the classification model. For example, given a data set containing 20 positive and 80 negative genes, a model that simply guesses all genes as negative agrees with the true labels on 80% of cases, as does another model that makes five mistakes in positive and 15 in negative genes. Taking into account the amount of agreement that we would expect by chance, the value of κ is 0.0 for the former model, while 0.47 for the latter.
It is known that κ depends on the class distribution and the number of categories in the test data [37]. This, however, was not a problem in our case, since the models in our test all had binary classes and the same class distributions. Another difficulty associated with κ is its lack of interpretability, although some relations between κ and model quality have been suggested [36]. Therefore, in addition to κ, we calculated sensitivity (SS) and specificity (SP) for each model (see Methods). SS is the proportion of positive genes that are correctly predicted by the model, i.e., the proportion of up- or down-regulated genes that can be explained by the regulatory elements identified. SP is the proportion of negative genes that are correctly predicted by the model, and 1 - SP represents the proportion of negative genes that cannot be separated from the positive genes based on the regulatory elements.
Methods for identifying DEG candidates
Microarray data are noisy and often measured with limited or no replication, which makes the identification of differentially expressed genes (DEGs) difficult. In most early microarray analysis, a fixed fold-change threshold (generally two-fold) was used to identify DEGs, while more sophisticated methods have emerged recently [38-41]. Although it is not the focus of this paper, we considered and compared several different DEG identification methods to show that our conclusions on the classification models are unlikely to be affected by the specific DEG identification method used.
For the first method (referred to as the vanilla method), we downloaded the transformed and normalized log ratios for the 250 microarray experiments. The data set had been corrected for background noises, and globally normalized by constant factors such that the mean log2(cy5/cy3) value is zero within each slide [40]. For expression data in cell cycle conditions, the log ratios were further normalized such that the mean log ratio for each gene across all cell cycle conditions is zero [28]. Several rules were also applied to remove spurious data points [28]. For each column (condition) of this data set, we selected the genes with log2(cy5/cy3) ≥ 1 (more than two-fold induction) as up-regulated genes. In cases there were more than 50 up-regulated genes, we selected the top 50 with the highest fold changes, or until ties were broken. Likewise, down-regulated genes with log2(cy5/cy3) ≤ - 1 were selected. Genes with |log2(cy5/cy3)| ≤ 0.6 (less than 1.5 fold expression change) were considered as non-DEGs and were used to sample negative instances. Note that we intentionally used two different thresholds for DEGs and non-DEGs, in order to exclude genes whose labels may be ambiguous.
For the second to fifth methods, we downloaded the raw intensity data. However, we only found raw data for 216 of the 250 conditions, and it was sometimes unclear how to match the name of a raw data file with a column in the log ratio data. The intensity data were background corrected, without any other normalization or transformation. We removed low quality data points that were annotated by the authors with failed status or non-zero flags.
The second approach (referred to as the global normalization method [40]) was similar to the vanilla approach, except that no per-gene normalization was made. The third approach (referred to the lowess normalization method) was similar to the second approach, except that within-slide normalization factors were intensity dependent, obtained through a locally weighted regression approach [40]. The fourth method (referred to as the vsn method) transformed the intensities with a generalized logarithm function, in order to stabilize the variances which were originally intensity dependent [39]. We applied the same thresholds as in the vanilla approach to the transformed data. The fifth method (referred to as the EDGE method) quantified the heteroscedasticity as a function of intensity and then taking it into account to identify DEGs [41]. The EDGE method assigned each gene a p-value based on its distance from the line of equivalence (cy5 = cy3), corrected for multiple tests. We ranked genes according to their corrected p-values, and selected up to 50 up-regulated or down-regulated genes with false discovery rate (FDR) < 0.002 [42]. The FDR threshold was used to ensure that the number of DAGs selected by EDGE was approximately equal to those chosen by the other approaches. It had no impact for most experiments, where all top 50 genes had FDRs less than 0.002. Genes with FDR > 0.5 were considered as non-DEGs. The programs for vsn and EDGE were obtained from their original authors.
We used each of the five methods to select DEGs and non-DEGs for each microarray. DEG sets with less than 20 genes were not used. Table 1 lists the average group size and the number of DEG sets identified by each method. The accession numbers for the genes in each set can be viewed on the supplementary website [43].
For each set of positive genes (up- or down-regulated), we randomly sampled threefold negative genes from the corresponding non-DEGs to construct a decision tree and performed a ten-fold cross validation. The random sampling was repeated for ten times for each DEG set. The cross-validation accuracies for all models are included in Supplementary Table 1 (see Additional file 1). Figure 2 shows the values of κ, SS and 1 - SP averaged across all gene sets obtained by each method. The five different methods showed similar accuracies in all three measures. Although two of the methods (global and lowess) seemed to have slightly better SS and kappa, the difference is not significant. We therefore restricted our subsequent analysis on the DEGs identified using the vanilla method, which resulted in the largest number of DEG groups.
A comparison of the prediction power of different features
We compared the cross-validation accuracies of the models built with three types of features that we discussed early: ChIP-chip data, predefined motifs and auto motifs. The combination of ChIP-chip data with predefined motifs was also tested. For each type of features or feature combinations, 446 × 10 = 4460 decision tree models were built (446 sets of DEGs and 10 sets of negative genes randomly sampled for each set of DEGs). We randomly exchanged the labels for positive and negative genes to serve as controls. Note that we carried out two types of cross-validations for models built with auto motifs. In the first method, promoter sequences of genes in both training data and test data were combined to find motifs. In other words, motifs were identified from all positive genes, and the same set of motifs was used to train models for each fold in a cross-validation. In the second method, genes were first divided into ten subsets without constructing the actual feature vectors. A subset was chosen for testing, and the other nine subsets were used for motif finding. In other words, motifs were identified from only the training genes, and a different set of motifs was used to train models for each fold in a cross-validation. The second method provided a more stringent estimation of the generalization accuracy of a model, since it completely hided the test data from the learning algorithm until they were tested. The first method, however, was used in several previous studies [11,16], probably because it is simple to implement and convenient to test. Here we analyzed the results of both cross-validation methods to compare auto motifs with other feature types. In the next two subsections, we used only the first method to show other aspects of the models based on auto motifs.
A complete list of the cross-validation accuracies of models for each microarray experiment is included in Supplementary Table 2 (see Additional file 2). The mean cross-validation accuracies of models for genes in cell cycle and stress conditions are shown in Figure 3.
As shown in Figure 3A, when the first cross-validation method is used, the models using auto motifs have the highest kappa values (~0.53) among the three individual feature types. However, it is important to note that these models also have the highest kappa values on randomly selected genes (~0.4).
Furthermore, the accuracies measured by the second cross-validation method are much lower: the average kappa values are 0.15 for models in cell cycle and 0.22 for models in stress response experiments, and are approximately zero for models of randomly selected genes. Therefore, the first method considerably over-estimates the accuracies of the models built with auto motifs. This is because that, with the first cross-validation method, the feature set contains some information about the test instances, even though the models are built only on training instances. Consequently, although the results reported in previous studies utilizing automatically identified motifs [11,16] may still be valid qualitatively, the exact accuracies may need to be re-evaluated. Nevertheless, an apparent advantage of using automatically identified motifs is that it may be able to discover new features not included in predefined motifs and ChIP data.
In cell cycle experiments, the models using ChIP data or predefined motifs have similar kappa values (0.27 and 0.29, respectively; p-value = 0.15 in a paired t-test). In contrast, in stress response experiments, the models using ChIP data alone have very low kappa values than that using predefined motifs (0.15 vs. 0.33, p-value = 10-47 in a paired t-test). The ChIP data used in this study were measured under normal cell growth conditions. However, it is known that TF binding may change with environmental conditions [44]. While the cell cycle expression data were measured under conditions relatively similar to normal cell growth conditions, stress treatment dramatically changes the environmental conditions and thereby alters the binding of TFs. It is thus expected to observe lower prediction accuracies in stress response experiments than in cell cycle experiments when using ChIP data alone.
Figure 3B and Figure 3C show the mean SS and 1 - SP for the models built with different features.
Figure 3B for SS shows almost identical trends as Figure 3A, which means that with combined features, the models can better explain the co-regulation of the genes. On the other hand, Figure 3C shows that models built with ChIP data have higher specificity than those built with predefined or auto motifs (p-value = 10-68 in a paired t-test between ChIP data and predefined motifs). This can be explained by the fact that ChIP data are more reliable than motif scores as an indicator of co-regulations, since ChIP data explicitly measure the binding affinities of gene promoters to TFs. There are also other advantages in using ChIP-chip data as features. For example, the number of features is well bounded by the number of TFs, which is estimated to be around 200 in yeast [44], comparing to thousands of putative binding motifs. As a result, the correlations and redundancies among features are low in ChIP-chip data, which makes it possible to build simple models for better interpretability. Furthermore, the models built with ChIP-chip data directly suggest regulatory relations between TFs and genes, which can be used to construct regulatory networks. Our results indicate that, however, the conditions of microarray experiments and ChIP assays must be considered with special care.
In cell cycle experiments, the models built with a combination of predefined motifs and ChIP data have substantially better kappa values than those with ChIP data or predefined motifs alone (Figure 3D). The p-value is 10-14 in a paired t-test for results of ChIP data (0.27) and combined features (0.34), and 10-12 for predefined motifs (0.29) and combined features (0.34). In stress experiments, the kappa values for models based on combined features (0.35) are only marginally better than that for predefined motifs (0.33), due to limited usage of ChIP data as pointed out above. Nevertheless, the models using combined features perform significantly better (improvement of kappa > 0.1) in 33 cases, while the models using predefined motifs are never better by more than 0.1 in kappa (Figure 3E). A paired t-test yields a p-value 10-5, which is still statistically significant. Putative binding motifs and ChIP data represent two distinct and complementary sources of evidence of regulation. Therefore, a combination of them can provide a better discriminative power than either of them does.
Since the role of decision tree models is exploratory, there is no need to be restricted to any specific type of features. Indeed, one should try a variety of them to identify the most relevant features for a particular set of genes. A good strategy in choosing feature types is probably as follows: first use a combination of TF binding data and predefined motifs for modeling; if a model with good cross-validation accuracy can not be found, then consider using automatically identified motifs. Obviously when predefined motifs or TF binding data are unavailable or insufficient, using automatically identified motifs as features is the only option.
Stability of models
To analyze the stability of decision tree models, we introduced some noises into positive training data and tested whether the models can separate the true positives from the noises. For each microarray experiment we randomly selected 5 to 50 negative genes and deliberately mislabeled them as positive. The original data and the additional fake positive instances were then combined to build a decision tree model. We compared the results of the new models to the results of the original models in ten-fold cross-validations and counted the numbers of losses, rescues and false positives (FPs). According to [14], a loss is a positive instance that is mis-classified in the noisy data but correctly classified in the original data. A rescue is a positive instance that is correctly classified in the noisy data but mis-classified originally. An FP corresponds to a newly added noise gene that is classified as a (false) positive gene.
As shown in Table 2, the models built with predefined motifs are more robust than the models built with auto motifs. Even at 100% noise level (50 randomly selected genes added), almost 90% of the introduced noises can be correctly filtered out by the models built with predefined motifs. In comparison, the models built with auto motifs can only filter out ≈ 75% of the noises. This is because the motif finding program was distracted by the wrongly labeled genes so that the discovered motifs became less effective to discriminate the true positives from false ones.
Furthermore, the predictions made by the models built with predefined motifs are more stable, as reflected by the smaller values of loss + rescue in Table 2. The addition of a small amount of noises may result in, dramatic change to the predictions by the models built with auto motifs. When as few as 5 noisy instance were added, the models based on auto motif changed their predictions by ≈ 58%, while the models based on predefined motif changed their predictions by only ≈ 30%.
Functional enrichment in correctly classified positive genes
As mentioned early, there may be errors in the labels of training instances, i.e., some genes assigned with the same label may actually not be co-regulated. Here we test decision tree models' capabilities of identifying "true" regulons from such genes. Given a set of putatively co-regulated genes as positive instances and randomly sampled genes as negative instances, we used a leave-one-out strategy to predict which of the positive genes are indeed co-regulated. To this end, a positive instance was removed from the training data, and tested on the decision tree model built without it. This was repeated for each positive gene. Consequently, the positive genes were categorized into two groups: those that were predicted as true positives and those as negatives, which we refer to as Gp and Gn, respectively. We hypothesized that the genes in Gp are more functionally coherent than the genes in Gn, since according to the decision tree model, the genes in Gp share some common regulatory elements. To test this hypothesis, we counted the number of Gene Ontology (GO) functional categories [45] that are statistically enriched in each group with a false discovery rate < 0.05 (see Methods). The above procedure was applied to the up- and down-regulated genes in each of the 250 microarray experiments and the average results were calculated. As shown in Table 3, the models built with predefined motifs and with auto motifs both succeed in retaining functionally related genes and filtering out unrelated genes. For example, in the models built with predefined motifs for up-regulated genes in cell cycle, there are in average 15 enriched GO categories in 17 Gp genes, while there are only 8 enriched GO categories in 27 Gn genes. A similar trend can be observed for down-regulated genes and for the models built with auto motifs. A paired t-test between and combining all cases yields a p-value 10-43.
Furthermore, we tested whether a similar degree of functional enrichment can be achieved without decision tree models. Suppose a decision tree model predicted 25 out of 50 positive genes as true positives, and GO analysis showed that these 25 genes were functionally more coherent than the remaining 25 genes. We want to know whether there is another way to select 25 genes that have higher degree of functional enrichment than those predicted by decision trees. For example, we may simply pick genes that are top ranked according to their expressions. We used the following procedure to test this hypothesis. For each set of up-or down-regulated genes, A, which has p predicted true positive genes, we selected p top ranked genes from A based on the absolute log ratios of their expressions. We denoted this set of genes as Gp', which has the same number of genes as Gp, and counted the number of enriched GO categories in it. The average results over all microarray experiments are shown in the column of Table 3. As shown, the genes in Gp contain more enriched functional categories than genes in Gp'(). A paired t-test between and combining all cases has a p-value 10-16. Therefore, we can conclude that both models built with auto motifs and with predefined motifs are more effective in selecting functionally related genes than the naive differential expression model.
CAGER web server
Several previous studies have shown that decision tree is a valuable tool in analyzing transcriptional regulation of gene expressions [11-13,20]. Although there are many publicly available software packages for building decision trees (for example, [21,25,26]), they are not specifically designed for biological applications, and are not convenient for biologists to use. Therefore, to make a good use of the results from this study, we designed and implemented a user-friendly web server and interface for building decision trees to analyze transcriptional regulation. The server integrates several software components that allow the user to select from different types of features and to interact with the constructed models.
The interface for user inputs is shown in Figure 4A. To submit a job to CAGER, the user is first asked to provide positive and negative genes as either ORF identifies for one of the supported organisms, which currently includes yeast S. cerevisiae and Arabidopsis thaliana, or promoter sequences in FASTA format. These can be copy-pasted to the web form or uploaded from files in the user's local computer. Given ORF identifiers, the promoter sequences in a supported genome are retrieved automatically from a local database. Second, the user specifies some feature sets, which may be ChIP-chip data for yeast S. cerevisiae [17], predefined motifs from Pilpel et al. [7], motifs automatically identified from the promoter sequences by AlignACE [5], or a combination of them. The user can also specify whether features should be identified from negative instances as well as from positive instances, whether feature filters should be used, and the minimum number of instances per node. Finally, the user fills in his or her email address and submits the job. After a job is completed, the user will be notified by email for instructions about how to access the results.
On the output page, a decision tree is displayed as a portable network graph (PNG), along with related statistics for the tree in training and cross-validation processes (Figure 4B). The text inside an internal node of the tree gives the name of a feature, and the text inside a leaf node shows the predicted label for genes inside the node, as well as the number of supporting and counter-instances for the prediction. Each node of the decision tree can be clicked to show some details. For example, if an internal node contains a feature derived from ChIP-chip data for a TF in yeast, clicking on it leads the user to SGD [46] for detailed information about the TF. If the feature is a binding motif, a click opens a new window to display the sequence logo [47] and the position specific weight matrix of the motif. A click on a leaf node brings up a window for displaying the identifiers of the positive and negative genes in the leaf.
Application to ABA-responsive genes in Arabidopsis
Here we show an example of using the web server to study the regulation of genes expressed in response to abscisic acid (ABA) in Arabidopsis. ABA is a phytohormone that plays important roles in many stages of plants, such as seed development and stress responses (see [48] for a review). Seki et al. [49] identified about 250 genes in Arabidopsis that are induced by at least 5-fold after ABA treatment. Since Arabidopsis is one of the supported organisms in our current server, its promoter sequences are available in our database.
Therefore, we provided as positive instances the list of ORF names corresponding to the up-regulated genes, and as negative instances a list of randomly selected genes that are not up-regulated. Promoter sequences were retrieved for 152 of the positive genes. We used auto motifs identified from these promoters as features. The decision tree and the sequence logos for the most interesting motifs are shown in Figure 5. AlignACE identified a total of 37 motifs with default parameters, five of which were selected by the decision tree (Figure 5A). Three motifs, ace_m2, ace_m9, and ace_18 (Figure 5B) together correctly classified 35 (= 13 + 6 + 22 - 6) positive genes (the rightmost three leaves labeled with 'p'), while as many as 107 positive genes were classified as negative. This may be due to the fact that ABA triggers a lot of down-stream responsive genes, many of which are not co-regulated with direct targets of ABA. The motif ace_m2 has a conserved CACGTG core, which is very close to the known ABA responsive elements (ABREs) identified in many other plants [50-56]. It is known that an ABRE often functions together with a coupling element (CE), but the consensus sequence of CE in Arabidopsis is elusive. Here, the decision tree suggests that ace_m18 and ace_m9 may be two possible CEs. Motif ace_m18 has a CGTGTG core which partially resembles the CE in rice OSEM gene (GACGCGTGTC) [57], and CE3 in barley HVA1 gene (ACGCGTGTCCTC) [55]. Note that ace_m18 has a weak second copy of the GTGTG core, whicl may be important for enhanced binding activity. Motif ace_m18 is also remotely similar to CE3 in maize RAB28 gene (ACGCGCCTCCTC) [53], although in maize the CGTGTG core is replaced by CGCGCC. Motif ace_m9 is a weak but significant motif that consists of a series of G's separated by one or two A's. This motif is not similar to any known motif in the database of Plant Cis-acting Regulatory DNA Elements (PLACE) [58], and therefore may be a new motif for Arabidopsis.
Conclusion
In this research, we compared the effect of using different features to study transcriptional regulation of gene expressions by classification methods. We considered features based on ChIP-chip data, predefined motifs, automatically identified motifs, and their combinations. We found that TF binding data from ChIP assays are effective in modeling gene expressions only under the same conditions where ChIP-chip experiments were conducted. Our results also indicate that many previous studies may have over-estimal the cross-validation accuracies of models built with automatically identified motifs. Furthermore, the models built with automatically identified motifs are not robust with respect to noises, comparing to those built with predefined motifs. A combination of ChIP-chip data with predefined motifs seems to be superior to either one of them applied separately.
We also showed that the positive genes correctly predicted by decision tree models are more functionally related than those that are not correctly predicted. Therefore, decision tree models can be used to refine putative regulons and detect new genes of a regulon. Simonis et al. [14] have showed this by testing on known regulons, while we confirmed this through analyzing the functional enrichment of predicted regulons. We presented a web service that integrates motif finding and decision tree learning for analyzing transcriptional regulation of gene expressions. Its usefulness was illustrated with an example of studying the regulation of ABA-responsive genes in Arabidopsis. We identified two motifs that are similar to known ABA-responsive elements (ABREs) and coupling elements (CEs), and suggested a new CE, which may deserve further studies. As demonstrated by the example, the web interface combines a number of software components and hides most specific parameters from the user, while still allows some flexibilities. The graphical representation of a decision tree makes it easy to visualize and extract significant regulatory rules. We believe that it can significantly reduce the learning curve for those who are interested in applying classification methods to analyzing transcriptional regulation, and will be a useful tool to facilitate the discovery of transcriptional regulatory networks by combining multiple information sources.
Methods
Datasets
Yeast gene expression data were downloaded from from Expression Connection [59]. ABA-induced gene expression data for Arabidopsis were obtained from [49]. ChIP-chip data for yeast transcription factors were downloaded from [60]. Putative binding motifs for yeast genome were obtained from [61]. Upstream sequences of yeast and Arabidopsis ORFs were obtained from RSA-tools [62] by retrieving up to 800 and 1500 bases, respectively, from translation start sites. Overlaps with other ORFs were truncated, and upstream sequences shorter than 100 bases were discarded. For yeast ORFs, the upstream sequences were used as promoters. For Arabidopsis ORFs, the upstream sequences were used as inputs to a promoter prediction program, TSSP [63], to predict transcription start sites (TSSs). Promoters of Arabidopsis ORFs were defined as from 350 bases upstream to 50 bases downstream relative to the predicted TSSs, where most known ABREs and CEs were discovered.
Feature filters and decision tree learning
Given a training set, features were ranked according to the restricted information gains that can be achieved by using individual features to separate positive and negative instances. Suppose that there are p positive instances and n negative instances, and by selecting a splitting point x on feature fi, the numbers of positive and negative instances on the left (fi <x) and right (fi ≥ x) child nodes are p1, n1, p2, and n2, respectively. The restricted information gain due to feature fi with respect to this split can be calculated by:
where Igain is the normal information gain computed from entropies [26]. The restriction to the information gain calculation ensures that a selected feature is more over-represented in positive instances than in negative instances, which is necessary since negative instances were chosen randomly and therefore should not be co-regulated. Features were ranked by and the top d features were selected for decision tree building (d is denned by the user and is default to 10 in this study). Decision trees were built with C4.5 algorithm [26], except that instead of Igain were used.
Measuring model accuracy
Let TP, TN, FP, and FN be the numbers of true positive, true negative, false positive and false negative predictions made by a binary classifier, respectively, and N = TP + TN + FP + FN. Sensitivity is defined as . Specificity is defined as . The kappa static κ of the classifier is defined as , where is the percentage of correctly predicted instances, and C is the expected percentage of instances that a classifier can predict correctly by chance.
Note that when A = 100% and C ≠ 100%, κ = 1.0, corresponding to a perfect classifier; when A ≤ C, κ ≤ 0, meaning that the classifier does not perform better than random guessing.
GO functional enrichment
GO annotations were retrieved from SGD (version September 2004) [46]. Go functional enrichment were calculated with accumulative hyper-geometric distribution. GO::TermFinder perl module [64] was used to search for significantly enriched functional categories with a false discovery rate (FDR) < 0.05 [42].
Software and implementation
Motifs were identified with AlignACE [5] and scanned against promoter sequences with ScanACE [5]. Decision trees were built with the J48 program, which is a java implementation of the C4.5 decision tree learning algorithm [26], included in the WEKA machine learning package [21]. Decision trees were drawn with the dot program in Graphviz 1.0 [65] and displayed with webdot in the same package. Sequence logos of motifs were drawn with the seqlogo program [47]. The CAGER web service was implemented in perl and run on an apache web server with dual AMD Athlon 1600 MHz CPUs and 2 GB of RAM.
Authors' contributions
JR designed the experiments, performed the analysis, and prepared the manuscript. WZ provided essential guidance during the study and revised the manuscript critically. All authors read and approved the final manuscript.
List of abbreviations
TF: transcription factor. ChIP: Chromatin Immunoprecipitation. GO: gene ontology. ABA: abscisic acid. ABRE: ABA-responsive element. CE: coupling element. ORF: open reading frame. DEG: differentially expressed gene. SS: sensitivity. SP: specificity. FDR: false discovery rate. TP: true positive. FP: false positive. TN: true negative. FN: false negative.
Supplementary Material
Additional File 1
This Excel file contains Supplementary Table 1 that lists the cross-validation accuracies of decision tree models built for genes selected by different DEG identification methods.
Click here for file
Additional File 2
This Excel file contains Supplementary Table 2 that lists the cross-validation accuracies of decision tree models built with different types of features.
Click here for file
Acknowledgements
This research was supported in part by NSF grants IIS-0196057 and EIA-0113618 under the ITR program, and a grant from Monsanto Corporation. We thank two anonymous reviewers for their very useful comments. JR also wishes to thank Alexander V Loguinov for the EDGE software.
Figures and Tables
Figure 1 Effects of class distribution and feature selection on the accuracies of models. Models were built with ChIP-chip data for up- and down-regulated genes in yeast cell cycle. Each data point is an average of the kappa values of 10 × 77 = 770 models. (A). Effects of class distribution. X-axis ratio of the number of negative instances to the number of positive instances; y-axis: kappa in ten-fold cross-validations; (B). Effects of feature selection. X-axis: number of features selected; y-axis: kappa in ten-fold cross-validations.
Figure 2 The accuracies of models built for DEGs identified by different methods. The three groups of bars show the mean SS, 1 - SP and kappa values of the models built for genes selected by five different DEG identification methods. The error bars represent individual 95% confider interval for the means.
Figure 3 The accuracies of models built with different features. (A), (B) and (C) show kappa, SS and 1 - SP for models built with different features, respectively. ChIP: ChIP-chip data, pre: predefined motifs, auto: automatically identified motifs using AlignACE. The first cross-validation method was used for models in autol and the second cross-validation method was used for models in auto2 (see text). White bar: average accuracy of models for true DEGs. Grey bar: average accuracy of models for random genes. The error bars represent individual 95% confident interval for the means. (D) and (E) show kappa values for models built with different types of features for stress conditions and cell cycle conditions, respectively. X-axis: kappa for models built with predefined motifs; y-axis: kappa for models built with a combination of ChIP-chip data and predefined motifs.
Figure 4 Screenshots of the input and output interfaces of CAGER. (A). The input consists of four steps. First, the user provides ORF identifies or promoter sequences for positive and negative genes. Second, the user selects a type of features or some of their combinations. Third, the user can change parameters for decision tree learning. The user then provides an email address for notification of results and finally submits the job. (B). On the top of the output page is a graphical representation of a decision tree. Each oval of the decision tree represents an internal node and each box represents a leaf node. The text inside an internal node is the name of a feature, and the text associated with an edge is a test of the feature. The text inside a leaf node gives the predicted label (p: positive, n: negative) for genes inside the leaf, and the number of supporting and counter-instances, if any. A path from the root to a leaf node defines a possible regulatory rule. For example, the rightmost path can be read as "if the binding affinity of Mbpl to a gene's promoter is at least 1.75, the gene is positive (i.e., up-regulated under the condition that the decision tree models)." The numbers "16/1" enclosed in parenthesis means that 16 training instances have feature Mbpl_bind ≥ 1.75, of which 15 are positive and one is negative. On the bottom of the output page are related statistics for training and cross-validation of the decision tree model.
Figure 5 Decision tree and motifs learned for ABA-responsive genes in Arabidopsis. (A). A decision tree model learned for the transcriptional regulation of abscisic acid-induced genes in Arabidopsis. (B). Sequence logos of the three most significant motifs used by the decision tree model.
Table 1 Statistics of DEG sets identified by different methods. The ranges represent one standard deviation from the means.
EDGE global lowess vanilla vsn
Number of down-regulated sets 215 202 194 223 204
Number of down-regulated sets 215 210 200 223 195
Number of genes per set 48 ± 3 47 ± 8 45 ± 6 49 ± 7 46 ± 6
Table 2 Stability of decision tree models. Up: the models built for up-regulated genes. Down: the models built for down-regulated genes. Noise: the number of noisy instances added into the training set. TP: the number of true positive genes predicted by models built on the original data. TP': the number of true positive genes predicted by models built on the noisy data. Loss: the number of positive instances correctly classified in the original data but mis-classified in the noisy data. Rescue: the number of positive instances correctly classified in the noisy data but mis-classified originally. FP: the number of newly added noise genes classified as positive. Each value is an average across 223 up-regulated or 223 down-regulated gene sets. The standard errors for loss, rescue and FP are all less than 0.2.
Predefined motifs Auto motifs
Noise TP TP' Loss Rescue FP TP TP' Loss Rescue FP
Up 0 16.6 16.6 0.0 0.0 0.0 26.4 26.4 0.0 0.0 0.0
5 16.6 17.1 2.5 2.4 0.6 26.4 27.7 7.8 7.6 1.5
10 16.6 17.4 3.6 3.2 1.1 26.4 28.9 8.1 7.7 2.9
15 16.6 17.7 4.0 3.5 1.6 26.4 30.3 8.2 7.7 4.4
20 16.6 18.8 4.2 4.1 2.4 26.4 31.7 8.3 7.8 5.8
25 16.6 19.3 4.5 4.2 3.0 26.4 32.6 8.5 7.8 6.8
50 16.6 21.2 5.8 4.6 5.7 26.4 37.8 9.1 7.2 13.3
0 19.1 19.1 0.0 0.0 0.0 27.9 27.9 0.0 0.0 0.0
Down 5 19.1 19.6 2.0 2.0 0.5 27.9 29.1 7.3 7.0 1.5
10 19.1 20.5 2.7 2.9 1.1 27.9 29.6 8.0 6.8 2.9
15 19.1 20.7 3.2 3.2 1.6 27.9 30.6 8.1 6.6 4.1
20 19.1 20.9 3.8 3.5 2.1 27.9 30.9 8.8 6.5 5.3
25 19.1 21.2 4.1 3.6 2.6 27.9 32.8 8.5 6.6 6.9
50 19.1 22.6 5.3 3.6 5.2 27.9 37.8 9.7 6.2 13.4
Table 3 Functional enrichment in correctly predicted positive genes. Predefined motifs: the models built with predefined motifs. Auto motifs: the models built with auto motifs. Up: the models built for up-regulated genes. Down: the models built for down-regulated genes. #Gp: the number of Gp genes (positive genes that are predicted as positive by a model). #Gn: the number of Gn genes (positive genes that are predicted as negative by a model). : the number of enriched functional categories in the set of Gp genes. : the number of enriched functional categories in the set of Gn genes. : the number of enriched functional categories in the set of Gp' genes (see text). The sample sizes for up- and down-regulated genes in cell cycle are 56 and 59, respectively, and 167 and 164 in stress conditions. The ranges represent individual 95% confidence interval for the means.
Model #Gp #Gn
Predefined motifs Cell cycle Up 17.2 ± 2.2 27.1 ± 2.7 15.1 ± 2.9 8.2 ± 2.6 12.3 ± 3.5
Down 14.0 ± 2.1 31.0 ± 2.5 14.5 ± 3.7 9.6 ± 3.4 8.4 ± 3.3
Stress Up 16.4 ± 1.0 33.1 ± 1.1 13.4 ± 2.3 10.6 ± 1.8 10.3 ± 1.9
Down 21.0 ± 2.0 30.7 ± 1.7 17.1 ± 2.4 7.8 ± 1.7 14.7 ± 2.3
Auto motifs Cell cycle Up 24.9 ± 1.8 19.3 ± 1.8 16.9 ± 3.2 7.9 ± 2.5 13.9 ± 3.2
Down 23.6 ± 1.9 21.4 ± 1.8 15.6 ± 3.7 8.4 ± 2.9 11.8 ± 3.5
Stress Up 27.0 ± 0.8 22.6 ± 0.8 16.2 ± 2.3 9.1 ± 1.9 15.2 ± 2.4
Down 29.4 ± 1.3 22.2 ± 1.1 18.2 ± 2.3 6.6 ± 1.5 16.4 ± 2.4
==== Refs
Eisen M Spellman P Brown P Botstein D Cluster analysis and display of genome-wide expression patterns Proc Natl Acad Sci USA 1998 95 14863 8 9843981 10.1073/pnas.95.25.14863
Tavazoie S Hughes J Campbell M Cho R Church G Systematic determination of genetic network architecture Nat Genet 1999 22 281 5 10391217 10.1038/10343
Tamayo P Slonirn D Mesirov J Zhu Q Kitareewan S Dmitrovsky E Lander E Golub T Interpreting patterns of gene expression with self-organizing maps: methods and application to hematopoietic differentiation Proc Natl Acad Sci USA 1999 96 2907 12 10077610 10.1073/pnas.96.6.2907
Bailey T Elkan C Fitting a mixture model by expectation maximization to discover motifs in biopolymers Proc Int Conf Intell Syst Mol Biol 1994 2 28 36 7584402
Roth F Hughes J Estep P Church G Finding DNA regulatory motifs within unaligned noncoding sequences clustered by whole-genome mRNA quantitation Nat Biotechnol 1998 16 939 45 9788350 10.1038/nbt1098-939
van Helden J Andre B Collado-Vides J Extracting regulatory sites from the upstream region of yeast genes by computational analysis of oligonucleotide frequencies J Mol Biol 1998 281 827 42 9719638 10.1006/jmbi.1998.1947
Pilpel Y Sudarsanam P Church G Identifying regulatory networks by combinatorial analysis of promoter elements Nat Genet 2001 29 153 9 11547334 10.1038/ng724
Bussemaker H Li H Siggia E Regulatory element detection using correlation with expression Nat Genet 2001 27 167 71 11175784 10.1038/84792
Keles S van der Laan M Eisen M Identification of regulatory elements using a feature selection method Bioinformatics 2002 18 1167 75 12217908 10.1093/bioinformatics/18.9.1167
Conlon E Liu X Lieb J Liu J Integrating regulatory motif discovery and genome-wide expression analysis Proc Natl Acad Sci USA 2003 100 3339 44 12626739 10.1073/pnas.0630591100
Hu Y Sandmeyer S McLaughlin C Kibler D Combinatorial motif analysis and hypothesis generation on a genomic scale Bioinformatics 2000 16 222 32 10869015 10.1093/bioinformatics/16.3.222
Phuong T Lee D Lee K Regression trees for regulatory element identification Bioinformatics 2004 20 750 7 14751992 10.1093/bioinformatics/btg480
Middendorf M Kundaje A Wiggins C Freund Y Leslie C Predicting genetic regulatory response using classification Bioinformatics 2004 20 I232 I240 15262804 10.1093/bioinformatics/bth923
Simonis N Wodak S Cohen G van Helden J Combining pattern discovery and discriminant analysis to predict gene co-regulation Bioinformatics 2004 20 2370 9 15073004 10.1093/bioinformatics/bth252
Segal E Yelensky R Koller D Genome-wide discovery of transcriptional modules from DNA sequence and gene expression Bioinformatics 2003 19 i273 82 12855470 10.1093/bioinformatics/btg1038
Beer M Tavazoie S Predicting gene expression from sequence Cell 2004 117 185 98 15084257 10.1016/S0092-8674(04)00304-6
Lee T Rinaldi N Robert F Odom D Bar-Joseph Z Gerber G Hannett N Harbison C Thompson C Simon I Zeitlinger J Jennings E Murray H Gordon D Ren B Wyrick J Tagne J Volkert T Fraenkel E Gilford D Young R Transcriptional regulatory networks in Saccharomyces cerevisiae Science 2002 298 799 804 12399584 10.1126/science.1075090
Banerjee N Zhang M Identifying cooperativity among transcription factors controlling the cell cycle in yeast Nucleic Acids Res 2003 31 7024 31 14627835 10.1093/nar/gkg894
Gao F Foat B Bussemaker H Defining transcriptional networks through integrative modeling of mRNA expression and transcription factor binding data BMC Bioinformatics 2004 5 31 15113405 10.1186/1471-2105-5-31
Ruan J Zhang W Discovering transcriptional regulatory rules from gene expression and TF-DNA binding data by decision tree learning Tech Rep 43, Department of Computer Science and Engineering, Washington University in St Louis 2004
Witten I Frank E Data Mining: Practical Machine Learning Tools and Techniques with Java Implementations 1999 San Mateo, CA: Morgan Kaufmann
Roven C Bussemaker H REDUCE: An online tool for inferring cis-regulatory elements and transcriptional module activities from microarray data Nucleic Acids Res 2003 31 3487 90 12824350 10.1093/nar/gkg630
Stormo G DNA binding sites: representation and discovery Bioinformatics 2000 16 16 23 10812473 10.1093/bioinformatics/16.1.16
Buchler N Gerland U Hwa T On schemes of combinatorial transcription logic Proc Natl Acad Sci USA 2003 100 5136 41 12702751 10.1073/pnas.0930314100
Breiman L Friedman J Stone ROC Classification and Regression Trees 1984 Belmont, CA: Wadsworth Int. Group
Quinlan R C4-5: Programs for Machine Learning 1993 San Mateo, CA: Morgan Kaufmann
Murthy SK Automatic construction of decision trees from data: a multi-disciplinary survey Data Mining and Knowledge Discovery 1998 2 345 389 10.1023/A:1009744630224
Spellman P Sherlock G Zhang M Iyer V Anders K Eisen M Brown P Botstein D Futcher B Comprehensive identification of cell cycle-regulated genes of the yeast Saccharomyces cerevisiae by microarray hybridization Mol Biol Cell 1998 9 3273 97 9843569
Cho R Campbell M Winzeler E Steinmetz L Conway A Wodicka L Wolfsberg T Gabrielian A Landsman D Lockhart D Davis R A genome-wide transcriptional analysis of the mitotic cell cycle Mol Cell 1998 2 65 73 9702192 10.1016/S1097-2765(00)80114-8
Gasch A Spellman P Kao C Carmel-Harel O Eisen M Storz G Botstein D Brown P Genomic expression programs in the response of yeast cells to environmental changes Mol Biol Cell 2000 11 4241 57 11102521
Weiss G Provost F The effect of class distribution on classifier learning: an empirical study Tech Rep ML-TR-44, Department of Computer Science, Rutgers University 2001
Mewes H Amid C Arnold R Frishman D Guldener U Mannhaupt G Munsterkotter M Pagel P Strack N Stumpflen V Warfsmann J Ruepp A MIPS: analysis and annotation of proteins from whole genomes Nucleic Acids Res 2004 32 D41 4 14681354 10.1093/nar/gkh092
Kohavi R John G Wrappers for feature subset selection Artificial Intelligence 1997 97 273 324 10.1016/S0004-3702(97)00043-X
Jain A Zonker D Feature selection: Evaluation, application, and small sample performance IEEE Trans, on Pattern Analysis and Machine Intelligence 1997 19 153 158 10.1109/34.574797
Xing E Jordan M Karp R Feature selection for high-dimensional genomic microarray data Proceedings of the Eighteenth International Conference on Machine Learning 2001 601 608
Landis J Koch G The measurement of observer agreement for categorical data Biometrics 1977 33 159 74 843571
Altman DG Practical Statistics for Medical Research 1991 Chapman & Hall/CRC
Quackenbush J Microarray data normalization and transformation Nat Genet 2002 32 496 501 12454644 10.1038/ng1032
Huber W von Heydebreck A Sultmann H Poustka A Vingron M Variance stabilization applied to microarray data calibration and to the quantification of differential expression Bioinformatics 2002 18 S96 104 12169536
Yang Y Dudoit S Luu P Lin D Peng V Ngai J Speed T Normalization for cDNA microarray data: a robust composite method addressing single and multiple slide systematic variation Nucleic Acids Res 2002 30 el5 10.1093/nar/30.4.e15
Loguinov A Mian I Vulpe C Exploratory differential gene expression analysis in microarray experiments with no or limited replication Genome Biol 2004 5 R18 15003121 10.1186/gb-2004-5-3-r18
Benjamini Y Hochberg Y Controlling the false discovery rate: a practical and powerful approach to multiple testing Journal of the Royal Statistical Society 1995 57 289 300
Supplementary Website
Harbison C Gordon D Lee T Rinaldi N Macisaac K Danford T Hannett N Tagne J Reynolds D Yoo J Jennings E Zeitlinger J Pokholok D Kellis M Rolfe P Takusagawa K Lander E Gifford D Fraenkel E Young R Transcriptional regulatory code of a eukaryotic genome Nature 2004 431 99 104 15343339 10.1038/nature02800
Harris M Clark J Ireland A Lomax J Ashburner M Foulger R Eilbeck K Lewis S Marshall B Mungall C Richter J Rubin G Blake J Bult C Dolan M Drabkin H Eppig J Hill D Ni L Ringwald M Balakrishnan R Cherry J Christie K Costanzo M Dwight S Engel S Fisk D Hirschman J Hong E Nash R Sethuraman A Theesfeld C Botstein D Dolinski K Feierbach B Berardini T Mundodi S Rhee S Apweiler R Barrell D Camon E Dimmer E Lee V Chisholm R Gaudet P Kibbe W Kishore R Schwarz E Sternberg P Gwinn M Hannick L Wortman J Berriman M Wood V de la Cruz N Tonellato P Jaiswal P Seigfried T White R The Gene Ontology (GO) database and informatics resource Nucleic Acids Res 2004 32 D258 61 14681407
Dwight S Balakrishnan R Christie K Costanzo M Dolinski K Engel S Feierbach B Fisk D Hirschman J Hong E Issel-Tarver L Nash R Sethuraman A Starr B Theesfeld C Andrada R Binkley G Dong Q Lane C Schroeder M Weng S Botstein D Cherry J Saccharomyces genome database: underlying principles and organisation Brief Bioinform 2004 5 9 22 15153302 10.1186/1471-2105-5-9
Crooks G Hon G Chandonia J Brenner S WebLogo: a sequence logo generator Genome Res 2004 14 1188 90 15173120 10.1101/gr.849004
Leung J Giraudat J Abscisic acid signal transduction Annu Rev Plant Physiol Plant Mol Biol 1998 49 199 222 15012233 10.1146/annurev.arplant.49.1.199
Seki M Ishida J Narusaka M Fujita M Nanjo T Umezawa T Kamiya A Nakajima M Enju A Sakurai T Satou M Akiyama K Yamaguchi-Shinozaki K Carninci P Kawai J Hayashizaki Y Shinozaki K Monitoring the expression pattern of around 7,000 Arabidopsis genes under ABA treatments using a full-length cDNA microarray Fund Integr Genomics 2002 2 282 91 10.1007/s10142-002-0070-6
Ono A Izawa T Chua N Shimamoto K The rab16B promoter of rice contains two distinct abscis acid-responsive elements Plant Physiol 1996 112 483 91 8883374 10.1104/pp.112.2.483
Hattori T Totsuka M Hobo T Kagaya Y Yamamoto-Toyoda A Experimentally determined sequenc requirement of ACGT-containing abscisic acid response element Plant Cell Physiol 2002 43 136 40 11828032 10.1093/pcp/pcf014
Garcia-Garrido J Menossi M Puigdomenech P Martinez-Izquierdo J Delseny M Characterization of a gene encoding an abscisic acid-inducible type-2 lipid transfer protein from rice FEBS Lett 1998 428 193 9 9654133 10.1016/S0014-5793(98)00529-8
Busk P Pages M Protein binding to the abscisic acid-responsive element is independent of VIVIPAROUS1 in vivo Plant Cell 1997 9 2261 70 11407411 10.1105/tpc.9.12.2261
Busk P Jensen A Pages M Regulatory elements in vivo in the promoter of the abscisic acid responsive gene rab17 from maize Plant J 1997 11 1285 95 9225468 10.1046/j.1365-313X.1997.11061285.x
Shen Q Zhang P Ho T Modular nature of abscisic acid (ABA) response complexes: composite promoter units that are necessary and sufficient for ABA induction of gene expression in barley Plant Cell 1996 8 1107 19 8768371 10.1105/tpc.8.7.1107
Guiltinan M WR JM Quatrano R A plant leucine zipper protein that recognizes an abscisic ack response element Science 1990 250 267 71 2145628
Hobo T Asada M Kowyama Y Hattori T ACGT-containing abscisic acid response element (ABRE) and coupling element 3 (CE3) are functionally equivalent Plant J 1999 19 679 89 10571853 10.1046/j.1365-313x.1999.00565.x
Higo K Ugawa Y Iwamoto M Korenaga T Plant cis-acting regulatory DNA elements (PLACE) database Nucleic Acids Res 1999 27 297 300 9847208 10.1093/nar/27.1.297
Expression Connection
Saccharomyces cerevisiae ChlP-chip data
Known and putative regulatory motifs
van Helden J Regulatory sequence analysis tools Nucleic Acids Res 2003 31 3593 6 12824373 10.1093/nar/gkg567
Solovyev V DB Statistical approaches in Eukaryotic gene prediction Handbook of Statistical genetic 2001 John Wiley and Sons 83 127
Boyle E Weng S Gollub J Jin H Botstein D Cherry J Sherlock G GO::TermFinder – open source software for accessing Gene Ontology information and finding significantly enriched Gene Ontology terms associated with a list of genes Bioinformatics 2004 20 3710 3715 D258-61 15297299 10.1093/bioinformatics/bth123
Graphviz – Graph Visualization Software
| 15890068 | PMC1174863 | CC BY | 2021-01-04 16:02:50 | no | BMC Bioinformatics. 2005 May 12; 6:114 | utf-8 | BMC Bioinformatics | 2,005 | 10.1186/1471-2105-6-114 | oa_comm |
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-1311592462610.1186/1471-2105-6-131Research ArticleGene finding in the chicken genome Eyras Eduardo [email protected] Alexandre [email protected] Robert [email protected] Jacqueline M [email protected] Francisco [email protected] Paul [email protected] Elizabeth J [email protected] Genis [email protected] David D [email protected] Carine [email protected] Jane [email protected] Stylianos E [email protected] Ewan [email protected] Roderic [email protected] Michael R [email protected] Research Group in Biomedical Informatics, Institut Municipal d'Investigacio Medica/Universitat Pompeu Fabra/Centre de Regulacio Genomica, E08003 Barcelona, Catalonia, Spain2 Department of Genetic Medicine and Development, University of Geneva, Medical School and University Hospital of Geneva, CMU, 1, rue Michel Servet, 1211 Geneva, Switzerland3 Center for Integrative Genomics, University of Lausanne, 1015 Lausanne, Switzerland4 The Wellcome Trust Sanger Institute, Wellcome Trust Genome Campus, Hinxton, Cambridge CB10 1SA, UK5 Laboratory for Computational Genomics and Department of Computer Science, Campus Box 1045, Washington University, One Brookings Drive, St Louis, Missouri 63130, USA6 EBI, Wellcome Trust Genome Campus, Hinxton, Cambridge CB10 1SD, UK2005 30 5 2005 6 131 131 26 10 2004 30 5 2005 Copyright © 2005 Eyras et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Despite the continuous production of genome sequence for a number of organisms, reliable, comprehensive, and cost effective gene prediction remains problematic. This is particularly true for genomes for which there is not a large collection of known gene sequences, such as the recently published chicken genome. We used the chicken sequence to test comparative and homology-based gene-finding methods followed by experimental validation as an effective genome annotation method.
Results
We performed experimental evaluation by RT-PCR of three different computational gene finders, Ensembl, SGP2 and TWINSCAN, applied to the chicken genome. A Venn diagram was computed and each component of it was evaluated. The results showed that de novo comparative methods can identify up to about 700 chicken genes with no previous evidence of expression, and can correctly extend about 40% of homology-based predictions at the 5' end.
Conclusions
De novo comparative gene prediction followed by experimental verification is effective at enhancing the annotation of the newly sequenced genomes provided by standard homology-based methods.
==== Body
Background
The draft sequence of the chicken (Gallus gallus) genome has been recently obtained and an initial analysis completed [1]. This genome sequence will be extremely valuable for vertebrate development, avian biology, and agriculture. To reap these benefits, however, we must be able to determine with reasonable accuracy the amino acid sequences of the proteins encoded by this genome. We therefore set out to adapt three leading gene prediction systems to the chicken genome sequence and to evaluate their predictions experimentally. Some features of the chicken genome facilitate gene prediction while others hinder it. Generally, gene prediction tends to be more accurate in more compact genomes than in the large mammalian genomes [2-6]. The chicken genome is about 40% the size of the human genome, but about three times the size of the Takifugu genome [7-11]. This translates into a substantial reduction in repeats and pseudogenes. Interspersed repeats cover about 9% of the genome, far less than the 40–50% found in mammals [1]. The small number of processed pseudogenes is beneficial to automated annotation, as this technology often misclassifies pseudogenes as functional. De novo prediction methods need full open reading frame (ORF) mRNAs to train their statistical models and homology-based methods rely on these sequences in order to predict a core set of high quality genes. However, an abundant set of chicken full-length cDNAs was not available at the start of this project, as only about 1,800 putatively full-length cDNAs and 340,000 ESTs were deposited in GenBank [12].
We set out to determine how well the computational methods used for annotating protein-coding genes in the mouse and rat genomes [9,10] would perform in this avian context, where transcriptome sequencing was much less advanced and where sequence divergence may be beyond the optimal distance. An assessment of the computational methods was carried out by a large-scale experimental verification by RT-PCR. The gene prediction tools tested were Ensembl, SGP2 and TWINSCAN. Ensembl is a homology-based method, which builds gene-models using species-specific known sequences and proteins from other species aligned to the genome [13]. SGP2 and TWINSCAN are de novo comparative gene predictors whose only inputs are the genome to be annotated and a second informant genome [3,4,14-16]. In this case, the informant genome selected for comparison was human. Since Ensembl relies on mapping known genes from chicken and other organisms to the chicken genome, we surmised that its prediction set would contain fewer false positives (predicted genes that are not real) and more accurate gene structures, as well as more false negatives (real genes that are not predicted) than the de novo methods. Conversely, SGP2 and TWINSCAN should be effective in detecting bona fide genes missed by the Ensembl pipeline. In particular, we were able to verify gene structures simultaneously predicted by both de novo methods about 50% of the time. In this respect, de novo comparative methods complement homology-based methods, which in general miss genes for which there is no pre-existing evidence of transcription.
To test and compare the performance of these prediction systems under the new conditions presented by the chicken genome, we tested a large number of predicted genes by RT-PCR and direct sequencing. We sampled individually from the predictions for which either one or two of the three methods agreed, and the remainder did not. We thus focused our tests on the most difficult genes to predict and on the differences between the various prediction systems. We present here the largest experimental comparison of multiple gene-prediction programs to date as well as the first to use this kind of differential design. We aimed to evaluate how de novo and homology-based gene finding methods perform in a newly sequenced genome for which a small number of gene sequences are known. In particular we have (1) evaluated the three prediction methods, (2) investigated TWINSCAN and SGP2 as possible effective methods to complement the Ensembl prediction pipeline and (3) tested the overall specificity of the Ensembl prediction set.
Results and discussion
Experimental evaluation of gene finding in chicken
We aimed to estimate the accuracy of each of the gene prediction sets, which consisted of 29,430 SGP2 (S) and 29,052 TWINSCAN (T) gene predictions, with one transcript per gene, and 17,709 Ensembl (E) genes containing 28,416 transcripts. We classified the predictions according to the Venn diagram defined by the three-way intersections of the sets and their complements (Figure 1A, Table 1). The subsets were populated with intron assemblies (IAs), defined as a list of exons and introns contiguous in a predicted transcript (see Figure 2 and Methods for details), and can be classified into three types: (1) the orphan subsets, containing those elements that are in one set but not in any of the other two, (2) the two-way intersection subsets, containing those elements that are in two sets and not in the third, and (3) the triple intersection, containing those elements that are in all three sets. Subsequently we tested pairs of adjacent exons from each of the subsets.
Complementing homology-based gene prediction with de novo methods
Selected IAs belonging to the three possible two-way intersections were experimentally tested. The results are summarized in Figure 1 and detailed in Table 2. After RT-PCR, gel purification, and direct sequencing, about 50% of the tested transcripts predicted by both SGP2 and TWINSCAN, but not by Ensembl yielded spliced alignments to the gene targeted (Figure 1C). This rate is higher than that reported for human using a combination of homology-based and single-genome predictors [16,17] in spite of the lack of available known gene sequences for the chicken genome. There is a total of 4,769 IAs in the '(S and T) not E' subset, corresponding to a total of 13,470 exons. Projecting these results onto genes and using an average distribution of coding exons per gene from other vertebrates (Human, Mouse and Rat), we estimate that approximately 740 to 840 bona fide chicken genes that are not in the currently predicted Ensembl set can be found by the de novo comparative methods followed by direct amplification and sequencing.
Considering the set of IAs unique to one prediction set (orphans), 39% of these have one single intron and 80% have 1 or 2 introns. On the other hand, after testing experimentally 96 orphans from TWINSCAN, 88 orphans from SGP2 and 30 orphans from Ensembl, we found that about 77% of the Ensembl orphans are real genes, compared to an average 18% for TWINSCAN and SGP2 (see Figure 1, see Table 2). Thus while Ensembl orphans are more likely to be real genes not predicted by the other methods, orphan de novo predictions are more likely to be false positives.
Extending homology-based gene predictions with de novo methods
As Ensembl predictions often fail to correctly predict one or both ends of a gene [13], we reasoned that de novo prediction methods could help in extending the homology-based predicted transcripts. To test this hypothesis, we identified candidate 5' extensions: exons predicted by TWINSCAN, SGP2 or both to the 5' side of Ensembl genes on the same strand. We found that 8,368 (47%) Ensembl genes have such candidate extensions. However, not all these extensions were as likely to correspond to real exons. From this total, 7,630 genes had extensions suggested by de novo predictions that overlap the Ensembl gene (linked, see Figure 3A), and 738 (4%) had extra exons from de novo predictions that did not overlap the Ensembl gene (unlinked, see Figure 3B). As 99% of Ensembl introns were no longer than 100 kb, we considered only exon extensions that were no further than 100 kb from the 5'-most Ensembl exon. Interestingly, we found that 93% of the linked extensions were to multiexonic Ensembl transcripts, the remainder being extensions to single-exon Ensembl transcripts; however, for unlinked extensions, only 58% were to multiexonic Ensembl transcripts.
We investigated experimentally 60 linked and 29 unlinked extensions by designing one primer in the 5'-most exon of the Ensembl prediction and the second primer within one of the upstream exon suggested by TWINSCAN and/or SGP2 to the 5' side. The RT-PCR results (see Table 3) show that de novo methods-suggested linked extensions were correct in about 40% of the cases. This rate dropped to a mere 7% for the unlinked extensions.
Separating the extensions according to whether the extra exon was predicted either by SGP2 or by TWINSCAN showed that both methods had a comparable contribution. From the 60 linked tested extensions, 46 were predicted by SGP2 with 21 (46%) RT-PCR positives, 36 were predicted by TWINSCAN with 15 (42%) positives, and 22 were predicted by both programs with 12 (54%) positives. On the other hand, from the 29 tested unlinked extensions 15 were predicted by SGP2, with 2 positives, and 17 were predicted by TWINSCAN, with 1 positive. Finally, there were 3 cases where both, SGP2 and TWINSCAN, predicted the unlinked extra exon of which one was positive.
Testing Ensembl specificity
A randomly selected set of Ensembl predictions was assayed to evaluate Ensembl's specificity. This test measured a false positive rate of 4% (see Table S5 of the supplementary material for more details). On the other hand, the tested exon-pairs for the two-way intersection sets that included Ensembl ('(E and S) not T' and '(T and E) not S') had an average false positive rate of 19% and 35%, respectively (see Figure 1). The disparity is greater with the two-way intersection set that excludes Ensembl ('(S and T) not E'), which shows a false positive rate of 53%. One explanation for this difference is the observation that most of the Ensembl predictions have exons predicted by both SGP2 and TWINSCAN. Indeed, 25,222 (89%) of the 28,416 Ensembl transcripts have at least one exon, which is also in the SGP2 and TWINSCAN sets, and 82% of these transcripts have 2 or more exons in common with both de novo methods. Thus, Ensembl predictions are most likely to fall within a triple intersection, resulting in an increased rate of true positives. Based on previous experiments [16], we expected the triple intersection to give a yield close to 100% positive rate. We tested 20 cases of triple intersection and found 85% positive rate (see Figure 1C). Moreover, 94% of these positive cases had the exon-intron boundaries correctly predicted (see Figure 1D).
We compared the accuracy of the human-based predictions with the accuracy of a fish-based set of predictions. We predicted genes in chicken with SGP2 and TWINSCAN using Tetraodon nigroviridis as informant genome, and found that they are less accurate than the human-based ones (see Table S7 of the supplementary material). Additionally, we found that 85% and 98% of the Tetraodon-based predictions from TWINSCAN and SGP2, respectively, overlap the corresponding human-based ones. Interestingly, 85% of the TWINSCAN orphans overlap TWINSCAN Tetraodon-based predictions, and 80% of the SGP2 orphans overlap SGP2 Tetraodon-based predictions. On the other hand, 99% of the IAs common to TWINSCAN and SGP2 but not in Ensembl, overlap TWINSCAN or SGP2 Tetraodon-based predictions.
Conclusions
In this paper we have evaluated how effective purely computational approaches for genome annotation can be, even in the absence of a large collection of previously known genes, by means of the largest attempt so far to experimentally compare several gene finders. After testing the accuracy of Ensembl, SGP2 and TWINSCAN on the chicken genome we have shown that de novo comparative methods followed by experimental verification remain a successful approach in the annotation of newly sequenced genomes from which little is known.
We found that approximately 50% of predictions that were in TWINSCAN and SGP2 but not in Ensembl could be experimentally verified (Figure 1). These experiments demonstrate that de novo comparative prediction methods are effective at complementing homology-based methods and confirm that a combination of methods can improve the prediction accuracy [18-22]. Moreover, in spite of the limited gene sequence data available for chicken, the combination of TWINSCAN and SGP2 achieves better accuracy than previous attempts to verify by RT-PCR computational predictions that fall outside a set of annotations [17,23]. On the other hand, looking at the intron assemblies unique to one prediction set, the proportion of positives is largely reduced for predictions not in Ensembl. The predictions unique to one of the de novo methods show an abundance of gene models with 2 and 3 exons, which may be artefacts due to genome misassemblies. These results are in contrast with the high success rate (77%) of the predictions unique to Ensembl. This is a reasonable observation considering that the Ensembl prediction pipeline has access to genes that do not follow a 'standard' gene-grammar (e.g., unusual codon usage), but which may nevertheless be represented in the cDNA/protein databases used.
The Ensembl chicken gene set has been found to have a 96% positive rate, whereas the IAs from the two-way intersections that include Ensembl, '(E and S) not T' and '(T and E) not S', and the Ensembl orphans, have a lower positive rate, 81%, 65% and 77%, respectively, which stems from the fact that most exons predicted by Ensembl are also predicted by both SGP2 and TWINSCAN. Additionally, de novo comparative methods are useful for extending partial predictions from homology-based methods. Ensembl may generate predictions based on protein fragments or on partial homology from other species, and TWINSCAN and SGP2 predictions can add bona fide exons to the Ensembl predictions they overlap with. For the 5' end we show that 40% of the tested cases, where either TWINSCAN or SGP2 predicted at least one additional exon, were verified (Table 3). To our knowledge, this is the first time that experimental evidence is provided for extensions to homology-based models produced by de novo methods.
We observed that the subsets containing SGP2 IAs (e.g., '(S and E) not T)') have in general a higher proportion of RT-PCR positives than those containing TWINSCAN IAs (e.g., '(T and E) not S)') (Figure 1C, Table 2). There are two factors that may contribute to this difference. The first is an intrinsic difference between TWINSCAN and SGP2 – SGP2 uses TBLASTX (translated) alignments between human and chicken to reward exons overlapping aligned regions, whereas TWINSCAN uses BLASTN (nucleotide) alignments to influence the scores of exons, splice sites, and translation initiation and termination sites. Human and chicken are sufficiently diverged that translated alignments may be more sensitive, whereas nucleotide alignments fail to cover many known exons. The other factor is incidental to the way TWINSCAN was trained and run to produce the predictions tested. TWINSCAN used 525 chicken RefSeqs to estimate parameters for its probability model. This training set was probably too small to produce optimal parameter values. SGP2, on the other hand, was run with a combination of parameters estimated from the much larger set of known human genes (for its model of chicken DNA sequence) and parameters were hand tuned using the same 525 chicken genes (for its scoring of human-chicken alignments). Although a larger fraction of SGP2 predictions yielded positive experimental results, we found that TWINSCAN tends to be more accurate than SGP2 in the prediction of the intron boundaries (Figure 1D, Table 2). This difference stems from the intron model used by TWINSCAN, as opposed to SGP2, which does not model introns explicitly. TWINSCAN was re-run after completion of the experiments with an improved intron-length model, yielding a prediction set that was substantially smaller and more accurate (see Table S6) than the set tested. In spite these differences, comparing the gene predictions with a set of coding cDNAs released after the completion of these analyses, we found that all three methods have similar sensitivity (79%) (see Methods for details), hence the de novo comparative methods cover a fraction of the transcriptome similar to homology-based methods with a minimal initial amount of genome-specific expression data.
The experimentally verified IAs represent a fraction of the actual number of chicken genes that can be eventually found using our methods. If we extrapolate the proportions of experimentally verified IAs (Figure 1C) to all the generated IAs in the Venn diagram (Figure 1B) and using an average distribution of coding exons per gene from other vertebrates (Human, Mouse and Rat), we estimate a range of 14,600 to 17,500 experimentally verifiable chicken genes from our computational predictions. In this paper we only analysed intron assemblies and deliberately left out a number of chicken protein-coding intronless predictions (3,049 from SGP2, 2,727 from TWINSCAN and 1,855 from Ensembl). The triple intersection of these intronless genes contains 148 genes, which are worth investigating and for which techniques different from the ones applied here will be required.
Considering all 2-way IAs (see Table 1), one would need 11,274 RT-PCR reactions to experimentally confirm about 7,232 (64%) genes. This number of experiments compares favourably to large-scale EST projects with the added benefit of having almost no redundancy (only gene fissions and misassemblies will contribute to redundancy). The biggest drawback to EST sequencing is its large redundancy and extensive overlap. The falling cost of primers and the increased flexibility of large-scale molecular biology centers make this approach of computational prediction followed by experimental verification cost effective and scalable [5,6]. As RT-PCR primers can be designed with appropriate linker sites such an approach could also provide a physical resource of clonable fragments. We conclude that de novo comparative gene predictions followed by experimental verification is an effective way to carry out the annotation of a newly sequenced genome for which little gene sequence information is known. In particular, as our results show, performing RT-PCR and sequencing for all the predicted novel genes, starting with those predicted by multiple de novo methods, should enhance the quality of the annotation in forthcoming eukaryote genome sequencing projects.
Methods
Generation of predictions
The initial lack of an abundant set of known chicken gene sequences forced us to adapt the methodology for training of the de novo methods and running the Ensembl gene build. The Ensembl prediction pipeline [13] builds gene models from known vertebrate proteins and cDNA sequences, whereas gene models based solely on ESTs are usually kept separated as ESTGenes [24]. For the chicken genome, ESTs were combined with the standard gene build to include additional genes and transcripts. The alignments from approximately 400,000 ESTs and 24,000 full-length cDNAs [12,25] from multiple tissues were conciliated into non-redundant transcript structures [24]. Those predicted models that fell on non-annotated loci and those that contributed with at least two new exons in previously annotated loci were added to the protein-based gene set. Single-exons transcripts produced by the EST/cDNA-based models were rejected, as they could not be distinguished from cloning artefacts without protein evidence. For our analyses we did not include the untranslated regions that Ensembl annotates combining the cDNA/EST and protein evidence [13]. Ensembl also annotates processed pseudogenes [13], which we did not consider either.
TWINSCAN was trained on 525 chicken RefSeq sequences [26] aligned to the chicken genomic sequence. This set is based on a set of 1266 provisional RefSeq mRNAs from GenBank (27 March 2004) March 27 filtered in the following way: The Refseq mRNA sequences were matched with the corresponding mRNAs placed on the Chicken genome assembly at UCSC. Any mRNA not placed was not considered. Additionally, sequences without an ungapped alignment between the entire CDS portion of the Refseq mRNA and the extracted unmasked genomic sequence were removed. Additionally, cases with in-frame stop codons and/or non-canonical splice sites were also removed. TWINSCAN uses nucleotide alignments and has specific models for how the alignments modify the scores of the gene signals. The BLASTN [27] alignments used by TWINSCAN covered 3.8% of the chicken assembly.
SGP2 training followed a hybrid approach. SGP2 was run with human parameters for the coding statistics and splice sites, whereas the score weights that reward the human-chicken homologies and penalize the lack of them were optimised using the same set of 525 Chicken RefSeqs used for the TWINSCAN training. SGP2 was then run on unsegmented chicken chromosomes using the TBLASTX [27] alignments with the human genomic sequence (assembly NCBI34). These alignments, which covered approximately 3% of the Chicken genome, were enriched with 391,610 extra HSPs obtained from the ungapped Exonerate [28] alignments of human proteins from the Ensembl predictions (release NCBI34c), the GeneId prediction set for the same human assembly and the set of vertebrate RefSeq proteins (version of April 2004). The extra alignments covered 43% of the nucleotides in TBLASTX HSPs and 5% of their sequence represented 5,840 non-redundant homology regions that had no overlap with the TBLASTX hits. These extra alignments produced a considerable improvement of the sensitivity and specificity at the gene level with respect to SGP2 predictions using only TBLASTX HSPs when tested against the Ensembl set and the aforementioned 525 RefSeqs. It also achieved a slight improvement of the sensitivity at the exon and nucleotide level.
Classification of predictions
We classified the predictions according to the Venn diagram defined by the intersections of the three sets: Ensembl (E), TWINSCAN (T) and SGP2 (S). Using the exact identity between transcripts is problematic because the exon-structures from each prediction are in general not identical. Thus we defined the intron assembly (IA) as the element of comparison. An IA is a list of exons and introns that are contiguous in a given predicted transcript. When comparing two predictions three differentiated sets are produced: the set of IAs that are identical in both transcripts, and the two sets of IAs that are present in one prediction and not in the other (see Figure 2). If the boundaries of the exons do not agree between two predictions the produced IA contains only the exonic sequence that is common to both predictions, i.e., an intersecting IA contains the sequence and exon-structure on which both predictors agree (see Figure 2). An intron assembly represents naturally the entities to be tested by RT-PCR, where an exon-pair, separated by one or more introns and by not more than about 1 Kb, is tested for amplification in cDNA tissue libraries. As each IA contains the maximal set of introns in a given genomic locus that fall in one of the different categories of the Venn diagram, we picked up one exon-pair from each IA, making sure to test different genes in each Venn diagram.
Combining the operations of set-difference 'not', set-intersection 'and', and set-union '+', we populated the Venn diagram. For instance: the subset 'E not (T+S)' was generated by first obtaining the union of T and S predictions (T+S) and then comparing the Ensembl predictions with this latter set, hence 'E not (T+S)' contains Ensembl IAs that are not in TWINSCAN or SGP2. Similarly, we obtained the subset '(S and T) not E' by first calculating the intersection of S and T, and then calculating the set-difference against E. Note that the set-difference is non-commutative, as we keep elements from one set and use the other for comparison, hence S not T ? T not S. These operations divide the subsets into three types: (1) orphan subsets, formed by those elements that are in one set but not in any of the other two, (2) the two-way intersection subsets, formed by those elements that are in two sets and not in the third one, and (3) the triple intersection, formed by those elements that are in the three sets. Figure 1B shows the number of IAs in each of the subsets (see also Table 1).
When considering the Ensembl predictions we introduced a slight modification of the operations, as one Ensembl gene may have more than one transcript (see Figure 2b). We defined the intersecting IAs as the longest non-redundant common IA between the transcripts from either prediction. That is, from all the redundant intron assemblies in an Ensembl gene that are also in a prediction from SGP2 or TWINSCAN, we took the longest one. By the concept of redundancy of two intron assemblies we mean that they have the same splicing structure or one is included in the structure of the other, allowing for mismatches in the exon edges. Likewise, for the set difference involving Ensembl predictions we took the longest IA from all the ones in an Ensembl gene that were novel with respect to the de novo prediction. The inverse case works similarly: an IA in TWINSCAN or SGP2 which is not in an Ensembl gene is the longest IA that is novel with respect to all the Ensembl transcripts in that gene.
We classified IAs according to their position relative to the predictions from the other set against which they are compared. We considered an IA to be:
• Intergenic: if it falls outside the genomic extension of any of the predictions from the other set.
• Bridge: if it bridges between two different predictions in the other set.
• Intronic: if it extends one or more introns of the excluded prediction, i.e., the exons of the IA fall in the introns of the other prediction.
• External: if it extends the 5' or 3' of the other prediction.
In Table S1 we give the distribution of the IAs according to their relative position (see Figure 4). We observe that while SGP2 produces approximately the same proportion of external and intronic orphan IAs, TWINSCAN produces many more intronic ones.
Finally, IAs were further classified according to whether they are complete, i.e., they represent a complete ATG-to-STOP prediction in at least one set (Figure 4). Table S2 presents the number of complete IAs from each set. TWINSCAN predictions produced more complete novel genes within introns than SGP2, whereas SGP2 predicts more complete novel genes that extend the genomic span of other genes. Furthermore, more than 70% of the complete intergenic orphan IAs from all programs have either two or three exons. Finally, analysing the triple intersection of the prediction sets we find 10,650 IAs that are common to the three sets from which 8,837 (83%) are complete. In contrast to the complete IAs predicted by only one program, 55% of those predicted by all programs have more than 3 exons.
Comparison to chicken expression data
We compared the prediction sets against 13,880 chicken cDNAs annotated as coding for protein and 4,154 cDNAs annotated as non-coding [12,25]. We used Exonerate in ungapped mode to directly compare the nucleotide sequence of the gene predictions against the cDNA set. We considered a cDNA to be 'found' if there was an alignment with more than 90% identity over more than 60 bp. The results showed that Ensembl and SGP2 included 79% and TWINSCAN included 77% of the coding cDNAs, whereas they included a 17%, 20% and 30% of non-coding cDNAs, respectively. Additionally, we found that 8,126 (28%) SGP2 genes, 7,037 (40%) Ensembl genes and 7,728 (27%) TWINSCAN genes aligned against the coding cDNAs; whereas 838 (3%) SGP2 genes, 1,186 (7%) Ensembl genes and 722 (2.5%) TWINSCAN genes aligned against the cDNAs labelled as non-coding. The three prediction sets together found a total of 12,009 (86.5%) of coding cDNAs, with a common set of 9,597 (69%) coding cDNAs found by all three.
Primer design
Primers used in the PCR reactions were designed using primer3 and filtered using ePCR. For each IA in each tested set, two exons were selected for primer placement. This selection was made ensuring that the exons selected for primer placement were of adequate length to generate likely primers, and that there was a sufficient amount of coding sequence to give an amplicon length in the desired range. Additional checks were performed in cases where one prediction method included a coding exon not suggested by the other two prediction methods in the same transcript. In those cases, the length of the tentative exons was accounted for by what exons were selected for primer placement, helping reduce failure due to primers being placed too far apart in the actual transcript to get successful amplification. All designed primers and primer pairs were filtered for mispriming using the entire chicken genome to find potential priming locations. Primers returning priming locations not overlapping coordinates of the target gene were rejected. From all passing primer pairs for a given IA, the pair giving the longest expected product in the requested range was selected for amplification. The selection of targets appearing on each plate was done completely randomly among all targets returning at least one designed primer pair passing all criteria. The primer selection parameters used for the design of Plate 1 were consistent with primer3 defaults, except: Tm range was set to (63, 65), product length range was set to (300, 600), and primer length range was set to (23, 28) with 24 being optimal. The primer selection parameters used for the design of Plates 2–5 were also consistent with primer3 defaults, except: GC content range was set to (30, 70) with 50 being optimal, Tm range was set to (59, 62), product length range was set to (150, 500), and primer length range was set to (17, 27) with 20 being optimal.
Experimental verification of predictions by RT-PCR
cDNA preparation
Multiple organs (brain, liver, heart, spleen, lung, kidney, muscle, tongue, trachea, crop, proventriculus, gizzard, gall bladder, small-intestine, pancreas, caeca, mesentery, ovary, oviduct and testis) of an adult male and two egg-laying females of the "Bleue de Hollande" strain were collected soon after sacrifice. Total RNA was prepared from frozen tissues using TRIzol Reagent (Invitrogen) according to manufacturers' instructions. The quality of all RNA samples was checked using an Agilent 2100 Bioanalyzer (Agilent Technologies) and by PCR using pairs of oligos designed in four CNGs (Conserved Non-Genic Sequences) conserved between GGA1 and HSA21 [29,30], as indicators of possible genomic DNA contamination. Total RNA was converted to cDNA using Superscript II (Invitrogen) primed with random primers. For each tissue in the study, 5 μg of total RNA was converted to cDNA.
Experimental verification
Predictions of chicken genes were assayed experimentally by RT-PCR as previously described and modified [9,16,31]. Similar amounts of 12 Gallus gallus cDNAs (brain, liver, heart, spleen, lung, kidney, muscle, proventriculus, small intestine, caeca, ovary and testis, final dilution 1000x) were mixed with JumpStart REDTaq ReadyMix (Sigma) and 4 ng/ul primers (Sigma-Genosys) with a BioMek 2000 robot (Beckman). The ten first cycles of PCR amplification were performed with a touchdown annealing temperatures decreasing from 60 to 50°C; annealing temperature of the next 30 cycles was carried out at 50°C. Amplimers were separated on "Ready to Run" precast gels (Pharmacia) and sequenced. This procedure was used to experimentally assay 456 exon-exon junctions of chicken predictions. The later are representative of each subsets of predictions found in the Venn diagram of the three sets of predictions studied, i.e., the Ensembl, TWINSCAN and SGP2 set (see Table S3, and Table S4 for the transcripts used as internal controls). The sequences of the amplified exon-exon junctions can be obtained from the web site .
Authors' contributions
EE with help from AR and MRB led the writing and preparation of the manuscript. RC, EB, SEA, PF, DDS and RG contributed to the text and overall layout. MRB conceived the Venn diagram approach for the experiments in chicken and chose the target sets, with input from DDS and EE. EE generated the chicken intron assemblies and analysed the experimental results. RC, PF, EE, GP, FC generated the gene predictions and performed several analyses on them. The generation of primers was carried out by DDS, MRB and EB for chicken. AR and CW collected the tissues and prepared the RNAs/cDNAs. EJH, JMB, CW carried out the RT-PCR experiments with supervision from AR, JR, EB and SEA. The analysis of the chicken predictions was coordinated by RG, MRB and EB.
Supplementary Material
Additional File 1
Supplementary information: distribution of intron-assemblies according to relative positions; details on the RT-PCR test and control sets; accuracy evaluation of the human-based and tetraodon-based chicken gene predictions.
Click here for file
Acknowledgements
We thank J.F. Abril, O. Gonzalez, C. Rossier, M. Ruedi and C. Ucla for assistance. This work was supported by grants from the Jérôme Lejeune, Childcare and Désirée and Niels Yde Foundations, the European Union, the Swiss National Science Foundation and the NCCR Frontiers in Genetics. E Eyras is funded by the Institució Catalana de Recerca I Estudis Avançats (ICREA). R Guigo's lab acknowledges support from grants from the National Plan for R&D (Spain), QLK3-CT-2002-02062 from the European Community and HG003150-01 from the National Institutes of Health. Research on gene prediction by vertebrate genome comparison in the Brent lab is supported by grant R01 HG02278 from the National Institutes of Health. This work was also supported by EMBL and the Wellcome Trust.
Figures and Tables
Figure 1 Venn diagram of the prediction sets. Venn diagram obtained from the comparison of the three prediction sets: Ensembl (E), SGP2 (S) and TWINSCAN (T). (A) Description of each subset in the Venn diagram. (B) Total number of intron assemblies (IAs) populating each subset. (C) Percentage of experimentally verified IAs for each subset (top) and number of assayed IAs (bottom). (D) Percentage of correctly predicted splice junctions (top) from the experimentally verified IAs (bottom).
Figure 2 Comparison of two predictions. From the comparison of two predictions (a) we obtain three differentiated sets of intron assemblies (IAs): the set of IAs that are identical in both transcripts ('A and B'), and two set of the IAs that are in one prediction but not in the other ('A not B' and 'B not A'). When two sets have the same intron with different outside boundaries for the flanking exons these boundaries are taken from the intersection of the exons. Ensembl predictions (b) have in general more than one transcript per gene (two top yellow tracks). The intersecting intron assemblies (IAs) are therefore defined as the longest non-redundant IAs common between the transcripts from either prediction. For the novel IAs we take the longest non-redundant IAs in one that are not present in the other set.
Figure 3 Ensembl extensions. Exon extensions to Ensembl predictions can be obtained from exons predicted by TWINSCAN and/or SGP2. These exons can either (a) be part of a transcript with exons in common with the Ensembl transcript (linked) or (b) be part of a close but non-overlapping transcript (unlinked).
Figure 4 Classification of intron assemblies. We classified the intron assemblies that were novel with respect to a reference set according to their position relative to the other set against which we do the comparison. A novel IA can (a) fall between the genomic extent of two predictions (intergenic), (b1) bridge across two predictions (bridge), (c1) overlap the 5' or the 3' end of one prediction (external), and (d1) fall within one or more introns of another prediction (intronic). Additionally, novel IAs are labelled as complete when they are a complete ATG-to-STOP prediction: (b2), (c2) and (d2).
Table 1 distribution of intron assemblies (IAs) for each of the 7 subsets of the Venn diagram of the three prediction sets: Ensembl (E), SGP2 (S) and TWINSCAN (T) (see also Figure 1). The number of transcripts from each prediction set participating in the intron assemblies is indicated on the right.
Distribution of intron assemblies
Subsets Number of IAs Transcripts involved
E S T
E and S and T 10650 10282 8974 9888
(S and T) not E 4769 0 3930 3924
(E and S) not T 4757 3636 3273 0
(T and E) not S 1748 1592 0 1507
S not (E+T) 25119 0 20740 0
T not (S+E) 27592 0 0 22239
E not (T+S) 13514 11014 0 0
Table 2 Experimentally verified of intron assemblies (see also Figure 1)
Experimentally verified intron assemblies
Total tested No amplimer Amplimer correctly predicted Amplimer but junction not correctly predicted
S and T and E 20 3 (15%) 16 (80%) 1 (5%)
(S and T) not E 76 40 (53%) 27 (35%) 9 (12%)
(E and S) not T 64 12 (19%) 44 (69%) 8 (12%)
(T and E) not S 40 14 (35%) 22 (55%) 4 (10%)
S not (T + E) 88 67 (76%) 6 (7%) 15 (17%)
T not (S + E) 96 83 (86%) 9 (9%) 4 (5%)
E not (T + S) 30 7 (23.3%) 16 (53.3%) 7 (23.3%)
Table 3 Experimental verification of IAs corresponding to Ensembl 5' extensions. The extensions are separated according to whether the 5'-most Ensembl exon also existed in TWINSCAN and/or SGP2 (linked) or not (unlinked) (see Figure 3).
Experimentally verified Ensembl extensions
Ensembl extensions Total tested No amplimer Amplimer correctly predicted Amplimer but junction not correctly predicted
linked 60 36 (60%) 11 (18%) 13(22%)
unlinked 29 27 (93%) 2 (7%) 0
==== Refs
International Chicken Genome Sequencing Consortium Sequence and Comparative Analysis of the Chicken Genome provide unique perspectives on vertebrate evolution Nature 2004 432 695 716 Erratum in: Nature 433(7027):777 15592404 10.1038/nature03154
Alexandersson MS Cawley Pachter L SLAM: cross-species gene finding and alignment with a generalized pair hidden Markov model Genome Res 2003 13 496 502 12618381 10.1101/gr.424203
Flicek PE Keibler P Hu I Korf Brent MR Leveraging the mouse genome for gene prediction in human: From whole-genome shotgun reads to a global synteny map Genome Res 2003 13 46 54 12529305 10.1101/gr.830003
Parra G Agarwal P Abril JF Wiehe T Fickett JW Guigo R Comparative Gene Prediction in Human and Mouse Genome Res 2003 13 108 117 12529313 10.1101/gr.871403
Tenney A Brown RH Vaske C Lodge JK Doering TL Brent MR Gene prediction and verification in a compact genome with numerous small introns Genome Res 2004 14 2330 2335 15479946 10.1101/gr.2816704
Wei C Lamesch P Arumugam M Rosenberg J Hu P Brent MR Vidal M Closing in on the C. elegans ORFeome by Cloning TWINSCAN Genome Res 2005 15 577 82 15805498 10.1101/gr.3329005
International Human Genome Sequencing Consortium Initial Sequencing and Analysis of the Human Genome Nature 2001 409 860 921 11237011 10.1038/35057062
Venter JC Adams MD Myers EW Li PW Mural RJ Sutton GG Smith HO Yandell M Evans CA Holt RA Gocayne JD Amanatides P Ballew RM Huson DH Wortman JR Zhang Q Kodira CD Zheng XH Chen L Skupski M Subramanian G Thomas PD Zhang J Gabor Miklos GL Nelson C Broder S Clark AG Nadeau J McKusick VA Zinder N Levine AJ Roberts RJ Simon M Slayman C Hunkapiller M Bolanos R Delcher A Dew I Fasulo D Flanigan M Florea L Halpern A Hannenhalli S Kravitz S Levy S Mobarry C Reinert K Remington K Abu-Threideh J Beasley E Biddick K Bonazzi V Brandon R Cargill M Chandramouliswaran I Charlab R Chaturvedi K Deng Z Di Francesco V Dunn P Eilbeck K Evangelista C Gabrielian AE Gan W Ge W Gong F Gu Z Guan P Heiman TJ Higgins ME Ji RR Ke Z Ketchum KA Lai Z Lei Y Li Z Li J Liang Y Lin X Lu F Merkulov GV Milshina N Moore HM Naik AK Narayan VA Neelam B Nusskern D Rusch DB Salzberg S Shao W Shue B Sun J Wang Z Wang A Wang X Wang J Wei M Wides R Xiao C Yan C Yao A Ye J Zhan M Zhang W Zhang H Zhao Q Zheng L Zhong F Zhong W Zhu S Zhao S Gilbert D Baumhueter S Spier G Carter C Cravchik A Woodage T Ali F An H Awe A Baldwin D Baden H Barnstead M Barrow I Beeson K Busam D Carver A Center A Cheng ML Curry L Danaher S Davenport L Desilets R Dietz S Dodson K Doup L Ferriera S Garg N Gluecksmann A Hart B Haynes J Haynes C Heiner C Hladun S Hostin D Houck J Howland T Ibegwam C Johnson J Kalush F Kline L Koduru S Love A Mann F May D McCawley S McIntosh T McMullen I Moy M Moy L Murphy B Nelson K Pfannkoch C Pratts E Puri V Qureshi H Reardon M Rodriguez R Rogers YH Romblad D Ruhfel B Scott R Sitter C Smallwood M Stewart E Strong R Suh E Thomas R Tint NN Tse S Vech C Wang G Wetter J Williams S Williams M Windsor S Winn-Deen E Wolfe K Zaveri J Zaveri K Abril JF Guigo R Campbell MJ Sjolander KV Karlak B Kejariwal A Mi H Lazareva B Hatton T Narechania A Diemer K Muruganujan A Guo N Sato S Bafna V Istrail S Lippert R Schwartz R Walenz B Yooseph S Allen D Basu A Baxendale J Blick L Caminha M Carnes-Stine J Caulk P Chiang YH Coyne M Dahlke C Mays A Dombroski M Donnelly M Ely D Esparham S Fosler C Gire H Glanowski S Glasser K Glodek A Gorokhov M Graham K Gropman B Harris M Heil J Henderson S Hoover J Jennings D Jordan C Jordan J Kasha J Kagan L Kraft C Levitsky A Lewis M Liu X Lopez J Ma D Majoros W McDaniel J Murphy S Newman M Nguyen T Nguyen N Nodell M Pan S Peck J Peterson M Rowe W Sanders R Scott J Simpson M Smith T Sprague A Stockwell T Turner R Venter E Wang M Wen M Wu D Wu M Xia A Zandieh A Zhu X The sequence of the human genome Science 2001 291 1304 51 Erratum in: Science 2001, 292(5523):1838 11181995 10.1126/science.1058040
Mouse Genome Sequence Consortium Initial sequencing and comparative analysis of the mouse genome Nature 2002 420 520 12466850 10.1038/nature01262
Rat Genome Sequencing Project Consortium Genome sequence of the Brown Norway rat yields insights into mammalian evolution Nature 2004 428 493 15057822 10.1038/nature02426
Aparicio S Chapman J Stupka E Putnam N Chia JM Dehal P Christoffels A Rash S Hoon S Smit A Gelpke MD Roach J Oh T Ho IY Wong M Detter C Verhoef F Predki P Tay A Lucas S Richardson P Smith SF Clark MS Edwards YJ Doggett N Zharkikh A Tavtigian SV Pruss D Barnstead M Evans C Baden H Powell J Glusman G Rowen L Hood L Tan YH Elgar G Hawkins T Venkatesh B Rokhsar D Brenner S Whole-genome shotgun assembly and analysis of the genome of Fugu rubripes Science 2002 297 1301 10 12142439 10.1126/science.1072104
Boardman PE Sanz-Esquerro J Overton IM Burt DW Bosch E Fong WT Tickle C Brown WRA Wilson SA Hubbart SJ A comprehensive collection of chicken cDNAs Curr Biol 2002 12 1965 1969 12445392 10.1016/S0960-9822(02)01296-4
Curwen V Eyras E Andrews DT Clarke L Mongin E Searle S Clamp M The Ensembl Automatic Gene Annotation System Genome Res 2001 14 942 10.1101/gr.1858004
Korf I Flicek P Duan D Brent MR Integrating Genomic Homology into Gene Structure Prediction Bioinformatics 2001 17 140 148
Wu JQ Shteynberg D Arumugam M Gibbs RA Brent MR Identification of rat genes by TWINSCAN gene prediction, RT-PCR and direct sequencing Genome Res 2004 14 665 671 15060008 10.1101/gr.1959604
Guigo R Dermitzakis ET Agarwal P Ponting CP Parra G Reymond A Abril JF Keibler E Lyle R Ucla C Antonarakis SE Brent MR Comparison of mouse and human genomes followed by experimental verification yields an estimated 1,019 additional genes PNAS 2003 100 1140 1145 12552088 10.1073/pnas.0337561100
Makalowska I Sood R Faruque MU Hu P Robbins CM Eddings EM Mestre JD Baxevanis AD Carpten JD Identification of six novel genes by experimental validation of GeneMachine predicted genes Gene 2002 284 203 13 11891061 10.1016/S0378-1119(01)00897-6
Rogic S Ouellette BF Mackworth AK Improving gene recognition accuracy by combining predictions from two gene-finding programs Bioinformatics 2002 18 1034 45 12176826 10.1093/bioinformatics/18.8.1034
Howe KL Chothia T Durbin R GAZE: a generic framework for the integration of gene-prediction data by dynamic programming Genome Res 2002 12 1418 27 12213779 10.1101/gr.149502
Zhang L Pavlovic V Cantor CR Kasif S Human-Mouse Gene Identification by Comparative Evidence Integration and Evolutionary Analysis Genome Res 2003 13 1190 1202 12743024 10.1101/gr.703903
Yada T Totoki Y Takaeda Y Sakaki Y Takagi T DIGIT: A Novel Gene Finding Program by Combining Gene-Finders Proceedings of the 8th Pacific Symposium on Biocomputing (PSB 2003), Lihue, Hawaii, USA 375 387 January 3–7, 2003
Allen JE Pertea M Salzberg SL Computational gene prediction using multiple sources of evidence Genome Res 2004 14 142 8 14707176 10.1101/gr.1562804
Das M Burge CB Park E Colinas J Pelletier J Assessment of the total number of human transcription units Genomics 2001 77 71 8 11543635 10.1006/geno.2001.6620
Eyras E Caccamo M Curwen V Clamp M ESTgenes: Alternative splicing from ESTs in Ensembl Genome Res 2004 14 976 15123595 10.1101/gr.1862204
Hubbard SJ Graham DV Beattie KJ Overton IA McLaren SR Croning MDR Boardman PE Bonfield JK Burnside J Davies RM Farrel ER Francis MD Griffiths-Jones S Humphray SJ Hyland C Scott CE Tang H Taylor RG Tickle C Brown WRA Birney E Rogers J Wilson SA Transcriptome Analysis for the Chicken Based on 19,626 Finished cDNA Sequences and 485,337 Expressed Sequence Tags Genome Res 2005 15 174 83 15590942 10.1101/gr.3011405
Pruitt KD Maglott DR RefSeq and LocusLink: NCBI gene-centered resources Nucleic Acids Res 2001 29 137 140 11125071 10.1093/nar/29.1.137
Altschul SF Gish W Miller W Myers EW Lipman DJ Basic local alignment search tool J Mol Biol 1990 215 403 410 2231712 10.1006/jmbi.1990.9999
Slater GS Birney E Automated generation of heuristics for biological sequence comparison BMC Bioinformatics 2005 6 31 15713233 10.1186/1471-2105-6-31
Dermitzakis ET Reymond A Lyle R Scamuffa N Ucla C Deutsch S Stevenson BJ Flegel V Bucher P Jongeneel CV Antonarakis SE Numerous potentially functional but non-genic conserved sequences on human chromosome 21 Nature 2002 420 578 82 12466853 10.1038/nature01251
Dermitzakis ET Reymond A Scamuffa N Ucla C Kirkness E Rossier C Antonarakis SE Evolutionary discrimination of mammalian conserved non-genic sequences (CNGs) Science 2003 302 1033 5 14526086 10.1126/science.1087047
Reymond A Camargo AA Deutsch S Stevenson BJ Parmigiani RB Ucla C Bettoni F Rossier C Lyle R Guipponi M de Souza S Iseli C Jongeneel CV Bucher P Simpson AJ Antonarakis SE Nineteen additional unpredicted transcripts from human chromosome 21 Genomics 2002 79 824 32 12036297 10.1006/geno.2002.6781
Scripts to generate intron assemblies from prediction sets:
Download page for SGP2 software and predictions:
Download page for TWINSCAN software and predictions:
Website for Ensembl predictions in chicken:
Website with supplementary data:
| 15924626 | PMC1174864 | CC BY | 2021-01-04 16:02:50 | no | BMC Bioinformatics. 2005 May 30; 6:131 | utf-8 | BMC Bioinformatics | 2,005 | 10.1186/1471-2105-6-131 | oa_comm |
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-1311592462610.1186/1471-2105-6-131Research ArticleGene finding in the chicken genome Eyras Eduardo [email protected] Alexandre [email protected] Robert [email protected] Jacqueline M [email protected] Francisco [email protected] Paul [email protected] Elizabeth J [email protected] Genis [email protected] David D [email protected] Carine [email protected] Jane [email protected] Stylianos E [email protected] Ewan [email protected] Roderic [email protected] Michael R [email protected] Research Group in Biomedical Informatics, Institut Municipal d'Investigacio Medica/Universitat Pompeu Fabra/Centre de Regulacio Genomica, E08003 Barcelona, Catalonia, Spain2 Department of Genetic Medicine and Development, University of Geneva, Medical School and University Hospital of Geneva, CMU, 1, rue Michel Servet, 1211 Geneva, Switzerland3 Center for Integrative Genomics, University of Lausanne, 1015 Lausanne, Switzerland4 The Wellcome Trust Sanger Institute, Wellcome Trust Genome Campus, Hinxton, Cambridge CB10 1SA, UK5 Laboratory for Computational Genomics and Department of Computer Science, Campus Box 1045, Washington University, One Brookings Drive, St Louis, Missouri 63130, USA6 EBI, Wellcome Trust Genome Campus, Hinxton, Cambridge CB10 1SD, UK2005 30 5 2005 6 131 131 26 10 2004 30 5 2005 Copyright © 2005 Eyras et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Despite the continuous production of genome sequence for a number of organisms, reliable, comprehensive, and cost effective gene prediction remains problematic. This is particularly true for genomes for which there is not a large collection of known gene sequences, such as the recently published chicken genome. We used the chicken sequence to test comparative and homology-based gene-finding methods followed by experimental validation as an effective genome annotation method.
Results
We performed experimental evaluation by RT-PCR of three different computational gene finders, Ensembl, SGP2 and TWINSCAN, applied to the chicken genome. A Venn diagram was computed and each component of it was evaluated. The results showed that de novo comparative methods can identify up to about 700 chicken genes with no previous evidence of expression, and can correctly extend about 40% of homology-based predictions at the 5' end.
Conclusions
De novo comparative gene prediction followed by experimental verification is effective at enhancing the annotation of the newly sequenced genomes provided by standard homology-based methods.
==== Body
Background
The draft sequence of the chicken (Gallus gallus) genome has been recently obtained and an initial analysis completed [1]. This genome sequence will be extremely valuable for vertebrate development, avian biology, and agriculture. To reap these benefits, however, we must be able to determine with reasonable accuracy the amino acid sequences of the proteins encoded by this genome. We therefore set out to adapt three leading gene prediction systems to the chicken genome sequence and to evaluate their predictions experimentally. Some features of the chicken genome facilitate gene prediction while others hinder it. Generally, gene prediction tends to be more accurate in more compact genomes than in the large mammalian genomes [2-6]. The chicken genome is about 40% the size of the human genome, but about three times the size of the Takifugu genome [7-11]. This translates into a substantial reduction in repeats and pseudogenes. Interspersed repeats cover about 9% of the genome, far less than the 40–50% found in mammals [1]. The small number of processed pseudogenes is beneficial to automated annotation, as this technology often misclassifies pseudogenes as functional. De novo prediction methods need full open reading frame (ORF) mRNAs to train their statistical models and homology-based methods rely on these sequences in order to predict a core set of high quality genes. However, an abundant set of chicken full-length cDNAs was not available at the start of this project, as only about 1,800 putatively full-length cDNAs and 340,000 ESTs were deposited in GenBank [12].
We set out to determine how well the computational methods used for annotating protein-coding genes in the mouse and rat genomes [9,10] would perform in this avian context, where transcriptome sequencing was much less advanced and where sequence divergence may be beyond the optimal distance. An assessment of the computational methods was carried out by a large-scale experimental verification by RT-PCR. The gene prediction tools tested were Ensembl, SGP2 and TWINSCAN. Ensembl is a homology-based method, which builds gene-models using species-specific known sequences and proteins from other species aligned to the genome [13]. SGP2 and TWINSCAN are de novo comparative gene predictors whose only inputs are the genome to be annotated and a second informant genome [3,4,14-16]. In this case, the informant genome selected for comparison was human. Since Ensembl relies on mapping known genes from chicken and other organisms to the chicken genome, we surmised that its prediction set would contain fewer false positives (predicted genes that are not real) and more accurate gene structures, as well as more false negatives (real genes that are not predicted) than the de novo methods. Conversely, SGP2 and TWINSCAN should be effective in detecting bona fide genes missed by the Ensembl pipeline. In particular, we were able to verify gene structures simultaneously predicted by both de novo methods about 50% of the time. In this respect, de novo comparative methods complement homology-based methods, which in general miss genes for which there is no pre-existing evidence of transcription.
To test and compare the performance of these prediction systems under the new conditions presented by the chicken genome, we tested a large number of predicted genes by RT-PCR and direct sequencing. We sampled individually from the predictions for which either one or two of the three methods agreed, and the remainder did not. We thus focused our tests on the most difficult genes to predict and on the differences between the various prediction systems. We present here the largest experimental comparison of multiple gene-prediction programs to date as well as the first to use this kind of differential design. We aimed to evaluate how de novo and homology-based gene finding methods perform in a newly sequenced genome for which a small number of gene sequences are known. In particular we have (1) evaluated the three prediction methods, (2) investigated TWINSCAN and SGP2 as possible effective methods to complement the Ensembl prediction pipeline and (3) tested the overall specificity of the Ensembl prediction set.
Results and discussion
Experimental evaluation of gene finding in chicken
We aimed to estimate the accuracy of each of the gene prediction sets, which consisted of 29,430 SGP2 (S) and 29,052 TWINSCAN (T) gene predictions, with one transcript per gene, and 17,709 Ensembl (E) genes containing 28,416 transcripts. We classified the predictions according to the Venn diagram defined by the three-way intersections of the sets and their complements (Figure 1A, Table 1). The subsets were populated with intron assemblies (IAs), defined as a list of exons and introns contiguous in a predicted transcript (see Figure 2 and Methods for details), and can be classified into three types: (1) the orphan subsets, containing those elements that are in one set but not in any of the other two, (2) the two-way intersection subsets, containing those elements that are in two sets and not in the third, and (3) the triple intersection, containing those elements that are in all three sets. Subsequently we tested pairs of adjacent exons from each of the subsets.
Complementing homology-based gene prediction with de novo methods
Selected IAs belonging to the three possible two-way intersections were experimentally tested. The results are summarized in Figure 1 and detailed in Table 2. After RT-PCR, gel purification, and direct sequencing, about 50% of the tested transcripts predicted by both SGP2 and TWINSCAN, but not by Ensembl yielded spliced alignments to the gene targeted (Figure 1C). This rate is higher than that reported for human using a combination of homology-based and single-genome predictors [16,17] in spite of the lack of available known gene sequences for the chicken genome. There is a total of 4,769 IAs in the '(S and T) not E' subset, corresponding to a total of 13,470 exons. Projecting these results onto genes and using an average distribution of coding exons per gene from other vertebrates (Human, Mouse and Rat), we estimate that approximately 740 to 840 bona fide chicken genes that are not in the currently predicted Ensembl set can be found by the de novo comparative methods followed by direct amplification and sequencing.
Considering the set of IAs unique to one prediction set (orphans), 39% of these have one single intron and 80% have 1 or 2 introns. On the other hand, after testing experimentally 96 orphans from TWINSCAN, 88 orphans from SGP2 and 30 orphans from Ensembl, we found that about 77% of the Ensembl orphans are real genes, compared to an average 18% for TWINSCAN and SGP2 (see Figure 1, see Table 2). Thus while Ensembl orphans are more likely to be real genes not predicted by the other methods, orphan de novo predictions are more likely to be false positives.
Extending homology-based gene predictions with de novo methods
As Ensembl predictions often fail to correctly predict one or both ends of a gene [13], we reasoned that de novo prediction methods could help in extending the homology-based predicted transcripts. To test this hypothesis, we identified candidate 5' extensions: exons predicted by TWINSCAN, SGP2 or both to the 5' side of Ensembl genes on the same strand. We found that 8,368 (47%) Ensembl genes have such candidate extensions. However, not all these extensions were as likely to correspond to real exons. From this total, 7,630 genes had extensions suggested by de novo predictions that overlap the Ensembl gene (linked, see Figure 3A), and 738 (4%) had extra exons from de novo predictions that did not overlap the Ensembl gene (unlinked, see Figure 3B). As 99% of Ensembl introns were no longer than 100 kb, we considered only exon extensions that were no further than 100 kb from the 5'-most Ensembl exon. Interestingly, we found that 93% of the linked extensions were to multiexonic Ensembl transcripts, the remainder being extensions to single-exon Ensembl transcripts; however, for unlinked extensions, only 58% were to multiexonic Ensembl transcripts.
We investigated experimentally 60 linked and 29 unlinked extensions by designing one primer in the 5'-most exon of the Ensembl prediction and the second primer within one of the upstream exon suggested by TWINSCAN and/or SGP2 to the 5' side. The RT-PCR results (see Table 3) show that de novo methods-suggested linked extensions were correct in about 40% of the cases. This rate dropped to a mere 7% for the unlinked extensions.
Separating the extensions according to whether the extra exon was predicted either by SGP2 or by TWINSCAN showed that both methods had a comparable contribution. From the 60 linked tested extensions, 46 were predicted by SGP2 with 21 (46%) RT-PCR positives, 36 were predicted by TWINSCAN with 15 (42%) positives, and 22 were predicted by both programs with 12 (54%) positives. On the other hand, from the 29 tested unlinked extensions 15 were predicted by SGP2, with 2 positives, and 17 were predicted by TWINSCAN, with 1 positive. Finally, there were 3 cases where both, SGP2 and TWINSCAN, predicted the unlinked extra exon of which one was positive.
Testing Ensembl specificity
A randomly selected set of Ensembl predictions was assayed to evaluate Ensembl's specificity. This test measured a false positive rate of 4% (see Table S5 of the supplementary material for more details). On the other hand, the tested exon-pairs for the two-way intersection sets that included Ensembl ('(E and S) not T' and '(T and E) not S') had an average false positive rate of 19% and 35%, respectively (see Figure 1). The disparity is greater with the two-way intersection set that excludes Ensembl ('(S and T) not E'), which shows a false positive rate of 53%. One explanation for this difference is the observation that most of the Ensembl predictions have exons predicted by both SGP2 and TWINSCAN. Indeed, 25,222 (89%) of the 28,416 Ensembl transcripts have at least one exon, which is also in the SGP2 and TWINSCAN sets, and 82% of these transcripts have 2 or more exons in common with both de novo methods. Thus, Ensembl predictions are most likely to fall within a triple intersection, resulting in an increased rate of true positives. Based on previous experiments [16], we expected the triple intersection to give a yield close to 100% positive rate. We tested 20 cases of triple intersection and found 85% positive rate (see Figure 1C). Moreover, 94% of these positive cases had the exon-intron boundaries correctly predicted (see Figure 1D).
We compared the accuracy of the human-based predictions with the accuracy of a fish-based set of predictions. We predicted genes in chicken with SGP2 and TWINSCAN using Tetraodon nigroviridis as informant genome, and found that they are less accurate than the human-based ones (see Table S7 of the supplementary material). Additionally, we found that 85% and 98% of the Tetraodon-based predictions from TWINSCAN and SGP2, respectively, overlap the corresponding human-based ones. Interestingly, 85% of the TWINSCAN orphans overlap TWINSCAN Tetraodon-based predictions, and 80% of the SGP2 orphans overlap SGP2 Tetraodon-based predictions. On the other hand, 99% of the IAs common to TWINSCAN and SGP2 but not in Ensembl, overlap TWINSCAN or SGP2 Tetraodon-based predictions.
Conclusions
In this paper we have evaluated how effective purely computational approaches for genome annotation can be, even in the absence of a large collection of previously known genes, by means of the largest attempt so far to experimentally compare several gene finders. After testing the accuracy of Ensembl, SGP2 and TWINSCAN on the chicken genome we have shown that de novo comparative methods followed by experimental verification remain a successful approach in the annotation of newly sequenced genomes from which little is known.
We found that approximately 50% of predictions that were in TWINSCAN and SGP2 but not in Ensembl could be experimentally verified (Figure 1). These experiments demonstrate that de novo comparative prediction methods are effective at complementing homology-based methods and confirm that a combination of methods can improve the prediction accuracy [18-22]. Moreover, in spite of the limited gene sequence data available for chicken, the combination of TWINSCAN and SGP2 achieves better accuracy than previous attempts to verify by RT-PCR computational predictions that fall outside a set of annotations [17,23]. On the other hand, looking at the intron assemblies unique to one prediction set, the proportion of positives is largely reduced for predictions not in Ensembl. The predictions unique to one of the de novo methods show an abundance of gene models with 2 and 3 exons, which may be artefacts due to genome misassemblies. These results are in contrast with the high success rate (77%) of the predictions unique to Ensembl. This is a reasonable observation considering that the Ensembl prediction pipeline has access to genes that do not follow a 'standard' gene-grammar (e.g., unusual codon usage), but which may nevertheless be represented in the cDNA/protein databases used.
The Ensembl chicken gene set has been found to have a 96% positive rate, whereas the IAs from the two-way intersections that include Ensembl, '(E and S) not T' and '(T and E) not S', and the Ensembl orphans, have a lower positive rate, 81%, 65% and 77%, respectively, which stems from the fact that most exons predicted by Ensembl are also predicted by both SGP2 and TWINSCAN. Additionally, de novo comparative methods are useful for extending partial predictions from homology-based methods. Ensembl may generate predictions based on protein fragments or on partial homology from other species, and TWINSCAN and SGP2 predictions can add bona fide exons to the Ensembl predictions they overlap with. For the 5' end we show that 40% of the tested cases, where either TWINSCAN or SGP2 predicted at least one additional exon, were verified (Table 3). To our knowledge, this is the first time that experimental evidence is provided for extensions to homology-based models produced by de novo methods.
We observed that the subsets containing SGP2 IAs (e.g., '(S and E) not T)') have in general a higher proportion of RT-PCR positives than those containing TWINSCAN IAs (e.g., '(T and E) not S)') (Figure 1C, Table 2). There are two factors that may contribute to this difference. The first is an intrinsic difference between TWINSCAN and SGP2 – SGP2 uses TBLASTX (translated) alignments between human and chicken to reward exons overlapping aligned regions, whereas TWINSCAN uses BLASTN (nucleotide) alignments to influence the scores of exons, splice sites, and translation initiation and termination sites. Human and chicken are sufficiently diverged that translated alignments may be more sensitive, whereas nucleotide alignments fail to cover many known exons. The other factor is incidental to the way TWINSCAN was trained and run to produce the predictions tested. TWINSCAN used 525 chicken RefSeqs to estimate parameters for its probability model. This training set was probably too small to produce optimal parameter values. SGP2, on the other hand, was run with a combination of parameters estimated from the much larger set of known human genes (for its model of chicken DNA sequence) and parameters were hand tuned using the same 525 chicken genes (for its scoring of human-chicken alignments). Although a larger fraction of SGP2 predictions yielded positive experimental results, we found that TWINSCAN tends to be more accurate than SGP2 in the prediction of the intron boundaries (Figure 1D, Table 2). This difference stems from the intron model used by TWINSCAN, as opposed to SGP2, which does not model introns explicitly. TWINSCAN was re-run after completion of the experiments with an improved intron-length model, yielding a prediction set that was substantially smaller and more accurate (see Table S6) than the set tested. In spite these differences, comparing the gene predictions with a set of coding cDNAs released after the completion of these analyses, we found that all three methods have similar sensitivity (79%) (see Methods for details), hence the de novo comparative methods cover a fraction of the transcriptome similar to homology-based methods with a minimal initial amount of genome-specific expression data.
The experimentally verified IAs represent a fraction of the actual number of chicken genes that can be eventually found using our methods. If we extrapolate the proportions of experimentally verified IAs (Figure 1C) to all the generated IAs in the Venn diagram (Figure 1B) and using an average distribution of coding exons per gene from other vertebrates (Human, Mouse and Rat), we estimate a range of 14,600 to 17,500 experimentally verifiable chicken genes from our computational predictions. In this paper we only analysed intron assemblies and deliberately left out a number of chicken protein-coding intronless predictions (3,049 from SGP2, 2,727 from TWINSCAN and 1,855 from Ensembl). The triple intersection of these intronless genes contains 148 genes, which are worth investigating and for which techniques different from the ones applied here will be required.
Considering all 2-way IAs (see Table 1), one would need 11,274 RT-PCR reactions to experimentally confirm about 7,232 (64%) genes. This number of experiments compares favourably to large-scale EST projects with the added benefit of having almost no redundancy (only gene fissions and misassemblies will contribute to redundancy). The biggest drawback to EST sequencing is its large redundancy and extensive overlap. The falling cost of primers and the increased flexibility of large-scale molecular biology centers make this approach of computational prediction followed by experimental verification cost effective and scalable [5,6]. As RT-PCR primers can be designed with appropriate linker sites such an approach could also provide a physical resource of clonable fragments. We conclude that de novo comparative gene predictions followed by experimental verification is an effective way to carry out the annotation of a newly sequenced genome for which little gene sequence information is known. In particular, as our results show, performing RT-PCR and sequencing for all the predicted novel genes, starting with those predicted by multiple de novo methods, should enhance the quality of the annotation in forthcoming eukaryote genome sequencing projects.
Methods
Generation of predictions
The initial lack of an abundant set of known chicken gene sequences forced us to adapt the methodology for training of the de novo methods and running the Ensembl gene build. The Ensembl prediction pipeline [13] builds gene models from known vertebrate proteins and cDNA sequences, whereas gene models based solely on ESTs are usually kept separated as ESTGenes [24]. For the chicken genome, ESTs were combined with the standard gene build to include additional genes and transcripts. The alignments from approximately 400,000 ESTs and 24,000 full-length cDNAs [12,25] from multiple tissues were conciliated into non-redundant transcript structures [24]. Those predicted models that fell on non-annotated loci and those that contributed with at least two new exons in previously annotated loci were added to the protein-based gene set. Single-exons transcripts produced by the EST/cDNA-based models were rejected, as they could not be distinguished from cloning artefacts without protein evidence. For our analyses we did not include the untranslated regions that Ensembl annotates combining the cDNA/EST and protein evidence [13]. Ensembl also annotates processed pseudogenes [13], which we did not consider either.
TWINSCAN was trained on 525 chicken RefSeq sequences [26] aligned to the chicken genomic sequence. This set is based on a set of 1266 provisional RefSeq mRNAs from GenBank (27 March 2004) March 27 filtered in the following way: The Refseq mRNA sequences were matched with the corresponding mRNAs placed on the Chicken genome assembly at UCSC. Any mRNA not placed was not considered. Additionally, sequences without an ungapped alignment between the entire CDS portion of the Refseq mRNA and the extracted unmasked genomic sequence were removed. Additionally, cases with in-frame stop codons and/or non-canonical splice sites were also removed. TWINSCAN uses nucleotide alignments and has specific models for how the alignments modify the scores of the gene signals. The BLASTN [27] alignments used by TWINSCAN covered 3.8% of the chicken assembly.
SGP2 training followed a hybrid approach. SGP2 was run with human parameters for the coding statistics and splice sites, whereas the score weights that reward the human-chicken homologies and penalize the lack of them were optimised using the same set of 525 Chicken RefSeqs used for the TWINSCAN training. SGP2 was then run on unsegmented chicken chromosomes using the TBLASTX [27] alignments with the human genomic sequence (assembly NCBI34). These alignments, which covered approximately 3% of the Chicken genome, were enriched with 391,610 extra HSPs obtained from the ungapped Exonerate [28] alignments of human proteins from the Ensembl predictions (release NCBI34c), the GeneId prediction set for the same human assembly and the set of vertebrate RefSeq proteins (version of April 2004). The extra alignments covered 43% of the nucleotides in TBLASTX HSPs and 5% of their sequence represented 5,840 non-redundant homology regions that had no overlap with the TBLASTX hits. These extra alignments produced a considerable improvement of the sensitivity and specificity at the gene level with respect to SGP2 predictions using only TBLASTX HSPs when tested against the Ensembl set and the aforementioned 525 RefSeqs. It also achieved a slight improvement of the sensitivity at the exon and nucleotide level.
Classification of predictions
We classified the predictions according to the Venn diagram defined by the intersections of the three sets: Ensembl (E), TWINSCAN (T) and SGP2 (S). Using the exact identity between transcripts is problematic because the exon-structures from each prediction are in general not identical. Thus we defined the intron assembly (IA) as the element of comparison. An IA is a list of exons and introns that are contiguous in a given predicted transcript. When comparing two predictions three differentiated sets are produced: the set of IAs that are identical in both transcripts, and the two sets of IAs that are present in one prediction and not in the other (see Figure 2). If the boundaries of the exons do not agree between two predictions the produced IA contains only the exonic sequence that is common to both predictions, i.e., an intersecting IA contains the sequence and exon-structure on which both predictors agree (see Figure 2). An intron assembly represents naturally the entities to be tested by RT-PCR, where an exon-pair, separated by one or more introns and by not more than about 1 Kb, is tested for amplification in cDNA tissue libraries. As each IA contains the maximal set of introns in a given genomic locus that fall in one of the different categories of the Venn diagram, we picked up one exon-pair from each IA, making sure to test different genes in each Venn diagram.
Combining the operations of set-difference 'not', set-intersection 'and', and set-union '+', we populated the Venn diagram. For instance: the subset 'E not (T+S)' was generated by first obtaining the union of T and S predictions (T+S) and then comparing the Ensembl predictions with this latter set, hence 'E not (T+S)' contains Ensembl IAs that are not in TWINSCAN or SGP2. Similarly, we obtained the subset '(S and T) not E' by first calculating the intersection of S and T, and then calculating the set-difference against E. Note that the set-difference is non-commutative, as we keep elements from one set and use the other for comparison, hence S not T ? T not S. These operations divide the subsets into three types: (1) orphan subsets, formed by those elements that are in one set but not in any of the other two, (2) the two-way intersection subsets, formed by those elements that are in two sets and not in the third one, and (3) the triple intersection, formed by those elements that are in the three sets. Figure 1B shows the number of IAs in each of the subsets (see also Table 1).
When considering the Ensembl predictions we introduced a slight modification of the operations, as one Ensembl gene may have more than one transcript (see Figure 2b). We defined the intersecting IAs as the longest non-redundant common IA between the transcripts from either prediction. That is, from all the redundant intron assemblies in an Ensembl gene that are also in a prediction from SGP2 or TWINSCAN, we took the longest one. By the concept of redundancy of two intron assemblies we mean that they have the same splicing structure or one is included in the structure of the other, allowing for mismatches in the exon edges. Likewise, for the set difference involving Ensembl predictions we took the longest IA from all the ones in an Ensembl gene that were novel with respect to the de novo prediction. The inverse case works similarly: an IA in TWINSCAN or SGP2 which is not in an Ensembl gene is the longest IA that is novel with respect to all the Ensembl transcripts in that gene.
We classified IAs according to their position relative to the predictions from the other set against which they are compared. We considered an IA to be:
• Intergenic: if it falls outside the genomic extension of any of the predictions from the other set.
• Bridge: if it bridges between two different predictions in the other set.
• Intronic: if it extends one or more introns of the excluded prediction, i.e., the exons of the IA fall in the introns of the other prediction.
• External: if it extends the 5' or 3' of the other prediction.
In Table S1 we give the distribution of the IAs according to their relative position (see Figure 4). We observe that while SGP2 produces approximately the same proportion of external and intronic orphan IAs, TWINSCAN produces many more intronic ones.
Finally, IAs were further classified according to whether they are complete, i.e., they represent a complete ATG-to-STOP prediction in at least one set (Figure 4). Table S2 presents the number of complete IAs from each set. TWINSCAN predictions produced more complete novel genes within introns than SGP2, whereas SGP2 predicts more complete novel genes that extend the genomic span of other genes. Furthermore, more than 70% of the complete intergenic orphan IAs from all programs have either two or three exons. Finally, analysing the triple intersection of the prediction sets we find 10,650 IAs that are common to the three sets from which 8,837 (83%) are complete. In contrast to the complete IAs predicted by only one program, 55% of those predicted by all programs have more than 3 exons.
Comparison to chicken expression data
We compared the prediction sets against 13,880 chicken cDNAs annotated as coding for protein and 4,154 cDNAs annotated as non-coding [12,25]. We used Exonerate in ungapped mode to directly compare the nucleotide sequence of the gene predictions against the cDNA set. We considered a cDNA to be 'found' if there was an alignment with more than 90% identity over more than 60 bp. The results showed that Ensembl and SGP2 included 79% and TWINSCAN included 77% of the coding cDNAs, whereas they included a 17%, 20% and 30% of non-coding cDNAs, respectively. Additionally, we found that 8,126 (28%) SGP2 genes, 7,037 (40%) Ensembl genes and 7,728 (27%) TWINSCAN genes aligned against the coding cDNAs; whereas 838 (3%) SGP2 genes, 1,186 (7%) Ensembl genes and 722 (2.5%) TWINSCAN genes aligned against the cDNAs labelled as non-coding. The three prediction sets together found a total of 12,009 (86.5%) of coding cDNAs, with a common set of 9,597 (69%) coding cDNAs found by all three.
Primer design
Primers used in the PCR reactions were designed using primer3 and filtered using ePCR. For each IA in each tested set, two exons were selected for primer placement. This selection was made ensuring that the exons selected for primer placement were of adequate length to generate likely primers, and that there was a sufficient amount of coding sequence to give an amplicon length in the desired range. Additional checks were performed in cases where one prediction method included a coding exon not suggested by the other two prediction methods in the same transcript. In those cases, the length of the tentative exons was accounted for by what exons were selected for primer placement, helping reduce failure due to primers being placed too far apart in the actual transcript to get successful amplification. All designed primers and primer pairs were filtered for mispriming using the entire chicken genome to find potential priming locations. Primers returning priming locations not overlapping coordinates of the target gene were rejected. From all passing primer pairs for a given IA, the pair giving the longest expected product in the requested range was selected for amplification. The selection of targets appearing on each plate was done completely randomly among all targets returning at least one designed primer pair passing all criteria. The primer selection parameters used for the design of Plate 1 were consistent with primer3 defaults, except: Tm range was set to (63, 65), product length range was set to (300, 600), and primer length range was set to (23, 28) with 24 being optimal. The primer selection parameters used for the design of Plates 2–5 were also consistent with primer3 defaults, except: GC content range was set to (30, 70) with 50 being optimal, Tm range was set to (59, 62), product length range was set to (150, 500), and primer length range was set to (17, 27) with 20 being optimal.
Experimental verification of predictions by RT-PCR
cDNA preparation
Multiple organs (brain, liver, heart, spleen, lung, kidney, muscle, tongue, trachea, crop, proventriculus, gizzard, gall bladder, small-intestine, pancreas, caeca, mesentery, ovary, oviduct and testis) of an adult male and two egg-laying females of the "Bleue de Hollande" strain were collected soon after sacrifice. Total RNA was prepared from frozen tissues using TRIzol Reagent (Invitrogen) according to manufacturers' instructions. The quality of all RNA samples was checked using an Agilent 2100 Bioanalyzer (Agilent Technologies) and by PCR using pairs of oligos designed in four CNGs (Conserved Non-Genic Sequences) conserved between GGA1 and HSA21 [29,30], as indicators of possible genomic DNA contamination. Total RNA was converted to cDNA using Superscript II (Invitrogen) primed with random primers. For each tissue in the study, 5 μg of total RNA was converted to cDNA.
Experimental verification
Predictions of chicken genes were assayed experimentally by RT-PCR as previously described and modified [9,16,31]. Similar amounts of 12 Gallus gallus cDNAs (brain, liver, heart, spleen, lung, kidney, muscle, proventriculus, small intestine, caeca, ovary and testis, final dilution 1000x) were mixed with JumpStart REDTaq ReadyMix (Sigma) and 4 ng/ul primers (Sigma-Genosys) with a BioMek 2000 robot (Beckman). The ten first cycles of PCR amplification were performed with a touchdown annealing temperatures decreasing from 60 to 50°C; annealing temperature of the next 30 cycles was carried out at 50°C. Amplimers were separated on "Ready to Run" precast gels (Pharmacia) and sequenced. This procedure was used to experimentally assay 456 exon-exon junctions of chicken predictions. The later are representative of each subsets of predictions found in the Venn diagram of the three sets of predictions studied, i.e., the Ensembl, TWINSCAN and SGP2 set (see Table S3, and Table S4 for the transcripts used as internal controls). The sequences of the amplified exon-exon junctions can be obtained from the web site .
Authors' contributions
EE with help from AR and MRB led the writing and preparation of the manuscript. RC, EB, SEA, PF, DDS and RG contributed to the text and overall layout. MRB conceived the Venn diagram approach for the experiments in chicken and chose the target sets, with input from DDS and EE. EE generated the chicken intron assemblies and analysed the experimental results. RC, PF, EE, GP, FC generated the gene predictions and performed several analyses on them. The generation of primers was carried out by DDS, MRB and EB for chicken. AR and CW collected the tissues and prepared the RNAs/cDNAs. EJH, JMB, CW carried out the RT-PCR experiments with supervision from AR, JR, EB and SEA. The analysis of the chicken predictions was coordinated by RG, MRB and EB.
Supplementary Material
Additional File 1
Supplementary information: distribution of intron-assemblies according to relative positions; details on the RT-PCR test and control sets; accuracy evaluation of the human-based and tetraodon-based chicken gene predictions.
Click here for file
Acknowledgements
We thank J.F. Abril, O. Gonzalez, C. Rossier, M. Ruedi and C. Ucla for assistance. This work was supported by grants from the Jérôme Lejeune, Childcare and Désirée and Niels Yde Foundations, the European Union, the Swiss National Science Foundation and the NCCR Frontiers in Genetics. E Eyras is funded by the Institució Catalana de Recerca I Estudis Avançats (ICREA). R Guigo's lab acknowledges support from grants from the National Plan for R&D (Spain), QLK3-CT-2002-02062 from the European Community and HG003150-01 from the National Institutes of Health. Research on gene prediction by vertebrate genome comparison in the Brent lab is supported by grant R01 HG02278 from the National Institutes of Health. This work was also supported by EMBL and the Wellcome Trust.
Figures and Tables
Figure 1 Venn diagram of the prediction sets. Venn diagram obtained from the comparison of the three prediction sets: Ensembl (E), SGP2 (S) and TWINSCAN (T). (A) Description of each subset in the Venn diagram. (B) Total number of intron assemblies (IAs) populating each subset. (C) Percentage of experimentally verified IAs for each subset (top) and number of assayed IAs (bottom). (D) Percentage of correctly predicted splice junctions (top) from the experimentally verified IAs (bottom).
Figure 2 Comparison of two predictions. From the comparison of two predictions (a) we obtain three differentiated sets of intron assemblies (IAs): the set of IAs that are identical in both transcripts ('A and B'), and two set of the IAs that are in one prediction but not in the other ('A not B' and 'B not A'). When two sets have the same intron with different outside boundaries for the flanking exons these boundaries are taken from the intersection of the exons. Ensembl predictions (b) have in general more than one transcript per gene (two top yellow tracks). The intersecting intron assemblies (IAs) are therefore defined as the longest non-redundant IAs common between the transcripts from either prediction. For the novel IAs we take the longest non-redundant IAs in one that are not present in the other set.
Figure 3 Ensembl extensions. Exon extensions to Ensembl predictions can be obtained from exons predicted by TWINSCAN and/or SGP2. These exons can either (a) be part of a transcript with exons in common with the Ensembl transcript (linked) or (b) be part of a close but non-overlapping transcript (unlinked).
Figure 4 Classification of intron assemblies. We classified the intron assemblies that were novel with respect to a reference set according to their position relative to the other set against which we do the comparison. A novel IA can (a) fall between the genomic extent of two predictions (intergenic), (b1) bridge across two predictions (bridge), (c1) overlap the 5' or the 3' end of one prediction (external), and (d1) fall within one or more introns of another prediction (intronic). Additionally, novel IAs are labelled as complete when they are a complete ATG-to-STOP prediction: (b2), (c2) and (d2).
Table 1 distribution of intron assemblies (IAs) for each of the 7 subsets of the Venn diagram of the three prediction sets: Ensembl (E), SGP2 (S) and TWINSCAN (T) (see also Figure 1). The number of transcripts from each prediction set participating in the intron assemblies is indicated on the right.
Distribution of intron assemblies
Subsets Number of IAs Transcripts involved
E S T
E and S and T 10650 10282 8974 9888
(S and T) not E 4769 0 3930 3924
(E and S) not T 4757 3636 3273 0
(T and E) not S 1748 1592 0 1507
S not (E+T) 25119 0 20740 0
T not (S+E) 27592 0 0 22239
E not (T+S) 13514 11014 0 0
Table 2 Experimentally verified of intron assemblies (see also Figure 1)
Experimentally verified intron assemblies
Total tested No amplimer Amplimer correctly predicted Amplimer but junction not correctly predicted
S and T and E 20 3 (15%) 16 (80%) 1 (5%)
(S and T) not E 76 40 (53%) 27 (35%) 9 (12%)
(E and S) not T 64 12 (19%) 44 (69%) 8 (12%)
(T and E) not S 40 14 (35%) 22 (55%) 4 (10%)
S not (T + E) 88 67 (76%) 6 (7%) 15 (17%)
T not (S + E) 96 83 (86%) 9 (9%) 4 (5%)
E not (T + S) 30 7 (23.3%) 16 (53.3%) 7 (23.3%)
Table 3 Experimental verification of IAs corresponding to Ensembl 5' extensions. The extensions are separated according to whether the 5'-most Ensembl exon also existed in TWINSCAN and/or SGP2 (linked) or not (unlinked) (see Figure 3).
Experimentally verified Ensembl extensions
Ensembl extensions Total tested No amplimer Amplimer correctly predicted Amplimer but junction not correctly predicted
linked 60 36 (60%) 11 (18%) 13(22%)
unlinked 29 27 (93%) 2 (7%) 0
==== Refs
International Chicken Genome Sequencing Consortium Sequence and Comparative Analysis of the Chicken Genome provide unique perspectives on vertebrate evolution Nature 2004 432 695 716 Erratum in: Nature 433(7027):777 15592404 10.1038/nature03154
Alexandersson MS Cawley Pachter L SLAM: cross-species gene finding and alignment with a generalized pair hidden Markov model Genome Res 2003 13 496 502 12618381 10.1101/gr.424203
Flicek PE Keibler P Hu I Korf Brent MR Leveraging the mouse genome for gene prediction in human: From whole-genome shotgun reads to a global synteny map Genome Res 2003 13 46 54 12529305 10.1101/gr.830003
Parra G Agarwal P Abril JF Wiehe T Fickett JW Guigo R Comparative Gene Prediction in Human and Mouse Genome Res 2003 13 108 117 12529313 10.1101/gr.871403
Tenney A Brown RH Vaske C Lodge JK Doering TL Brent MR Gene prediction and verification in a compact genome with numerous small introns Genome Res 2004 14 2330 2335 15479946 10.1101/gr.2816704
Wei C Lamesch P Arumugam M Rosenberg J Hu P Brent MR Vidal M Closing in on the C. elegans ORFeome by Cloning TWINSCAN Genome Res 2005 15 577 82 15805498 10.1101/gr.3329005
International Human Genome Sequencing Consortium Initial Sequencing and Analysis of the Human Genome Nature 2001 409 860 921 11237011 10.1038/35057062
Venter JC Adams MD Myers EW Li PW Mural RJ Sutton GG Smith HO Yandell M Evans CA Holt RA Gocayne JD Amanatides P Ballew RM Huson DH Wortman JR Zhang Q Kodira CD Zheng XH Chen L Skupski M Subramanian G Thomas PD Zhang J Gabor Miklos GL Nelson C Broder S Clark AG Nadeau J McKusick VA Zinder N Levine AJ Roberts RJ Simon M Slayman C Hunkapiller M Bolanos R Delcher A Dew I Fasulo D Flanigan M Florea L Halpern A Hannenhalli S Kravitz S Levy S Mobarry C Reinert K Remington K Abu-Threideh J Beasley E Biddick K Bonazzi V Brandon R Cargill M Chandramouliswaran I Charlab R Chaturvedi K Deng Z Di Francesco V Dunn P Eilbeck K Evangelista C Gabrielian AE Gan W Ge W Gong F Gu Z Guan P Heiman TJ Higgins ME Ji RR Ke Z Ketchum KA Lai Z Lei Y Li Z Li J Liang Y Lin X Lu F Merkulov GV Milshina N Moore HM Naik AK Narayan VA Neelam B Nusskern D Rusch DB Salzberg S Shao W Shue B Sun J Wang Z Wang A Wang X Wang J Wei M Wides R Xiao C Yan C Yao A Ye J Zhan M Zhang W Zhang H Zhao Q Zheng L Zhong F Zhong W Zhu S Zhao S Gilbert D Baumhueter S Spier G Carter C Cravchik A Woodage T Ali F An H Awe A Baldwin D Baden H Barnstead M Barrow I Beeson K Busam D Carver A Center A Cheng ML Curry L Danaher S Davenport L Desilets R Dietz S Dodson K Doup L Ferriera S Garg N Gluecksmann A Hart B Haynes J Haynes C Heiner C Hladun S Hostin D Houck J Howland T Ibegwam C Johnson J Kalush F Kline L Koduru S Love A Mann F May D McCawley S McIntosh T McMullen I Moy M Moy L Murphy B Nelson K Pfannkoch C Pratts E Puri V Qureshi H Reardon M Rodriguez R Rogers YH Romblad D Ruhfel B Scott R Sitter C Smallwood M Stewart E Strong R Suh E Thomas R Tint NN Tse S Vech C Wang G Wetter J Williams S Williams M Windsor S Winn-Deen E Wolfe K Zaveri J Zaveri K Abril JF Guigo R Campbell MJ Sjolander KV Karlak B Kejariwal A Mi H Lazareva B Hatton T Narechania A Diemer K Muruganujan A Guo N Sato S Bafna V Istrail S Lippert R Schwartz R Walenz B Yooseph S Allen D Basu A Baxendale J Blick L Caminha M Carnes-Stine J Caulk P Chiang YH Coyne M Dahlke C Mays A Dombroski M Donnelly M Ely D Esparham S Fosler C Gire H Glanowski S Glasser K Glodek A Gorokhov M Graham K Gropman B Harris M Heil J Henderson S Hoover J Jennings D Jordan C Jordan J Kasha J Kagan L Kraft C Levitsky A Lewis M Liu X Lopez J Ma D Majoros W McDaniel J Murphy S Newman M Nguyen T Nguyen N Nodell M Pan S Peck J Peterson M Rowe W Sanders R Scott J Simpson M Smith T Sprague A Stockwell T Turner R Venter E Wang M Wen M Wu D Wu M Xia A Zandieh A Zhu X The sequence of the human genome Science 2001 291 1304 51 Erratum in: Science 2001, 292(5523):1838 11181995 10.1126/science.1058040
Mouse Genome Sequence Consortium Initial sequencing and comparative analysis of the mouse genome Nature 2002 420 520 12466850 10.1038/nature01262
Rat Genome Sequencing Project Consortium Genome sequence of the Brown Norway rat yields insights into mammalian evolution Nature 2004 428 493 15057822 10.1038/nature02426
Aparicio S Chapman J Stupka E Putnam N Chia JM Dehal P Christoffels A Rash S Hoon S Smit A Gelpke MD Roach J Oh T Ho IY Wong M Detter C Verhoef F Predki P Tay A Lucas S Richardson P Smith SF Clark MS Edwards YJ Doggett N Zharkikh A Tavtigian SV Pruss D Barnstead M Evans C Baden H Powell J Glusman G Rowen L Hood L Tan YH Elgar G Hawkins T Venkatesh B Rokhsar D Brenner S Whole-genome shotgun assembly and analysis of the genome of Fugu rubripes Science 2002 297 1301 10 12142439 10.1126/science.1072104
Boardman PE Sanz-Esquerro J Overton IM Burt DW Bosch E Fong WT Tickle C Brown WRA Wilson SA Hubbart SJ A comprehensive collection of chicken cDNAs Curr Biol 2002 12 1965 1969 12445392 10.1016/S0960-9822(02)01296-4
Curwen V Eyras E Andrews DT Clarke L Mongin E Searle S Clamp M The Ensembl Automatic Gene Annotation System Genome Res 2001 14 942 10.1101/gr.1858004
Korf I Flicek P Duan D Brent MR Integrating Genomic Homology into Gene Structure Prediction Bioinformatics 2001 17 140 148
Wu JQ Shteynberg D Arumugam M Gibbs RA Brent MR Identification of rat genes by TWINSCAN gene prediction, RT-PCR and direct sequencing Genome Res 2004 14 665 671 15060008 10.1101/gr.1959604
Guigo R Dermitzakis ET Agarwal P Ponting CP Parra G Reymond A Abril JF Keibler E Lyle R Ucla C Antonarakis SE Brent MR Comparison of mouse and human genomes followed by experimental verification yields an estimated 1,019 additional genes PNAS 2003 100 1140 1145 12552088 10.1073/pnas.0337561100
Makalowska I Sood R Faruque MU Hu P Robbins CM Eddings EM Mestre JD Baxevanis AD Carpten JD Identification of six novel genes by experimental validation of GeneMachine predicted genes Gene 2002 284 203 13 11891061 10.1016/S0378-1119(01)00897-6
Rogic S Ouellette BF Mackworth AK Improving gene recognition accuracy by combining predictions from two gene-finding programs Bioinformatics 2002 18 1034 45 12176826 10.1093/bioinformatics/18.8.1034
Howe KL Chothia T Durbin R GAZE: a generic framework for the integration of gene-prediction data by dynamic programming Genome Res 2002 12 1418 27 12213779 10.1101/gr.149502
Zhang L Pavlovic V Cantor CR Kasif S Human-Mouse Gene Identification by Comparative Evidence Integration and Evolutionary Analysis Genome Res 2003 13 1190 1202 12743024 10.1101/gr.703903
Yada T Totoki Y Takaeda Y Sakaki Y Takagi T DIGIT: A Novel Gene Finding Program by Combining Gene-Finders Proceedings of the 8th Pacific Symposium on Biocomputing (PSB 2003), Lihue, Hawaii, USA 375 387 January 3–7, 2003
Allen JE Pertea M Salzberg SL Computational gene prediction using multiple sources of evidence Genome Res 2004 14 142 8 14707176 10.1101/gr.1562804
Das M Burge CB Park E Colinas J Pelletier J Assessment of the total number of human transcription units Genomics 2001 77 71 8 11543635 10.1006/geno.2001.6620
Eyras E Caccamo M Curwen V Clamp M ESTgenes: Alternative splicing from ESTs in Ensembl Genome Res 2004 14 976 15123595 10.1101/gr.1862204
Hubbard SJ Graham DV Beattie KJ Overton IA McLaren SR Croning MDR Boardman PE Bonfield JK Burnside J Davies RM Farrel ER Francis MD Griffiths-Jones S Humphray SJ Hyland C Scott CE Tang H Taylor RG Tickle C Brown WRA Birney E Rogers J Wilson SA Transcriptome Analysis for the Chicken Based on 19,626 Finished cDNA Sequences and 485,337 Expressed Sequence Tags Genome Res 2005 15 174 83 15590942 10.1101/gr.3011405
Pruitt KD Maglott DR RefSeq and LocusLink: NCBI gene-centered resources Nucleic Acids Res 2001 29 137 140 11125071 10.1093/nar/29.1.137
Altschul SF Gish W Miller W Myers EW Lipman DJ Basic local alignment search tool J Mol Biol 1990 215 403 410 2231712 10.1006/jmbi.1990.9999
Slater GS Birney E Automated generation of heuristics for biological sequence comparison BMC Bioinformatics 2005 6 31 15713233 10.1186/1471-2105-6-31
Dermitzakis ET Reymond A Lyle R Scamuffa N Ucla C Deutsch S Stevenson BJ Flegel V Bucher P Jongeneel CV Antonarakis SE Numerous potentially functional but non-genic conserved sequences on human chromosome 21 Nature 2002 420 578 82 12466853 10.1038/nature01251
Dermitzakis ET Reymond A Scamuffa N Ucla C Kirkness E Rossier C Antonarakis SE Evolutionary discrimination of mammalian conserved non-genic sequences (CNGs) Science 2003 302 1033 5 14526086 10.1126/science.1087047
Reymond A Camargo AA Deutsch S Stevenson BJ Parmigiani RB Ucla C Bettoni F Rossier C Lyle R Guipponi M de Souza S Iseli C Jongeneel CV Bucher P Simpson AJ Antonarakis SE Nineteen additional unpredicted transcripts from human chromosome 21 Genomics 2002 79 824 32 12036297 10.1006/geno.2002.6781
Scripts to generate intron assemblies from prediction sets:
Download page for SGP2 software and predictions:
Download page for TWINSCAN software and predictions:
Website for Ensembl predictions in chicken:
Website with supplementary data:
| 15927066 | PMC1174865 | CC BY | 2021-01-04 16:02:56 | no | BMC Biol. 2005 May 31; 3:15 | latin-1 | BMC Biol | 2,005 | 10.1186/1741-7007-3-15 | oa_comm |
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-561592462210.1186/1471-2407-5-56Research ArticleA new survival model for hyperthermic intraperitoneal chemotherapy (HIPEC) in tumor-bearing rats in the treatment of peritoneal carcinomatosis Pelz Joerg OW [email protected] Joerg [email protected] Werner [email protected] Thomas [email protected] Department of Surgery, University of Erlangen-Nuremberg, Germany2005 30 5 2005 5 56 56 2 8 2004 30 5 2005 Copyright © 2005 Pelz et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Cytoreduction followed by hyperthermic intraperitoneal chemotherapy (HIPEC) improves survival in patients with peritoneal carcinomatosis of colorectal origin. Animal models are important in the evaluation of new treatment modalities. The purpose of this study was to devise an experimental setting which can be routinely used for the investigation of HIPEC in peritoneal carcinomatosis.
Methods
A new peritoneal perfusion system in tumor bearing rats were tested. For this purpose CC531 colon carcinoma cells were implanted intraperitoneally in Wag/Rija rats. After 10 days of tumor growth the animals were randomized into three groups of six animals each: group 1: control (n = 6), group 2: HIPEC with mitomycin C in a concentration of 15 mg/m2 (n = 6), group III: mitomycin C i.p. as monotherapy in a concentration of 10 mg/m2 (n = 6). After 10 days, total tumor weight and the extent of tumor spread, as classified by the modified Peritoneal Cancer Index (PCI), were assessed by autopsy of the animals.
Results
No postoperative deaths were observed. Conjunctivitis, lethargy and loss of appetite were the main side effects in the HIPEC group. No severe locoregional or systemic toxity was observed. All control animals developed massive tumor growth. Tumor load was significantly reduced in the treatment group and was lowest in group II.
Conclusion
The combination of hyperthermia with MMC resulted in an increased tumoricidal effect in the rat model. The presented model provides an opportunity to study the mechanism and effect of hyperthermic intraperitoneal chemotherapy and new drugs for this treatment modality.
==== Body
Background
Peritoneal carcinomatosis (PC) represents an advanced form of cancer and is associated with a poor prognosis and quality of life. The peritoneal failure rate among patients who present with recurrence after colon cancer resection is approximately 25–35% [1]. The median survival time after manifestation of peritoneal carcinomatosis is about 6 months. If peritoneal carcinomatosis is present, normaly there is no curative treatment available for any of the tumors in the abdominal cavity.
Hyperthermic intraperitoneal chemotherapy (HIPEC) is a new treatment modality. Complete cytoreductive surgery plus HIPEC is effective in the prevention and treatment of peritoneal metastases and it should lead to long-term survival for serosa-invasive carcinoma patients [2]. The efficacy of intraperitoneal (i.p.) chemotherapy is synergistically by hyperthermia. Although the technique of hyperthermic intraperitoneal chemoperfusion in humans has been employed in cancer therapy, controlled studies using this approach are rare. Intraperitoneal chemotherapy results in a high local drug concentration with less systemic exposure compared to conventional i.v. drug administration [3,4]. Peritoneal lavage with mitomycin C (MMC) results in a drug exposure of the peritoneal surface that is 20 times higher than elsewhere in the body [5,6]. The additional toxic effect of MMC at temperatures higher than 39°C has been demonstrated both in animals and in in vitro models [7,8].
One disadvantage of HIPEC is the limited penetration depth of cytostatic drugs into the tissue [4]. Thus, HIPEC must be combined with cytoreductive surgery. However, complex surgical techniques in an experimental setting should be comparable to the clinical situation. We therefore developed a miniaturised animal model and herewith report the details of the design and its efficiency.
Methods
Animals
18 inbred male pathogen-free WAG/Rija rats weighing 200 to 240 g, obtained from Charles River, Sulzfeld, Germany, were used in this study. The animals were kept in individual cages during the experiment and 12 hours of light per day. They were fed a standard laboratory diet and tap water ad libidum. Maintenance and care of all experimental animals were carried out according to the guidelines of the local responsible Animal Protection Commission and carried out in compliance with national guidelines (National Institute of Health for Use of Laboratory Animals).
Tumor
The tumor cell line used was a moderately differentiated adenocarcinoma of the rat colon (tumor cell line CC531; 1,2-dimithylhydrazine-induced [9]) obtained from the German Cancer Research Centre, Heidelberg, Germany.
Intraperitoneal tumor application was performed with a tumor suspension produced in vitro. The tumor cell line was cultivated at 37°C and 5% CO2 in monolayer cultures in an incubator in 20 ml complete medium (RPMI 1640 [Gibco, Life Technologies, Eggenstein, Germany], 10% heat-inactivated fetal bovine serum [Seromed, Biochrom, Berlin, Germany] and 1% Pen/Strep [Seromed]). After three days, cells were washed twice with phosphate buffered saline (PBS) and were detached with 3 ml trypsin (0.25 %). Trypsin was deactivated by adding the complete medium. After centrifugation, washing and re-suspension with PBS, vitality was evaluated in a Bürker hematocytometer after the addition of trypan blue. After vital counting, the suspension had a density of 2,5 × 106 vital cells/200 μl suspension (after centrifugation and re-suspension in PBS) before being injected into the animals. To control for possible mutations of cell lines, only cultures that had undergone less than 10 passages were used in the experiments.
The tumor cells were implanted via laparotomy performed under general anesthesia. The rats were anaesthetized by using isoflurane. The abdomen was shaved and cleaned with 70% alcohol. The laparotomy was performed using a lower middle incision of 6 cm. The tumor cell suspension was injected under the capsule of the peritoneal surface in the right upper side of the abdomen. Hence, tumor size and localisation were comparable to the human situation, especially after extensive cytoreductive surgery (Fig 3).
The animals were randomised into three groups of six animals each: group 1: control (n = 6), and group 2: HIPEC with MMC in a concentration of 15 mg/m2 (n = 6), group 3: MMC i.p. in a concentration of 10 mg/m2 (n = 6).
The equipment consisted of a miniature heat exchanger and a roller pump. The peritoneal perfusate was warmed in the heat exchanger (which consists of two outer silicone tubes with diameter of 15 mm for the in- and out-flowing water (Fig. 1)). The warmed perfusate was driven by a roller pump with two synchronously running pump-heads on a single axis for the inflow and the outflow lines (Masterflex®).
The intraperitoneal temperature was maintained between 40.5 and 41.5°C. Perfusion was performed over 90 minutes after the perfusion fluid had reached the required temperature. In group II, mitomycin C was added to the perfusate in three divided doses at 30 min intervals in a drug concentration of 15 mg/m2. The body surface of the animals was calculated according the formula (A(m2) = mk0,425 × 1K0,725/139.315). The first dose contained 50% and the following administrations 25% of the total dose.
For temperature measurement during perfusion, a nickel-chrome-nickel thermocouple of 0.6 mm in diameter (Standard Integrated Thermocouple Thermocoax, Phillips, Hamburg, Germany) was placed near the macroscopic tumor margin. Temperature was also monitored in the rectal cavity. The thermocouple was calibrated before use in a high-precision water bath. Baseline temperature was recorded for 5 minutes before treatment. Temperature was continuously measured during application. In group I (control) and group III (MMC only) temperature measurement was not performed.
After the perfusion the perfusate was removed and the abdomen was irrigated with saline for 10 minutes. Thereafter the abdomen was closed in two layers.
Evaluation
All animals were kept in individual cages. The rats were weighed and inspected for side effects (lethargy, loss of appetite, fatigue syndrome, wound infection) daily. The animals were kept under standard conditions and were euthanised by an overdose of anesthetic and cervical dislocation 10 days after treatment.
The animals were autopsied and peritoneal carcinomatosis was evaluated qualitatively and quantitatively. The tumor nodules were counted macroscopically.
Tumor response
Tumor response was graded according to the classification system as described by Steller et al. [10]. The scoring ranged from 0 to 5 and was performed by 2 independent observers. A score of 0 meant that there was no tumor growth; a score of 1 indicated an estimated tumor diameter less than 0.5 cm; a score of 2, a tumor diameter between 0.5 and 1 cm; a score of 3, a tumor diameter between 1 and 2 cm; a score of 4, a tumor diameter between 2 and 3 cm; and a score of 5, a tumor diameter of more than 3 cm.
Statistical Analysis
Data were analyzed using SPSS/PC+ statistical software. The mean scores were calculated. The significance of differences was assessed by the Kruskal-Wallis-test. A p value < 0.05 was considered to be significant.
Results
All animals survived the operative procedure and could be evaluated 10 days following surgery. No unexpected death occured after surgery.
Perfusion characteristics
Intraoperatively the required tissue temperature was reached within 9–11 minutes in the HIPEC group. Stable temperatures were then maintained for a further 90 minutes with a mean of 40.4°C ± 0.5. The volume needed to fill the perfusion circuit was about 250 ml. To reach tissue temperature the perfusion fluid was maintained at 45,3°C ± 2.3. The calculated body surface of the animals was 0.03–0.04 m2.
The course of intraperitoneal and rectal temperatures are presented in Fig. 2.
Duration of the procedure
The operation time, including the time needed to install the perfusion circuit, varied from 112 to 135 min (mean 121) in group II.
Body weight
Postoperatively body weight decreased in all groups during the first 5 days with a maximum decrease of 5 per cent. In the following 3 days, body weight recovered up to 108 percent of the pre-operative level. No significant differences were found between the experimental groups.
Clinical appearance
Conjunctivitis, lethargy and loss of appetite were the main side effects observed in each of the treatment groups. These were particulary pronounced in the HIPEC group. After 10 days 4/6 animals in group II had minor bleeding in the peritoneum. Moderate toxic reactions of the peritoneum were found in two animals in group III (MMC i.p.). Control animals had no side effects in the gastrointestinal tract.
Macroscopic tumor growth
In group I, all animals developed extensive intraperitoneal tumor growth. The median tumor weight was 8,1 ± 3,4 g. In group II and III median tumor weight was 1,8 ± 0,9 g and 5,7 ± 2,4 g. The cancer indices were significantly lower in group III (2,6) than in group I (3,5). The lowest tumor load was observed in group II (1,4). Tumor nodules in group II were 4 ± 2. Tumor nodules in group II were significant more after 10 days (35 ± 12) The tumor scores are given in table 1.
In group I, liver metastases were observed in two of the rats.
Discussion
Hyperthermic intraperitoneal chemotherapy (HIPEC) with mitomycin C has been applied following cytoreductive surgery for various peritoneal surface malignancies. Spratt et al. first performed HIPEC in a patient with pseudomyxoma peritonei [11]. A significant survival benefit has been shown for HIPEC when compared to systemic chemotherapy alone [4]. The complete cytoreductive surgery is the most important prognostic factor. Incomplete cytoreduction results in limited survival [12,13].
The goal of this study was to investigate the feasibility of HIPEC surviving tumor-bearing rats model and to investigate the efficacy of HIPEC applied to a single tumor growing intraperitoneally.
Hyperthermia and drugs used during HIPEC procedures have a limited penetration depth [14,4]. Therefore, HIPEC can only be effective in patients with minimal residual disease after extensive cytoreductive surgery. In the majority of related animal studies, tumor cells were injected directly into the abdominal cavity. This procedure results in diffuse growth of peritoneal carcinomatosis. To simulate cytoreductive surgery in our animal model, the CC531 colon carcinoma was implanted in an intraperitonael fat pad as described by Veenhuizen et al. [15]. Using this technique, we prevented diffuse tumor spread in the abdominal cavity. This technique may better simulate the clinical situation after de-bulking. The development of the tumors was 100 %. In contrast, Los described that only 80% of rats developed carcinomatosis after i.p. inoculation [3,4].
HIPEC could be conducted without difficulty in all cases. None of the animals died during perfusion. During and after 10 days following intervention, no serious complications were observed. In contrast to the control group (no treatment), there were 3 cases of mild side effects in group II. However, all of these side effects were completely reversible after a few days. During the procedure the abdominal cavity was kept open, allowing the abdominal perfusate to be stirred, resulting in a more homogeneous distribution of heated drug solution throughout the entire abdominal cavity.
The fact that an average temperature (of the perfusate) of 45,6°C was required in order to ensure the necessary intraabdominal temperature of 40,5°C, demonstrates the high degree of cooling that occurred across the large surface area. This question the feasibility of a closed abdomen procedure, which has been both postulated and clinically practised. An additional benefit of such a procedure would be a reduction of the potential MMC exposure risk for the operating staff. However, a major concern associated with a closed abdominal procedure is the inhomogenous distribution of the cytostatic drug and the intraabdominal temperature, because the abdominal peritoneal perfusate cannot be manually stirred as in open perfusion. A homogenous distribution of temperature is reliably assured by the open approach only, together with measurement of the temperature directely at the tumor, i.e. the target site.
MMC was chosen as the chemotherapeutic agent because of its known direct cytotoxic effect in the treatment of colorectal cancer and the thermal enhancement of its activity [16]. Heated intraperitoneal MMC is used in the clinical HIPEC setting with a concentration of 30–35 mg/m2 [17]. This concentration is 50 % higher than the maximum tolerated intravenous dose of MMC (20 mg/m2) because HIPEC results in decreased drug absorption from the perfusate (approx. 70%) [5]. Compared to the clinical situation, the MMC ratio in our experimental trial was 1,5:1 (HIPEC versus i.p. therapy). In animal models, MMC proved highly potent in the experimental setting and completely prevented intraperitoneal tumor growth in the maximal tolerated dosis of 20 mg/m2 i.p. [18]. In comparison, the MMC concentrations chosen for the study were lower than the maximal tolerated dose, but also administered in a ratio of 1,5:1 (15 mg/m2 HIPEC versus 10 mg/m2 i.p.). Our results showed that MMC concentration of 10 mg/m2 was not effective enough to completely inhibit tumor growth in group III (therapy of solid tumor) in contrast to the study of HRIBASCHEK et al. (preventing of peritoneal carcinomatosis)(18).
Primary results regarding tumor response in this HIPEC animal model confirm the clinical data collected thus far. In comparison to the control group, a significant delay of tumor growth was demonstrated in group III. This shows that MMC, as a monotherapy, has a clear influence on tumor response. The fact that tumor growth was merely slowed (but not arrested) demonstrates that the chosen MMC concentration was not too high.
Conclusion
By the presented model, the procedure of HIPEC could be evaluated qualitatively and quantitatively. The direct effect of HIPEC could be recorded macroscopically. Long-term observations can be conducted without difficulty, assuming acceptable mortality and morbidity rates.
In conclusion, intraperitoneal macroscopic tumor growth was reduced after hyperthermic intraperitoneal chemoperfusion in comparison with MMC application alone. However, further experimental in vitro and in vivo studies are required to better charaterize the exact mechanism of HIPEC and to set up objective parameters for optimization of the procedure.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
JP carried out the treatment and drafted the manuscript.
JD carried out the treatment.
WH participiated in the design of the study
TM participiated in the design and coordination of the study.
All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study was supported by grants from ELAN (Erlanger Leistungsbezogene Anschubfinanzierung und Leistungsförderung)
Figures and Tables
Figure 1 Diagram of the experimental HIPEC circuit.
Figure 2 The temperature course during HIPEC in the abdomen and in the rectal cavity.
Figure 3 Limited tumor implantation. Injection of cell suspension under the capsule of the peritoneal surface preventing diffuse peritoneal carcinomatosis (a). Isolated tumour nodule without diffuse carcinoses 10 days after implantation (b). Representative gross sections of peritoneal carcinomatosis in group I (control), II (HIPEC) and group III (MMC only) 10 days postoperatively.
Figure 4 Diffuse intraperitoneal tumor spread in the control group (group I).
Figure 5 Small tumor load in HIPEC group (group II).
Figure 6 Tumor growth after MMC application (group III).
Table 1 Objective tumor response in groups 1 (control; n = 6), 2 (HIPEC; n = 6), and 3 (MMC only; n = 6) 10 days after intervention. * P < 0.05 versus groups 1 and 3; § P < 0.05 versus group 1 (Kruskal-Wallis test)
Group I Group II Group III
Tumor weight (g) 8,1 ± 3,4 1,8 ± 0,9 * 5,7 ± 2,4 §
Tumor nodules 35 ± 12 4 ± 2 * 16 ± 10 §
Cancer index 3,5 1,4 * 2,6
Ascites 2/6 0/6 0/6
Clinical CR 0/6 4/6 1/6
==== Refs
Minsky BD Mies C Rich TA Recht A Chaffey JT Potentially curative surgery of colon cancer: patterns of failure and survival J Clin Oncol 6 106 118 2826711
Verwaal VJ van Ruth S de Bree E van Slooten G van Tinteren H Boot H Zoetmulder FAN Randomized trial of cytoreduction and hyperthermic intra-peritoneal chemotherapy versus systemic chemotherapy and palliative surgery in patients with peritoneal carcinomatosis of colorectal cancer J Clin Oncol 2003 12 3737 3743 10.1200/JCO.2003.04.187
Los G Mutsaers PH van der Vijgh WJ Baldew GS de Graaf PW McVie JG Direct diffusion of cis-diamminedichloroplatinum(II) in intraperitoneal rat tumors after intraperitoneal chemotherapy: a comparison with systemic chemotherapy Cancer Res 49 3380 4 1989 Jun 15 2720692
Los G Sminia P Wondergem J Mutsaers PH Havemen J ten Bokkel Huinink D Smals O Gonzalez-Gonzalez D McVie JG Optimisation of intraperitoneal cisplatin therapy with regional hyperthermia in rats Eur J Cancer 1991 27 472 7 1827723
Fernandez-Trigo V Stuart OA Stephens AD Hoover LD Sugarbaker PH Surgically directed chemotherapy: heated intraperitoneal lavage with mitomycin C Cancer Treat Res 1996 81 51 61 8834575
Witkamp AJ de Bree E Kaag MM Boot H Beijnen JH van Slooten GW Extensive cytoreductive surgery followed by intra-operative hyperthermic intraperitoneal chemotherapy with mitomycin-C in patients with peritoneal carcinomatosis of colorectal origin Eur J Cancer 2001 37 979 84 11334722 10.1016/S0959-8049(01)00058-2
Cavaliere R Ciocatto EC Giovanella BC Heidelberger C Johnson RO Margottini M Selective heat sensitivity of cancer cells. Biochemical and clinical studies Cancer 1967 20 1351 81 4292219
Isacoff WH Borud K Chemotherapy for the treatment of patients with metastatic colorectal cancer: an overview World J Surg 1997 21 748 62 9276707 10.1007/s002689900301
Marquet RC Westbroek DL Jeekel J Interferon treatment of a transplantable colon adenocarcinoma, importance of tumor site Br J Cancer 1984 33 698 692
Steller EP Ottow RT Matthews W Sugarbaker PH Rosenberg SH Recombinant interleukin-2 and adoptively transferred lymphokine-activated killer cells in the treatment of experimental peritoneal carcinomatosis Surg Forum 1985 36 390 2
Spratt JS Adcock RA Muskovin M Sherrilll W McKeown J Clinical delivery system for intraperitoneal hyperthermic chemotherapy Cancer Res 1980 40 256 60 6766084
Glehen O Kwiatkowski F Sugarbaker PH Elias D Levine EA De Simone M Barone R Yonemura Y Cavaliere F Quenet F Gutman M Tentes AA Lorimier G Bernard JL Bereder JM Porcheron J Gomez-Portilla A Shen P Deraco M Rat P Cytoreductive surgery combined with perioperative intraperitoneal chemotherapy for the management of peritoneal carcinomatosis from colorectal cancer: a multi-institutional study J Clin Oncol 22 3284 92 2004 Aug 15 15310771 10.1200/JCO.2004.10.012
Glehen O Mohamed F Sugarbaker PH Incomplete cytoreduction in 174 patients with peritoneal carcinomatosis from appendiceal malignancy Ann Surg 2004 240 278 85 15273552 10.1097/01.sla.0000133183.15705.71
van Ruth S Verwaal VJ Hart AA van Slooten GW Zoetmulder FA Heat penetration in locally applied hyperthermia in the abdomen during intra-postoperative hyperthermic intraperitoneal chemotherapy Anticancer Res 2003 23 1501 8 12820416
Veenhuizen RB Marijnissen JP Kenemans P Ruevekamp-Helmers MC 't Mannetje LW Helmerhorst TJ Stewart FA Intraperitoneal photodynamic therapy of the rat CC531 adenocarcinoma Br J Cancer 1996 73 1387 92 8645584
Storm FK Clinical hyperthermia and chemotherapy. Radiol Clin N America 1989 27 621 627
Verwaal VJ van Tinteren H van Ruth S Zoetmulder FAN Toxicity of Cytoreductive Surgery and Hyperthermic Intra-Peritoneal Chemotherapy J Surg Oncol 2004 85 61 67 14755505 10.1002/jso.20013
Hribaschek A Kuhn R Pross M Lippert H Halangk W Boltze C Ridwelski K Prophylaxis of peritoneal carcinomatosis in experimental investigations Int J Colorectal Dis 2001 16 340 345 11686534 10.1007/s003840100321
| 15924622 | PMC1174866 | CC BY | 2021-01-04 16:03:03 | no | BMC Cancer. 2005 May 30; 5:56 | utf-8 | BMC Cancer | 2,005 | 10.1186/1471-2407-5-56 | oa_comm |
==== Front
BMC Fam PractBMC Family Practice1471-2296BioMed Central London 1471-2296-6-221594148310.1186/1471-2296-6-22Research ArticlePhysicians perceived usefulness of high-cost diagnostic imaging studies: results of a referral study in a German medical quality network Schneider Antonius [email protected] Thomas [email protected] Michel [email protected] Joachim [email protected] Department of General Practice and Health Services Research, University of Heidelberg2005 7 6 2005 6 22 22 28 10 2004 7 6 2005 Copyright © 2005 Schneider et al; licensee BioMed Central Ltd.2005Schneider et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Medical and technological progress has led to increased numbers of diagnostic tests, some of them inducing high financial costs. In Germany, high-cost diagnostic imaging is performed by a medical specialist after referral by a general practitioner (GP) or specialist in primary care. The aim of this study was to evaluate the physicians' perceived usefulness of high-cost diagnostic imaging in patients with different clinical conditions.
Methods
Thirty-four GPs, one neurologist and one orthopaedic specialist in ambulatory care from a Medical Quality Network documented 234 referrals concerning 97 MRIs, 96 CTs-scan and 41 intracardiac catheters in a three month period. After having received the test results, they indicated if these were useful for diagnosis and treatment of the patient.
Results
The physicians' perceived usefulness of tests was lowest in suspected cerebral disease (40% of test results were seen as useful), cervical spine problems (64%) and unexplained abdominal complaints (67%). The perceived usefulness was highest in musculoskeletal symptoms (94%) and second best in cardiological diseases (82%).
Conclusion
The perceived usefulness of high-cost diagnostic imaging was lower in unexplained complaints than in specific diseases. Interventions to improve the effectiveness and efficiency of test ordering should focus on clinical decision making in conditions where GPs perceived low usefulness.
==== Body
Background
The continuous medical and technological progress has led to a rising use of high tech diagnostic tests which are often expensive. For that reason many efforts have been undertaken to enhance the effectiveness and efficiency of referrals for diagnostic tests. Studies have pointed out a wide range of reasons for referrals including patients demand for extensive diagnostics. For instance, referral patterns were related to the physicians' attitudes to their role [1] and to the interaction between the physician and patient [2]. Also the social context seems to have a high influence for referral for further diagnostics [3]. Nevertheless, the variation of referral rates remains largely unexplained [4].
Concerning expensive diagnostic procedures, Robling et al. found that the referral for diagnosis with MRI had biomedical, personal and contextual reasons [5]. Particularly "vague complaints" were related to a high likelihood for test ordering. A large observational, cross-sectional study revealed the influence of patients' expectations about test ordering for further diagnostic to clarify vague symptoms [6]. As the variation of referrals seems to be related with high expenditures of health care systems, health policy makers are seeking improvement in this area [7].
In general, there is no formal gatekeeping role for the GP in the German health care system and ambulatory care comprises almost all specialties [8]. However, especially for expensive or invasive diagnostic procedures, referrals are requested from a GP or a specialist in ambulatory care. Normally, the GP or specialist in primary care decides on the indication and performance of MRI or CT scan and then refers the patients directly to the radiologist, who works in a community-based practice. The indication for an intracardiac catheter is set by a cardiologist at the hospital or in a community-based practice in most cases. For routine diagnostic or follow up, but not for emergency case, a referral from a GP is formally required.
This prospective observational study examined referrals for test ordering with respect to the physicians' perceived usefulness of high-cost diagnostic imaging for diagnosis and treatment in general practice. Our aim was to identify clinical conditions where GPs and specialists in ambulatory care perceived limited relevance of the tests so that targeted strategies could be planned to support medical decision making in these conditions. Because of the impact on the financial resources the study focused on high-cost imaging, including MRI scan, CT-scan and intracardiac catheter.
Methods
Design
The project was designed as a prospective observational study. 34 general practitioners of a medical network, one neurologist and one orthopaedic specialist in ambulatory care and one hospital of the region participated between January and March 2000. In this time period all referrals concerning MRI scan, CT-scan or intracardiac catheter from a total of 234 patients were documented by the physicians. The participating physicians were members of the Medical Quality Network Ried in the southern region of Hessia, Germany. The overall aim of this network is the implementation of continuous quality improvement in primary care. For instance, previous activities focused on the introduction of a patient-held medical record [9].
Measures
The participating physicians were instructed to document each referral for MRI, CT-scan or intracardiac catheter. The documentation included reason for referral, clinical symptoms and previously established diagnoses of each patient.
The main outcome measure was the assessment by the physician if the result of the diagnostic procedure represents an important contribution to the diagnostic process. This could be a confirmation or negotiation of the estimated diagnoses for providing optimal treatment respectively to reassure that there is no dangerous health problem. To clarify this concept, a workshop with the participating physicians at the start of the study was done, in which the physicians were instructed to use this operationalization. The statement could be given as "useful" or "not useful" after receiving the test result. The estimation of the usefulness was usually made on the same day, when the physician received the referral letter. In Germany, the specialist who performed the diagnostics is obliged to send such a letter to the referring physician where the test results are listed in detail. For every documentation of the referral process and estimation of its usefulness, the physicians received a small financial incentive.
Analysis
Every reason for a diagnostic referral was documented in a free text. The clinical symptoms were clustered into six different subgroups after analysis of the whole spectrum of referrals: cervical and lumbar spine, internal diseases (without cardiac problems), nervous system, musculoskeletal system, cardiac system and others. Baseline data were compared descriptively. The Chi-Square-Test was used for testing the relation between gender of patients and perceived usefulness statistically. The t-test was performed to calculate the relation between perceived usefulness (as independent variable) and age (as dependent variable). Statistical procedures and analysis of the data were done with SPSS 11.0.
Results
Referrals for test ordering from 132 women and 100 men were documented (table 1). In sum 80 % of the diagnostic referrals of the women were assessed as useful, while 75 % were assessed as useful for the male patients. The Chi-Square-Test showed that this difference is not statistically significant (p = 0.796). There was no association between patient age and perceived usefulness of the test (t-test, p = 0.307).
Table 1 Characteristics of Patients: (in brackets: useful / not useful / no assessment)
Sex N Age CT MRI Cath. All useful in % (95% CI)
Female 132 54.8 + 16.0 min 17; max 85 51 68 13 (106 / 23 / 3) 80,3 (72.7–86.2)
Male 100 56.0 + 16,6 min 13; max 85 44 28 28 (75 / 18 / 7) 75,0 (65.7–82.5)
No declaration 2 42 y, 51 y 1 1 0 (2 / 0 / 0) 100
Amount 234 55.3 + 16.2 min 13; max 85 96 97 41 (183 / 41 / 10) 78,2 (72.5–83.0)
Table 2 describes the perceived usefulness of tests in different clinical conditions. In cerebral diseases, MRI and CT-scans were perceived to be useful for diagnosis in apoplexia (100% resp. 83%, in sum 89%) and encephalitis disseminata (ED) (75%). A high amount of MRI was ordered by the neurologist to rule out a cerebral tumour by reason of persistent headache (perceived usefulness 89%). The usefulness was perceived to be low in unclear cerebral symptoms (suspected disease), mostly due to persistent dizziness, ataxia, tinnitus and other vague complaints (37% resp. 50%, in sum 40%).
Table 2 Performed Diagnostics: (in brackets: useful / not useful / no assessment)
Area GP / Specialist Indication CT MRI useful in % (95% CI)
Cerebral (n = 61) GP Apoplexia (3) 3 (3 / 0 / 0) 0 100
GP Cerebral Tumour (3) 1 (1 / 0 / 0) 2 (2 / 0 / 0) 100
GP ED* (4) 1 (1 / 0 / 0) 3 (2 / 1 / 0) 75.0
GP Suspected Disease (11) 5 (2 / 2 / 1) 6 (2 / 3 / 1) 36.4
Neuro. Apoplexia (6) 2 (1 / 1 / 0) 4 (4 / 0 / 0) 83.3
Neuro. Rule out tumour (27) 8 (6 / 2 / 0) 19 (18 / 1/ 0) 88.9
Neuro. Trauma / SVT** / Hydrozephalus (3) 1 (1 /0 / 0) 2 (2 /0 / 0) 100
Neuro. Suspected Disease (4) 1 (1 /0 / 0) 3 (1 /2 / 0) 50.0
Total Amount (61) 22 (16 / 5 / 1) 39 (31 / 7 / 1) 77.0 (65.1–85.8)
Vertebra (n = 71) GP Cervical spine (12) 5 (2/3/0) 7 (3 /2 / 2) 41.6
GP Lumbar spine (26) 20 (17/2/1) 6 (5 /0 / 1) 88.5
Orthop. Cervical spine (8) 3 (3 /0 / 0) 5 (3 /2 / 0) 75.0
Orthop. Lumbar spine (9) 6 (5 /1 / 0) 3 (2 /1 / 0) 77.8
Neuro. Cervical spine (8) 1 (1 /0 / 0) 7 (6 /1 / 0) 87.5
Neuro. Lumbar spine (8) 2 (2 /0 / 0) 6 (6 /0 / 0) 100
Total Amount (71) 37 (30 / 6 / 1) 34 (25 / 6 / 3) 77.5 (66.5–85.6)
Musculoskeletal Symptoms (n = 18) GP Knee (10) 0 10 (9 / 1 / 0) 90.0
GP Shoulder (3) 0 3 (3 / 0 / 0) 100
GP Elbow (1) 0 1 (1 / 0 / 0) 100
Orthop. Knee (1) 1 (1 / 0 / 0) 0 100
Orthop. Shoulder (2) 0 2 (2 / 0 / 0) 100
Orthop. Upper ankle (1) 0 1 (1 / 0 / 0) 100
Total Amount (18) 1 (1 / 0 / 0) 17 (16 / 1 / 0) 94.4 (74.2–99.0)
Area Indication CT MRI useful in % (95% CI)
Internal (n = 27) Abdomina l complaints (15) 15 (10 / 5 / 0) 0 66.7
Pulmonary (7) 6 (5 / 0 / 1) 1 (1 / 0 / 0) 85.7
Urogenital (2) 2 (2 / 0 / 0) 0 100
Other tumours (3) 2 (1 / 1 / 0) 1 (1 / 0 / 0) 66.7
Total Amount (27) 25 (18 / 6 / 1) 2 (2 / 0 / 0) 74.1 (55.3–86.8)
Others (n = 4) Chron. Sinusitis 1 (1 / 0 / 0) 0 100
Lymphhaemangioma axillae 0 1 (1 / 0 / 0) 100
Pain after Appendectomy 1 (1 / 0 / 0) 0 100
Thoracic pain of unknown origin 1 (1 / 0 / 0) 0 100
Total Amount(4) 3 (3 / 0 / 0) 1 (1 / 0 / 0) 100
Area Indication Intracardiac catheter MRI useful in % (95% CI)
Cardiac (n = 49) Coronary artery disease 48 (39 / 6 / 3) 0 81,3 (68.1–89.8)
Cardiac effusion 0 1 (1 / 0 / 0) 100
* ED = Encephalitits disseminate; ** SVT = sinus vein thrombosis GP = general practitioner, Neuro. = neurologist, Orthop. = orthopaedic specialist
The usefulness of tests for cervical spine problems was less than 42% for the GPs, for the orthopaedic specialist 88% and for the neurologist 75% (in sum 64%). In opposite to that, both the GPs and specialists perceived MRI and CT-scan as useful for lumbar spine problems in about 88–100%. Problems with the knees, shoulder, elbows and foot ankles had the best result regarding usefulness of test ordering (in sum 94 %).
In cases of abdominal complaints, where also cancer was suspected as a reason, 67% of the diagnostic referrals were perceived as useful. "Unexplained abdominal symptom" was the most frequent reason for ordering an abdominal CT-scan (seven out of 15 cases). The usefulness of tests for pulmonary diseases, mostly for suspected carcinoma, was higher, but the absolute number of cases was limited. The group of "other tumours" included two cases of breast cancer and one case of oesophageal cancer. In cases of cardiac problems most tests referred to suspected coronary artery disease (CAD), using intracardiac catheters. One MRI was done because of a suspected pericardial effusion. In sum, the cardiological diagnostics were perceived as useful in 82%.
Discussion
The aim of our study was to examine to what extent MRI, CT-Scan or intracardial catheter were perceived as useful for making diagnoses and treatment decisions from the point of view of the physicians. The usefulness was most limited in problems related to cervical spine, unclear cerebral symptoms and abdominal pain.
In general, there is a limited usefulness of MRI and CT-scan for diagnostics in cervical spine problems [10,11]. The challenge in general practice concerning the management of patients with cervical spine problems is the already known psychological co-morbidity in this disease [12]. Due to the psychological strain of the patients related to a suspected herniated disc or nerve inflammation, there is a high pressure to clarify the reason of pain or to rule out a severe disease. As the test result is often negative whilst the pain is persisting, physicians and patients could be both disappointed, thus leading to low perceived usefulness of high-cost imaging.
In opposite to that, the diagnostic procedures in low back pain were perceived as useful in most cases. However, clinical research showed little correlation between morphology and complaints in low back pain [13]. The value for physicians in these cases could be the possibility to reassure the patient that there is no "dangerous problem". The optimistic estimation could reflect the satisfaction with care and with the course of the disease, which is normally inherent benign as the moderate lumbago is often self limiting.
The comparatively low usefulness of CT-scans in abdominal complaints may express the complexity of diagnostics in these cases. Many patients with unexplained abdominal symptoms show a high psychological co-morbidity [14], which could influence the doctor-patient-interaction and lead to a higher rate of diagnostic tests [6]. The same may be the case in unexplained neurological symptoms, which are often related with anxiety or depressive disorder [15]. Even if a perceived usefulness of about 64% could be seen as high, it should be mentioned that MRI or CT as most expensive tests are often done at the end of diagnostic procedures. The value of these tests should also be considered critically as they may induce unnecessary somatisation and medicalisation of the patients. To summarize, these results underline the difficulties in clinical diagnostics, particularly regarding unexplained complaints, which are frequent causes for consultation [16]. The problem behind is the high psychological co-morbidity, which is associated with prolonged illness behaviour and provocation of high usage of diagnostic procedures with an additional risk to harm the patients [17,18].
The results concerning the tests for cardiac diseases and problems with the musculoskeletal seem to be more informative. Intracardiac catheter, MRI or CT-scan led in most of these cases to an effective medical decision making, so it was regarded as helpful and appropriate by the referring GP.
A limitation of our study was that we investigated a subjective estimation of the physicians for the usefulness of the diagnostic referral but not the appropriateness judged by an external observer. For example, in the field of cardiology the managing of coronary artery disease in Germany has been criticised [19]. This indicates that the appropriateness for cardiological diagnostics could be weak despite the optimistic estimation in our study. Another limitation is that we had no total control on the response rate. But as the physicians received financial incentives for every documentation we assume that the response rate was quite high. The participating physicans are members of a medical quality network with a high motivation for participating in quality improvement projects. It must be questioned if the results are representative or even if there could be an overestimation of the usefulness regarding test ordering. The number of clinical conditions was large, consequently the subdivision into six categories led to heterogeneous groups with comparatively high confidence intervals. Further research should confirm the results of this explorative study.
Nevertheless, our study suggests that quality improvement should focus on patients with unexplained complaints to avoid expensive, unnecessary or dangerous diagnostic investigations. A starting point for dealing with these problems could be an analysis together with the network of physicians and a subsequent implementation of evidence based guidelines, accompanied by training in risk communication with difficult patients. This implementation of change should be done in a multifaceted strategy using guidelines, feedback and social interaction [20,21].
Conclusion
The perceived usefulness of high-cost diagnostic imaging was lower in unexplained complaints than in specific diseases. Interventions to improve the effectiveness and efficiency of test ordering should focus on clinical decision making in conditions where GPS perceived low usefulness. Further research is necessary to identify patient factors underlying unexplained symptoms and to find methods to improve decision making regarding test ordering.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
AS performed the statistical analysis and wrote the manuscript. TR supported in writing the manuscript. MW supported in statistics and writing. JS designed and supervised the study. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We want to thank the medical doctors of the Medical Quality Network Ried for participating and documentation.
==== Refs
Roland M Grimshaw J Grol R Shanks D Johnson A Russell I Taylor R Do general practitioner attitudes and characteristics of their practices explain patterns of specialist referral? Eur J Gen Pract 1997 3 143 147
Evans A A study of the referral decision in general practice Fam Pract 1993 10 104 110 8359598
Dempsey OP Bekker HL 'Heads you win, tails I lose': a critical incident study of GPs' decisions about emergency admission referrals Fam Pract 2002 19 611 616 12429663 10.1093/fampra/19.6.611
O'Donnell CA Variation in GP referral rates: what can we learn from the literature? Fam Pract 2000 17 462 471 11120716 10.1093/fampra/17.6.462
Robling M Kinnersley P Houston H Hourihan M Cohen D Hale J An exploration of GPs' use of MRI: a critical incident study Fam Pract 1998 15 236 243 9694181 10.1093/fampra/15.3.236
van Der Weijden T van Velsen M Dinant GJ van Hasselt CM Grol R Unexplained complaints in general practice: prevalence, patients' expectations, and professionals' test-ordering behavior Med Decis Making 2003 23 226 231 12809320 10.1177/0272989X03023003004
Coulter A Managing demand at the interface between primary and secondary care BMJ 1998 316 1974 1976 9641944
Busse R Riesberg A Health Care in Transition: Germany European Observatory on Health Systems and Policies 2004 http://www.euro.who.int/Document/E85472.pdf
Szecsenyi J Klingenberg A Pelz J Magdeburg K Evaluation of a patient-held medical record by patients-results from the Ried Medical Quality Network Z Arztl Fortbild Qualitatssich 2001 95 407 412 11503559
Peterson C Bolton J Wood AR Humphreys BK A cross-sectional study correlating degeneration of the cervical spine with disability and pain in United Kingdom patients Spine 2003 28 129 133 12544928 10.1097/00007632-200301150-00007
Sengupta DK Kirollos R Findlay GF Smith ET Pearson JC Pigott T The value of MR imaging in differentiating between hard and soft cervical disc disease: a comparison with intraoperative findings Eur Spine J 1999 8 199 204 10413345 10.1007/s005860050157
van der WD Croft P Penninx B Neck and upper limb pain: more pain is associated with psychological distress and consultation rate in primary care J Rheumatol 2002 29 564 569 11908574
Borenstein DG O'Mara JWJ Boden SD Lauerman WC Jacobson A Platenberg C Schellinger D Wiesel SW The value of magnetic resonance imaging of the lumbar spine to predict low-back pain in asymptomatic subjects : a seven-year follow-up study J Bone Joint Surg Am 2001 83-A 1306 1311 11568190
Janssen HA Borghouts JA Muris JW Metsemakers JF Koes BW Knottnerus JA Health status and management of chronic non-specific abdominal complaints in general practice Br J Gen Pract 2000 50 375 379 10897534
Carson AJ Ringbauer B Stone J McKenzie L Warlow C Sharpe M Do medically unexplained symptoms matter? A prospective cohort study of 300 new referrals to neurology outpatient clinics J Neurol Neurosurg Psychiatry 2000 68 207 210 10644789 10.1136/jnnp.68.2.207
Reid S Wessely S Crayford T Hotopf M Medically unexplained symptoms in frequent attenders of secondary health care: retrospective cohort study BMJ 2001 322 767 11282861 10.1136/bmj.322.7289.767
Kroenke K Patients presenting with somatic complaints: epidemiology, psychiatric comorbidity and management Int J Methods Psychiatr Res 2003 12 34 43 12830308
Roy-Byrne PP Generalized anxiety and mixed anxiety-depression: association with disability and health care utilization J Clin Psychiatry 1996 57 Suppl 7 86 91 8690701
Dissmann W de Ridder M The soft science of German cardiology Lancet 2002 359 2027 2029 12076573 10.1016/S0140-6736(02)08825-6
Verstappen WH van Der WT Sijbrandij J Smeele I Hermsen J Grimshaw J Grol RP Effect of a practice-based strategy on test ordering performance of primary care physicians: a randomized trial JAMA 2003 289 2407 2412 12746365 10.1001/jama.289.18.2407
Wensing M van Der WT Grol R Implementing guidelines and innovations in general practice: which interventions are effective? Br J Gen Pract 1998 48 991 997 9624774
| 15941483 | PMC1174867 | CC BY | 2021-01-04 16:29:13 | no | BMC Fam Pract. 2005 Jun 7; 6:22 | utf-8 | BMC Fam Pract | 2,005 | 10.1186/1471-2296-6-22 | oa_comm |
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-811592979410.1186/1471-2164-6-81Methodology ArticleGenome-wide localization of mobile elements: experimental, statistical and biological considerations Martinez-Vaz Betsy M [email protected] Yang [email protected] Wei [email protected] Arkady B [email protected] Department of Biochemistry, Molecular Biology and Biophysics and Biotechnology Institute, University of Minnesota, Saint Paul, MN 55108, USA2 Biostatistics Department, School of Public Health, University of Minnesota, Minneapolis, MN 55434, USA2005 1 6 2005 6 81 81 16 3 2005 1 6 2005 Copyright © 2005 Martinez-Vaz et al; licensee BioMed Central Ltd.2005Martinez-Vaz et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The distribution and location of insertion elements in a genome is an excellent tool to track the evolution of bacterial strains and a useful molecular marker to distinguish between closely related bacterial isolates. The information about the genomic locations of IS elements is available in public sequence databases. However, the locations of mobile elements may vary from strain to strain and within the population of an individual strain. Tools that allow de novo localization of IS elements and are independent of existing sequence information are essential to map insertion elements and advance our knowledge of the role that such elements play in gene regulation and genome plasticity in bacteria.
Results
In this study, we present an efficient and reliable method for linear mapping of mobile elements using whole-genome DNA microarrays. In addition, we describe an algorithm for analysis of microarray data that can be applied to find DNA sequences physically juxtaposed with a target sequence of interest. This approach was used to map the locations of the IS5 elements in the genome of Escherichia coli K12. All IS5 elements present in the E. coli genome known from GenBank sequence data were identified. Furthermore, previously unknown insertion sites were predicted with high sensitivity and specificity. Two variants of E. coli K-12 MG1655 within a population of this strain were predicted by our analysis. The only significant difference between these two isolates was the presence of an IS5 element upstream of the main flagella regulator, flhDC. Additional experiments confirmed this prediction and showed that these isolates were phenotypically distinct. The effect of IS5 on the transcriptional activity of motility and chemotaxis genes in the genome of E. coli strain MG1655 was examined. Comparative analysis of expression profiles revealed that the presence of IS5 results in a mild enhancement of transcription of the flagellar genes that translates into a slight increase in motility.
Conclusion
In summary, this work presents a case study of an experimental and analytical application of DNA microarrays to map insertion elements in bacteria and gains an insight into biological processes that might otherwise be overlooked by relying solely on the available genome sequence data.
==== Body
Background
Insertion elements, the simplest bacterial transposons, are short DNA sequences (700–2500 bp) carrying only genetic information related to their transposition and its regulation [1]. IS elements are capable of transposition into many sites within and between bacterial chromosomes and extra-chromosomal elements. The movement of IS elements can cause activation or silencing of adjacent genes [2]; chromosomal rearrangements such as deletions, inversions and insertions are also common consequences of IS element activity [3]. Due to diverse genetic effects associated with the activity of insertion elements, developing tools to identify and map the location of these DNA sequences in bacterial genomes is essential to advance our understanding of the role IS elements play in gene regulation and genome plasticity.
Mapping insertion elements in microbial genomes is important for several reasons. First, the distribution and location of insertion elements in a genome is a potent tool to track the evolution of a bacterial strain [4-7]. Second, IS elements are often used as molecular markers to distinguish between closely related bacterial strains. This approach is helpful in epidemiological studies in which the presence and location of a particular insertion element have been used as a marker to track the epidemiology of microbial pathogens [8,9].
Although the information about the genomic locations of IS elements is available in public sequence databases, by definition, the locations of mobile elements may vary from strain to strain and within the population of an individual strain [3], and [10]. Thus we need a tool that would not be solely dependent on the existing information about the location of insertion elements, but instead would allow de novo mapping of the sequences.
A variety of molecular techniques have been used to map insertion elements in bacteria. These include Southern hybridizations, inverse PCR, and vectorette PCR [11,12]; and [13]. Inverse PCR and Southern hybridizations are very laborious techniques that require further sample processing to determine the location of the insertion sequences. Recently, vectorette PCR has been described as rapid and efficient method to map IS elements in the E. coli genome [13]. DNA microarrays provide a powerful alternative to the gel-based techniques and allow reliable determination of relative abundances of individual RNA or DNA species in complex mixtures. Most microarray applications attempt to assess the relative abundance of individual nucleic acids species by labeling it (along with others in the mixture) directly, in sequence-independent manner [14-17] and [18]. However, the identification of neighboring sequences using microarrays relies on a sequence-dependent labeling by primer extension from a known sequence [19] and [9]. Raychaudhuri [19] developed and later applied [9] an algorithm using non-parametric discriminant analysis to predict the location of IS6110 element within the Mycobacterium tuberculosis genome from microarray data. Their algorithm requires two sets of feature training examples: insertion sites and non-insertion sites and the authors used known insertions sites across multiple experiments to generate the examples. However in most real life cases, we do not have multiple experiments and do not know the sites of insertions or non-insertion sites in advance. In this paper, we propose a simple but reliable algorithm to predict the genomic location of insertion DNA sequences based on microarray data without using the prior knowledge about the location of mobile elements.
Most applications of spotted cDNA microarrays rely on two-color competitive DNA hybridization to assess the relative abundance of nucleic acid species represented on the array [14]. Often, however, it is quite difficult to choose an adequate biologically meaningful reference for the two-color hybridization. In such cases especially, the reference serves only to compound the error in the ratio calculation. In this article we investigate the applicability of single channel hybridization for making biologically relevant inferences from microarray data.
We applied the tools presented herein to map the locations of the IS5 elements in the genome of Escherichia coli K12. We observed heterogeneity in localization of the elements within a population of the strain. The biological implications of this finding are discussed.
Results
Genome-wide mapping of IS5 elements
Individual clones of the E. coli MG1655 were obtained by plating an ATCC culture of the strain whose complete genome sequence was published in 1997 [20]. Liquid cultures of single colonies were used as frozen stocks for subsequent experiments. We carried out mapping experiments with two sets of biological replicates, where each set is represented by several independently grown colonies from two separately frozen stocks of bacteria. We designated these stocks 'A' and 'B', and the sets of biological replicates were A1, A2, A3, A4, A5 and B1, B2, B3, B4, B5, B6. We also carried out a set of control experiments designated as C1, C2 and C3, where no neighbor sequences should've been detected because purified transposon DNA was used as a template in the probe preparation. The IS mapping data presented in this paper is publicly available in the NCBI GEO database under the following accession numbers: GSE2697.
Table 1 compares the performance of different test statistics and normalizations in three experiments (control, stocks 'A' and 'B') based on the available sequence data. In this table, the neighbours from sequence data were defined as 5 upstream and 5 downstream genes of an IS5 element and the IS5 itself, there are eleven known copies of IS5 in the E. coli genome. Combining this information, we get 11 neighbours per IS5 element × 11 IS5 copies; this equals 121 neighbours based on sequence data (11*11 = 121). From this table (Table 1), we can see that rank-based statistics (rank product and median rank) perform similarly (the overlap between the lists of top genes identified by rank-product and median rank is about 90%), and are better than intensity-based statistics (Mean, SAM and t statistics). This statement is based on the observation that using rank statistics allows us to identify more neighbours in the experimental samples (stocks A and stock B) and fewer neighbours in the controls. In contrast, intensity-based statistics identified more gene neighbours for the control samples and fewer neighbours for the experimental samples. Among the intensity-based test statistics, it is hard to tell which one performs better for these data; however, none of them is better than rank-based methods. Using different normalization methods did not affect the results of rank-based analysis but led to different results than using intensity-based statistics. Consistent with our expectation, whereas only the IS5 element itself could be detected in the control experiment, IS5 elements and their neighbours can be found when genomic DNA from samples 'A' or 'B' was used as a template. The results of microarray experiment with the genomic DNA from stock 'A' were somewhat more consistent with sequence data than those obtained using DNA from stock 'B'. Based on these comparisons, we used median rank as our test statistic.
Table 1 Comparisons of different methods used to identify IS5 neighbouring genes from microarray experiments. Results obtained when different test statistics (rank product – RP, median rank – MR, Mean, SAM and t-statistics) and normalization methods (global and scale normalization) were applied to experimental data from the control, isolate 'A' and isolate 'B'. These results were also compared to existing genome sequence data. Top 50 to 100 genes were predicted as IS5 neighbours from microarray experiments. The neighbours from sequence data were defined as 5 genes upstream and 5 genes downstream of the IS5 elements and IS5 itself, so there were 121 neighbours based on sequence data. The numbers in this table represent number of genes that overlap between the microarray experiment and the sequence data. The higher the number is, the better consistency between the microarray experiment and the sequence data.
Method Top 50 Top 100
control A B control A B
Rank – based RP 13 46 31 15 69 43
MR 13 45 32 15 70 47
Global Normalization Mean 20 37 20 23 65 29
SAM 20 38 24 23 65 29
t 20 38 24 23 65 31
Scale Normalization Mean 20 44 21 23 65 23
SAM 20 40 22 23 63 29
t 20 16 24 23 30 31
We used false discovery rate plots to decide the cut-off for claiming the significant genes for median rank statistic. Figure 1 shows that the false discovery rate increases much more rapidly in the control compared to the experiment as the number of claimed positive genes increases. This observation is consistent with our hypothesis that we should find more true positive genes in experiments (including IS5 elements and their neighbours) than in controls (where only IS5 elements were expected to be detected). For the same number of claimed positive genes, the FDR for stock 'A' was somewhat lower than for 'B': when claiming top 100 genes as significant genes, the FDR for stock 'A' was about 0.02 and for stock 'B' was about 0.06. We regarded these false discovery rates (6 genes per 100) as acceptable and we set the cut-off for selecting significant genes at the top 100 genes.
Figure 1 False discovery rate for median rank statistic. False discovery rate, FDR, as a function of the number of total (claimed) positive genes, TP. A, FDR in the control experiment where fluorescent probe was confined to the IS5 sequence. B, FDR in the experiment where floarescent probe extended into the IS5 neighboring regions.
To determine whether there is any relationship between the proximity of a neighbour to the IS5 element and its intensity we constructed "Rank vs. Distance" plots for all 11 known IS5 positions in the genome (Fig. 2). The open circles represent the neighbouring ORFs and the solid circles correspond to IS5 elements. Figure 2A demonstrates that most of the genes in immediate proximity to IS5 elements have consistently low ranks, with very few outliers with high ranks. We obtained similar results for both stocks 'A' and 'B' (data not shown). The "Rank vs. Distance" plots in Figure 2B clearly demonstrate that there is no association between the rank and the distance in the control. While the ranks of IS5 elements are very low, the ranks of expected neighbours are high and random.
Figure 2 The ranks of the true neighbours of IS5 elements in the E. coli genome. A, Experiment. B, Control. Rank of fluorescent intensities of target sequences as a function of their positions, relative to the position of an IS5 element. All known 11 locations of IS5 elements are shown as separate panels. The closed circles represent the IS5 elements and open circles represent the neighbours. The lower the rank, the higher probability that the gene is a neighbour of the IS5 element. Experiment and control are described in the legend to Figure 1 and in the text.
100 genes predicted as neighbours of IS5 elements by median rank for stock 'A' are listed in Table 2. They include 12 genes that are likely to be false positives as there are no neighbours of those genes in the list of top 100; 4 pairs of possible significant genes with one neighbour present on the list, 12 groups of likely true positive genes with at least 2 neighbours on the top ranking list. Among 12 groups of significant genes (total of 80 genes), 11 groups are expected neighbours of IS5 elements based on the genome sequence data. Only one group of genes including motB, motA, flhC, flhD, yecG, otsA, otsB are not known neighbours of an IS5. It is highly unlikely to find 7 neighbouring genes as significant genes by chance alone. Table 3 lists 100 genes predicted as neighbours of IS5 elements by median rank method for stock 'B'. Among these 100 genes, 33 were stand alone genes without a single neighbour on the gene list. We designated them as false significant genes. We identified 12 groups of genes (total of 55 genes) as significant genes, with at least 2 neighbours appearing on the list. Among these 12 groups, 11 were known neighbours of IS5 elements. Only one group of three genes (pspB, pspC, pspD) was not previously known to be in the vicinity of an IS5. Compared to stock 'A', the number of false significant genes identified as neighbours of IS elements in stock 'B' is much higher, which is consistent with the observed differences in false discovery rates (Fig. 1). However, we could identify all known 11 locations of IS5 in both stocks. The most pronounced difference between two stocks, in terms of a probability of finding the IS element in a new location by chance, was the apparent presence of the transposon sequence proximal to the flhDC/yecG group of genes.
Table 2 The 100 genes predicted as neighbours of IS5 elements in the genome of MG1655 isolate 'A'. False significant genes are the genes without any neighbours (5 up- or down-stream neighbours) on this gene list; possible significant genes are the genes with one neighbor on the list; significant genes are the genes with at least 2 neighbours on the gene list. The names in bold are the known IS5 elements.
False significant genes lpxC, yi22_1, b0878, ompF, b1297, flip, ogrK, fruA, b2442, b2639, cpxR, araH
Possible significant genes, strings of neighbours 1. yi22_2, tra8_2
2. b1578, rspB
3. yi22_4, yi21_4
4. yi21_5, yi22_5
Significant genes, strings of neighbours 1. b0255, tra8_1, ykfC, yi52_1, ykfD, yagD, insB_2
2. b0546, b0547, b0548, b0550, b0551, yi52_2, b0554, b0555, b0556
3. gltL, gltK, gltJ, ybeJ, yi52_3, b0658
4. b1328, b1329, b1330, yi52_4, b1332, fnr
5. b1361, b1362, trkG, b1368, yi52_5, b1371, b1372, b1374
6. motB, motA, flhC, flhD, yecG, otsA, otsB
7. nac, cobT, cobs, cobU, yi52_6, b1995, yi22_3
8. b2028, gnd, yi52_7, yefJ, yefI, yefH, yefG
9. yejM, yejO, b2191, yi52_8, narP, ccmH
10. glcD, glcC, b2981, yi52_9, b2983, b2984, b2986
11. yhcD, yhcE, yi52_10, b3219, yhcG
12. arsB, yhiS, yi52_11, slp, yhiF, yhiD
Table 3 The 100 genes predicted as neighbours of IS5 elements in the genome of MG1655 isolate 'B'. Refer to the legend in Table 2 for details.
False significant genes araA, b0607, lipA, ybfA, tolR, b0817, ycfJ, b1388, relB, insA_5, flip, b1971, yeeD, b2059, ogrK, ompC, ptsH, b2505, mltB, rpoS, relA, fucK, exuR, yhbY, crp, rhaA, yjdG, tra8_3, yjhG, yjhL, pssR, smf_1, b1965
Possible significant genes strings of neighbours 1. insA_1, yaaC
2. yi22_1, b0365
3. b1141, b1146
4. rspB, rspA
5. yi22_4, yi21_4
6. b3254, yhdT
Significant genes, strings of neighbours 1. ykfC, yi52_1, ykfD, insB_2, insA_2
2. b0546, b0547, b0550, b0551, yi52_2, b0555
3. gltK, gltJ, ybeJ, yi52_3
4. pspB, pspC, pspD
5. b1330, yi52_4, b1332, ydaA, fnr
6. b1368, yi52_5, b1371, b1372
7. cobs, cobU, yi52_6, yi22_3
8. gnd, yi52_7, yefJ, yefI
9. yejO, b2191,yi52_8, ccmH
10. glcC, b2981, yi52_9, b2983, pitB
11. yhcD, yhcE, yi52_10, b3219, b3221, sspB, b3238
12. yhiS, yi52_11, yhiF, tnaA
We confirmed the presence of an IS5 element in the flhDC/yecG intergenic region by PCR. A 2720 bp PCR product was obtained with primers specific for the flhDC/yecG intergenic region and chromosomal DNA from stock 'A' as a template. In contrast, when the same primers were used with DNA from stock 'B', only a 1525 bp DNA fragment was obtained (Fig. 3A and 3B). Further DNA sequence analysis confirmed the presence of IS5 upstream of flhDC, in 'A', while no insertion elements were identified upstream of flhDC in 'B'. The inverted repeats of the IS5 insertion element were located 319 bp upstream of the reported flhDC transcription start site [21] (Fig. 3A). The sequence of the inverted repeats was 100% conserved relative to the reported IS5 gene sequence [22]. In addition, a 4-bp target duplication site (5'-TTAG-3') was found flanking the inverted repeats at the IS5 insertion sites. These experiments showed that as predicted by our analysis, an IS5 element was present in the flhDC /yecG region.
Figure 3 A, A diagram of the flhDC-yecG region in 'A' (motile), isolate 'B' (less motile) and in the IS5:kanr mutant. The location of IS5 relative to flhDC transcriptional start site is shown. The position of the primers used for the verification PCR and the length of the expected PCR products are indicated. The wide arrow at the bottom of the diagram indicates the direction of transcription. B, Confirmation of the presence of an IS5 element in the flhDC-yecG region. 1.0% agarose gel shows the products of the PCR reactions confirming the IS5 mapping predictions as well as the replacement and deletion of the IS5 element by the lambda red recombinase system. The PCR products shown were obtained from reactions done with flhDCF3 and yecGR primers. Lane1: isolate 'A' (motile); lane2: isolate 'B' (less motile); lane3: IS5: kanr; lane4: isolate 'AΔIS5'-colony #1; lane5: isolate 'AΔIS5' – colony #2, lane6: no template negative control; lane7: Hi-lo DNA marker ladder. The size of the PCR reaction products is shown in kb.
We also examined one of the transposon locations that was likely to be a false significant as it has been suggested by our statistical analysis. PCR analysis of the fliP neighbourhood was done with primers specific for the gene regions of fliO/ fliP and fliP/fliQ. The size of the PCR products obtained (1000 bp) was the expected size for the gene regions examined based on genome sequence data [20] (data not shown). This indicates that there is no IS5 element adjacent to fliP and that our assertion that genes present on the list of top ranking candidates without neighbours are likely to be false calls is correct.
Thus the predicted neighbours of an IS5 element were consistent with the verification experiments results and the sequence data. If we assume 11 IS5 insertion sites based on sequence data and the flhDC/yecG site based on verification experiments are the only insertion sites in the chromosome of 'A', then there are 11(neighbours)*12 (sites) = 132 neighbours. Assuming all other genes are non-neighbours, and we designated 80 significant genes as predicted neighbours, then the sensitivity of our algorithm is 80/132 = 60.6%. Specificity is 1 and false discovery rate is 0, since all predicted genes are true neighbours.
Biological implications of the presence of an IS5 element upstream of flhDC
So far we have demonstrated that our method allowed determining genome-wide distribution of insertion elements and that our analysis is sensitive enough for the purposes of differential localization of transposons in genomes of E. coli isolates. Differential localization of insertion elements maybe reflective of different history of isolates or of heterogeneity in a cell population. We confirmed that isolates 'A' and 'B' are indistinguishable by genetic fingerpriniting [23] and by comparative genomic hybridization on arrays [15] (data not shown) and that original population of an ATCC strain was split almost one-to-one between an 'A' type cells containing an IS5 element in the vicinity of flhDC/yecG and a 'B' type without an element. Next, we examined whether two types of cells that we distinguished by a location of an IS5 marker may have phenotypic differences associated with an insertion element. We directly compared expression profiles of two isolates using whole-genome DNA microarrays. Following the lowess smoothing and ANOVA normalization [24], we identified 369 genes at a false discovery rate of less than one gene per list [25], whose transcripts appeared to be differentially abundant between the two isolates <the gene expression data presented in this paper is publicly available in the NCBI GEO database under the following accession numbers: GSE2694>. One functional group dominated the list of significant genes: 54 genes classified as being involved in motility and chemotaxis "consumed" 90% of the variance of the entire list of significant genes. On average the flagella genes in isolate 'A' were 9-fold more abundant than in isolate 'B' (Fig. 4). The levels of expression of flhD and flhC, genes encoding a master-regulator of the motility and chemotaxis regulon and situated in immediate proximity to an IS5 insertion element, were moderately increased by 1.7 and 3.0 fold, respectively (Fig. 4). These findings demonstrated that in addition to the presence of an insertion element, two isolates could be differentiated on the basis of expression profiles.
Figure 4 Relative abundance of flagella transcripts in isolate 'A' versus 'B' (open bars), in 'A' versus 'AΔIS5' (solid bars), and in 'AΔIS5' versus isolate 'B' (gray bars). The log2 (Ratio) values of transcript abundances obtained following lowess smoothing and ANOVA normalization in direct comparisons on whole-genome DNA microarrays are plotted along the Y-axis. The names of the genes are shown under the horizontal axis. Error bars indicate the standard error for at least six experimental replicas.
Previous reports indicated that IS5 could serve as a mobile transcriptional enhancer in E. coli [26]. To determine whether the presence of IS5 was responsible for the differences in transcription observed between isolates 'A' and 'B', an IS5 gene deletion mutant was constructed using the lambda red recombinase gene replacement system [27]. We set up the following direct pairwise multiple replicated comparisons of transcriptional profiles: 'A' vs. 'B ', 'A' vs. 'AΔIS5', 'AΔIS5' vs. 'B'. In all these comparisons, genes involved in motility and chemotaxis represented the bulk of significant transcriptional differences (more than 90% of all variance in each of the significant lists). Deleting a transposon element from an isolate 'A' reduced the activity of flagella regulon on average by a factor of two. In contrast, the flagella transcripts were still more abundant in ΔIS5 strain than in a 'B' isolate, approximately 4 fold on average, suggesting that the presence of an IS5 element does not fully explain the difference in transcriptional activity of the motility regulon between two isolates.
A variety of molecular mechanisms are known to regulate the expression of motility and chemotaxis genes in E. coli [28]. To investigate other possible molecular causes that might be responsible for the differences in transcriptional activity between isolates 'A' and 'B', we compared sequences of known transcriptional regulators and some of their expected target promoters in two isolates. We have determined that the sequence of flhD, flhC, fliA, flgM and lhrA are identical between two isolates (data not shown). Detailed PCR analysis of every single motility and chemotaxis operon and intergenic sequences also revealed no discernable differences between two isolates (data not shown).
Discussion
Herein we presented a reliable and efficient method for linear mapping of mobile elements using whole-genome DNA microarrays. In summary, following DNA microarray hybridization the algorithm to find DNA sequences physically juxtaposed with the sequence of interest is:
1. Calculate the median rank for each gene and sort them in ascending order;
2. Estimate false discovery rates and use them to select a cut-off for significant genes;
3. Filter significant genes by using information about their neighbours
We used sequence data and verification experiments to evaluate the performance of our method. The results were very encouraging. Without any prior knowledge, we could identify positions of all known IS5 elements, and could determine a previously unknown insertion site with high sensitivity and specificity.
Another unique property of this approach is that we used single channel cDNA microarray hybridization without a reference channel. Although having reference channel could potentially control the variation of measurement for each spot, which reference sample should be used is sometimes very problematic [29,30]. Using single channel array hybridization could circumvent the need for an often artificial and inadequate reference as well as it can substantially reduce the cost of microarray experiments. Based on our results, single channel microarray experiments can be used to reliably predict the location of insertion sequences on the chromosome, but we need to pay an additional attention to the choice of normalization methods and test statistics. Traditional global normalization and intensity based test statistics may not be applicable here. Based on the presented results, using rank based statistics (such as median of ranks or rank product) can yield the best results. In the future, we will explore more the applicability of single channel microarray experiments as well as the related normalization and analytical issues in different biological contexts.
Compared to the discriminant analysis by [9], a big advantage of the proposed algorithm is that we do not have to have any a priori knowledge about the location of insertion elements. Instead, our algorithm relies only on the microarray data to predict the locations of a sequence of interest. This property allowed us to identify a differentially localized IS5 element in one bacterial isolate relative to another, where most of IS5 elements had similar (if not identical) locations. The assumption behind the discriminant analysis [9] is that there exists a relatively fixed relationship between the distance from an insertion sequence and the fluorescent intensities of neighbouring probes. This assumption is stronger than ours. Additional advantage of our algorithm is the ease of implementation. Unlike our analysis, the discriminant analysis gives a score for each gene representing the probability that a gene corresponds to a true insertion site. However, we can use estimated FDR to decide the cut-off and use the filtering procedure to screen out possible false positive genes that greatly improves predictive power of the analysis.
DNA microarrays have been successfully used to study relative abundances of RNA and DNA, and in this paper we presented a microarray application for mapping insertion elements in microbial genomes. Here, we discuss how knowledge about the distribution of insertion elements in a bacterial genome leads to important biological insights. Mapping and comparing the distribution of IS5 elements in a laboratory strain E. coli MG1655 led to the identification of two variants of this strain; these isolates were later found to be phenotypically distinct. Moreover, data obtained from IS mapping experiments allowed us to investigate and quantify the effect of IS5 on the transcriptional activity of motility and chemotaxis genes in the genome of E. coli MG1655.
Initial IS5 mapping data predicted two main variants of E. coli strain MG1655. These isolates were obtained by analyzing randomly selected colonies from frozen stocks of the strain MG1655 from the ATCC. The only significant difference between these two strains was the presence of an IS5 element upstream of the main flagella regulator, flhDC. Further biological analyses confirmed the presence of this insertion and showed phenotypic differences between these isolates; one isolate 'A' with an IS5 element upstream of the flhDC was more motile than isolate 'B', which did not contain such an insertion. Interestingly, when a bacterial population from an original ATCC stock was analyzed for the presence of an IS5 in that location we found that about half of the colonies contained an IS5 element in vicinity of the flhDC. In 100% of the examined cases the presence of the element correlated with higher motility of the cells. Following up on this observation, we investigated the frequency of a spontaneous loss of a motile phenotype. Five thousand colonies descended from a motile variant have been screened and no less motile revertants were found (data not shown). This observation suggests that the heterogeneity observed in the E. coli population did not result from an unusually frequent transposition event. Regardless of their origin, such heterogeneities when not accounted for may influence biological interpretation of experimental results where homogeneous offsprings are selected. For instance, before the discovery of the reported heterogeneity in a population of E. coli cells, we have observed on multiple occasions that knocking-out random genes results in differential expression of motility/flagella regulon. However, whenever we attempted to complement the knockouts we would discover that differential expression of motility genes was not affected by complementation. In fact, given the 50:50 heterogeneity of a population, there is a 25% chance of encountering motility related differences between a pure parent and a pure offspring and a 50% chance of encountering difference between a homogeneous offspring and an "impure" parent. We also observed instability of expression differences of the motility/chemotaxis regulon when compared biological replicates of liquid cultures grown from separate individual colonies. Given some published reports of motility differences as a major result of gene disruption [31], we would urge caution in interpreting such observations especially in the absence of convincing complementation data.
While this work was in progress, Barker and co-workers reported the finding of motility variants in E. coli strain MG1655 [32]. E. coli strains with increased motility contained IS5 or IS1 insertions in the promoter region of flhDC. The IS5 element present in these strains was located 96 to 99 nucleotides in the upstream direction relative to reported flhDC transcription start site. Moreover, the study reported increased expression of the flagella regulators flhD and fliA as well as of some other genes that are known to be under the control of flhDC. These data suggested that IS5 was responsible for increasing the level of expression of flagellar genes and producing cells with enhanced motility. Our IS mapping and motility assays results were in agreement with these findings. Furthermore, the IS5 identified in the strains with increased motility reported in our study were in a somewhat similar position (-311 to -314 bp relative to the flhDC transcription start site) and in the same orientation as the IS5 found in the strains reported by Barker [32]. Therefore, we decided to investigate whether the presence of IS5 upstream of flhDC influenced the expression of motility and chemotaxis genes in E. coli. Gene expression profiles were obtained for a motile isolate 'A' and its isogenic ΔIS5 derivative. The level of expression of flagellar genes in these strains was compared to the less motile isolate 'B'. Our analysis showed that deletion of the IS5 element situated upstream of the flhDC decreased the level of expression of flagellar genes but did not account for the differences in motility and gene expression observed between the motile and less motile isolates 'A' and 'B'. Moreover, we observed that cells in which an IS5 element had been deleted showed increased motility and up-regulation of flagellar genes when compared with less motile cells (isolate 'B'). These results are in agreement with the phenotypes observed in motility assays, strain 'AΔIS5' had a swarming rate that was not very different from the swarming rate of the wild type strain A (Table 4). These findings suggest that IS5 can be one of the components of a complex mechanism involved in up-regulation of flagella genes and the increased motility phenotype. If an IS5 were solely responsible for the heterogeneous motility observed in E. coli strains, then deletion of this transposon would have produced phenotypes similar, within error, to the ones observed in the less motile isolate 'B'. Regulation of the flhDC is under control of a variety of mechanisms involving CRP, H-NS, heat shock proteins and transcriptional regulators like LhrA. It is possible that point mutations in any of these regulators or in flagellar regulators not yet identified could explain the observed phenotypic differences. The possibility that genome rearrangements such as deletions or amplifications are responsible for the differential motility phenotype has been ruled out based on comparative genome hybridization and fingerprinting.
Table 4 Swarming rates for isolates 'A', 'B', 'A ΔflhDC:kanr' and 'A ΔIS5'. Swarm diameters were measured every hour for 8 hours for overnight colonies inoculated into soft tryptone agar and incubated at 30°C. Average swarm rates and standard errors are shown for at least 15 replicas.
Strain Swarm rate (mm h-1)
E. coli MG1655 motile (strain A) 4.1 ± 0.1
E. coli MG1655 less motile (strain B) 1.3 ± 0.2
E. coli MG1655 motile (strain A)(ΔflhDC::kanr) 0.0 ± 0.0
E. coli MG1655 motile (strain A) (ΔIS5) 3.2 ± 0.1
The results presented in this report provide evidence that supports the role of IS5 as a transcriptional enhancer of the flhDC-controlled operons. Previously, IS5 was linked to the activation and transcriptional regulation of the bgl operon in Escherichia coli [26]. In this case, transcriptional activation was dependent on the presence of IS5 and independent of its position and orientation relative to the promoter region. Our data demonstrated that an IS5 element 300 bp upstream of the transcription start site produces a mild enhancement of transcription of flagellar genes that translates into a slight increase in motility.
Interestingly, transcriptional differences between isolates 'A' and 'B' are largely confined to the motility/chemotaxis regulon. If activity of global regulators that are known to modulate transcription of this regulon were perturbed, we would have expected a much wider scope of transcriptional differences between two isolates. If we assume that these narrow regulatory differences are centered on the activity of FlhDC, then we could speculate that a modest 2-fold increase in the flhDC transcription results in more than 10-fold up-regulation of the fliA. This level of transcriptional amplification translates into phenotypic differences of three to four folds in the swarming rates between the isolates.
The IS mapping technique described in this study is a powerful tool that allows gaining insight into biological processes that might otherwise be overlooked when solely relied on available sequence data. This technique could be applied for de novo localization of mobile elements or of any other sequence determinant which position needs to be linearly mapped without any prior knowledge. In addition this type of genome-wide mapping combined with DNA cross-linking might provide an efficient way to study three-dimensional chromosomal organization [33].
Conclusion
Microarray analysis combined with rank statistics is an efficient and reliable method to localized mobile elements in a bacterial genome. Information obtained using this method allowed identification of two variants of Escherichia coli strain MG1655 in which the presence of an IS5 element produced an enhancement in transcription of flagellar genes and a slight increase in motility. The technique described here can be applied to the linear mapping of any target sequence without prior knowledge of sequence information.
Methods
Bacterial strains and growth conditions
Escherichia coli strain MG1655 obtained from ATCC was grown on Luria Bertani medium at 37°C in a shaker incubator at 250 rpm. Gene replacement mutants were grown in LB medium containing 30 μg/ml of kanamycin. Strains used for the gene replacement experiments: BW25113/pKD20, BW25141/pKD4 and BT340 were obtained from the E. coli Genetic Stock Center (CGSC) at Yale University. These strains were grown and maintained in LB medium with the appropriate antibiotics.
PCR and fluorescent labeling of IS5 probes
The probes to map the location of insertion elements in the E. coli genome consisted of a 300 bp fragment internal to the IS5 gene that was obtained by PCR with the following primers: IS5F-5'-TCGCCAGTTGGTTATCGTTT-3' and IS5R-5'-AGCTGGGTAATCTGCTGCAT-3'. This probe was fluorescently labeled directly by PCR in a reaction containing 500–1000 ng of genomic DNA, 0.5 μM concentration of each primer, 63 μls of sterile distilled water, 10 μls of 10X Thermopol DNA polymerase buffer (New England Biolabs), 5 units of Thermopol DNA Polymerase (New England Biolabs) and dNTPs at the following concentrations 200 μM dATP, 200 μM dGTP, 200 μM dCTP and 100 μM dTTP. Fluorescent labeling of the PCR products was accomplished by adding 100 μM of Cy5 or Cy3 dUTP (Amersham Pharmacia) to each reaction. DNA amplification was done using the following parameters: One cycle of 98°C for 5 minutes; 35 cycles of 98°C for 2 minutes, 55°C for 1 minute and 30 seconds, 72°C for 2 minutes; one last cycle of 72°C for 5 minutes. Reaction products were evaluated by agarose gel electrophoresis and purified using a Millipore Microcon 30 DNA concentration and purification system.
Microarray experiments
Whole-genome DNA microarrays of E. coli were designed, printed and probed as described previously [34]. The microarray consisted of PCR products, within the open reading frames of Escherichia coli strain MG1655, robotically spotted on poly-L-lysine-coated glass slides. To ensure the success of PCR amplification and to minimize cross-hybridization we redesigned more than 700 primer pairs from the original set of primers supplied by Sigma-Genosys and their sequences could be downloaded from
Relative transcript abundances in isolates 'A', 'B' and derivatives have been measured by direct pairwise comparisons in competitive two-color hybridizations. Total RNA samples were extracted from cultures grown on LB medium to OD600 of 0.5 using the hot-phenol method [34]. The experimental error of the measurements of RNA abundances was assessed from three independent replicates, where one replicate corresponded to RNA extracted from a culture grown from a separate colony. Following lowess smoothing and variance reduction, differentially expressed genes were identified using two-class comparisons of the adjusted relative expression values by SAM [25] at 1% false discovery rate at the 90th percentile.
Single and two-color hybridizations were carried out in 20–25 ul under a 20 × 20 flat coverslip in hybridization chambers (Monterey Industries) submerged in a 65°C water bath for 5 to 16 hours. Slides were washed sequentially in each of the following solutions: 0.1XSSC and 0.03% SDS, 0.5XSSC, and 0.25XSSC. After the washes, the slides were air dried by centrifugation and scanned with an Axon Genepix 4000B laser scanner at the resolution of 10 um per pixel.
Array CGH (Comparative Genomic Hybridization) was done as a competitive hybridization of fluorescently labeled genomic DNA from isolates 'A' and 'B'. Genomic DNA was extracted using standard procedures [35]. DNA fragments of 700–1000 bps obtained by sonication were used for direct labeling using DNA polymerase I (Klenow fragment). The labeling reaction consisted of 5 μg of DNA, 5 μg of random hexamer (pdN6), 1.5 μl of dNTP mix (0.5 mM dATP, 0.5 mM dCTP, 0.5 mM dGTP and 0.2 mM dTTP), 0.2 mM Cy3 or Cy5 dUTP, 1X Klenow buffer and 8 units of Klenow. Reactions were incubated at 37°C for 2 hours.
PCR to confirm IS element mapping predictions
PCR reactions to confirm the IS mapping predictions were done with primers specific for upstream region of flhDC and the gene regions of fliO/ fliP and fliP/fliQ. The following primers were used for the PCR reactions: flhDCF3-5'-AGCATGAACGTTTTGTTCCC-3'and yecGR1-5'-CACGCTGCGTAAATCTTCAA-3'; fliPF-5'-ACCTGTCCTTCTCTGGCTGA-3' and fliQR-5'-AACAAGGTGCGGACGTAATC-3'; fliOF-5'-CATTATTGCCCTGATCCTCG-3' and fliPR-5'-TTTAAAGGGCAGAGCAATGG-3'. PCR reactions were done using the following conditions: One cycle at 98°C for 5 minutes; 35 cycles at 98°C for 2 minutes, 60°C for 1 minute and 30 seconds, 72°C for 2 minutes; one last cycle at 72°C for 5 minutes. 500 ng of genomic DNA of isolates 'A' and 'B' was used as a template in these reactions.
Motility assays
Motility assays were done using tryptone soft agar [36]. This medium consisted of 1% tryptone peptone, 0.5% NaCl and 0.3% Bacto-agar. To assay for motility, either 5 μls of overnight cultures or a fresh E. coli colony were gently spotted onto soft agar and incubated at 30°C for 5 hours. Motility was evaluated based on the size of the ring formed by bacteria as they grew on tryptone agar.
Construction of gene deletion mutants
Deletion mutants for the flhDC genes (flhDC::kan) were constructed using the lambda red recombinase method described by Datsenko and Wanner [27]. Briefly, E. coli K-12 cells containing the pKD20 plasmid were grown in 50 ml SOB medium containing 100 μg/ml of ampicillin and 1 mM arabinose at 30°C. Once the cells reached OD600 = 0.6, the culture was concentrated 100 fold by centrifugation and washed once with an equal volume of cold sterile distilled water. Subsequent washes were done twice using half the volume of cold sterile distilled water. The final cell pellets were resuspended in 150 μl of ice cold sterile 10% glycerol. The linear DNA fragments used for gene disruptions were obtained by polymerase chain reaction with plasmid pKD4 as a template and the following primers: flhDCH1P1-5'-TTAAACAGCCTGTACTCTCTGTTCATCCAGCAGTTGTGGGGTGTAGGCTGGAGCTGCTTC-3' and flhDCH2P2C-5'-GTGGGAATAATGCATACCTCCGAGTTGCTGAAACACATTTCATATGAATATCCTCCTTAG-3'. Perkin Elmer high fidelity XL DNA polymerase was used for the amplification of the DNA fragments necessary for gene disruptions. The reaction mixtures and thermo-cycling parameters used have been described by the manufacturer (Perkin Elmer, CA). To do gene disruptions, 50 μl of electrocompetent cells and 3 μl (150 ng) of the linear PCR product were placed in a BioRad gene pulser 0.1 cm gap electroporation cuvette. Introduction of linear DNA into cells was accomplished by electroporation at 1.8 Kv for 5 seconds. Shocked cells were recovered by adding 1 ml of SOC medium and by incubating at 37°C for 90 minutes in a lab line "Cell-Grow" tissue culture rotator. After recovery cells were plated on LB plates containing 30 μg/ml of kanamycin. Candidate mutants were verified by PCR analysis using primers specific for the disrupted gene and the kanamycin cassette present in pKD4. IS5 gene deletion mutants were constructed using the procedure described above and the following primers: IS5H1P1-5'-CAGATAAGCTATTTTTAAACAGACACTTACCGCACAACAAGTGTAGGCTGGAGCTGCTTC-3' and IS5H2P2-5'-AACATTAAGTTGATTGTTGCCTTTCTTTGTATTTAATTAGCATATGAATATCCTCCTTAG-3'.
The kanamycin resistance cassette used to construct the flhDC and IS5 gene replacements was removed by expressing FLP recombinase from the helper plasmid (pCP20), after its transformation into the gene replacement mutants [27].
Data Analysis
General considerations
Based on the experiment procedure, we can hypothesize the following: (1) the neighbours of IS5 have higher intensities than non-neighbours, (2) the closer to IS5 a gene is, the higher the intensity of a corresponding element on the array, (3) in the control experiment where no primer extension outside of the element is allowed, only those elements on the array that correspond to the IS5 sequence will have high intensities, its neighbours should not have high intensities. We predicted the neighbours of IS5 elements scattered across the genome based on these hypotheses, i.e. we identified genes with significant increase in intensity. We used GenBank sequence information and the follow-up verification experiments to evaluate the predicted neighbours and thereby assessed the reliability of whole-genome microarray screening experiments in establishing linear linkages of defined sequence markers. In order to identify the neighbours, we first ranked the genes by a summary ranking statistic from higher intensities to lower intensities; then chose a cut-off value and regarded the genes above the cut-off as significant; finally we filtered some possible false positive genes by using information about neighboring genes.
Data pre-processing and ranking statistics
In our analysis we surveyed fluorescent signals from 4281 target sequences representing more than 98% of protein encoding genes of the E. coli genome. We used log2 transformation and without background correction. For normalization, we tried both global normalization (xnorm = x-median) and scale normalization (xnorm = (x-mean)/sd) where x is log2 (intensity) and sd is a standard deviation.
In microarray data analysis, different ranking statistics may give very different lists of significant gene [37], [38]. For traditional two-channel cDNA microarray analysis, SAM t-statistic [25] generally performs well in identifying differential expressed genes. We also evaluated the performance of other commonly used ranking statistics such as sample mean and Student t-statistic (standardized mean statistic). All of these three statistics use intensities directly. Raychaudhuri [19] showed that in their discriminant analysis using rank of intensities predicted the insertion site better than using intensities directly. Therefore we also tried using other two rank-based statistics: rank-product (the geometric mean of ranks of each replicate for each gene) and median rank (the median of ranks of each replicate). We used the sequence data to compare the quality of predictions that were made on the basis of these different statistics.
False discovery rate estimation
We used the false discovery rate (FDR) [39,40] to decide the cut-off values for significant genes. FDR is an alternative to controlling the false positive rate (type I error), and is defined as the expected proportion of false positive genes (FP) among total positive genes (TP); the observed FP/TP ratio is often used to estimate FDR. To calculate FDR, we used a permutation method to estimate the false positive number FP. Under the null hypothesis, we can generate a permuted data set. Specifically, for the rank based statistics, under null hypothesis, the rank of each gene is from a uniform distribution (1,G), where G is the total number of genes in the experiments. So the permutated data set is just a random sample from number 1 to G. For intensities based statistics, the permutation scheme can be found elsewhere [25,38]. We did 100 permutations for each gene. The false positive number from each permutation is the number of genes that counted as significant genes from the permuted data. The average of the false positive numbers over 100 permutations is calculated as FP. The number of genes that counted as significant genes from the original data is regarded as TP, and we estimate FDR = FP/TP [41]. Note that a more elaborate estimator of FDR, namely FDR = π0FP/TP with π0 as the prior probability of null hypothesis being true, has been proposed [42]. Since π0 is a constant and close to 1 in the context of our analysis, we reasoned that using this estimator would not significantly influence our results.
Filtering genes using information from the neighbours
Based on our hypothesis, for a true neighbour gene, not only it should have high rank intensity itself, but its neighbour should also have relatively high intensity, i.e. there should exist an association between the intensities of a true significant gene and its neighbours. If the gene is not truly significant, the intensities will be independent between the gene and its neighbour. By chance alone we can falsely identify the genes that are not true significant genes, but the chance of falsely identifying two or more neighbouring genes at the same time is greatly decreased. To formalize this idea, we used hypergeometric distribution to get the P-values. Suppose, we have G genes, top m genes are identified as significant genes, and we define the upstream and downstream 5 genes as the neighbours of gene i (so altogether there are 11 neighbour genes including the gene i itself). The number of genes from these 11 genes appeared in top m gene list will follow a hypergeometric distribution. If we choose m = 100, then the P-value for identifying one gene among these 11 genes will be 0.23, for identifying two genes will be 0.026 and for three genes will be 0.002. Thus we will deem the genes with at least two neighbours identified together in top m list as the significant genes (the neighbours of IS5), the genes with one neighbour identified in top m list as possible significant genes, and the genes identified alone without any neighbours appeared in top m list as possible false significant genes.
List of Abbreviations
FDR: false discovery rate; TP: total positives; RP: rank product; MR: median rank; IS: insertion elements
Authors' contributions
BMMV designed and performed all the biological experiments presented in this paper and prepared the manuscript. Yang Xie designed and performed statistical analysis and wrote the data analysis section of this article. WP provided advise and contributed to the statistical part of this work. ABK proposed and supervised the project, carried out parts of the statistical analysis, prepared and edited the final version of the manuscript.
Acknowledgements
We would like to thank Jason Beste and Shruthi Ravimohan for their help in the PCR analysis and the screening of motility variants of E. coli. We also thank Jae Ahn for help with RNA extraction experiments. ABK thanks Jon Bernstein for discussions. This work was supported in part by Grant GM066098 from the National Institutes of Health to ABK and by a Ford postdoctoral fellowship to BMMV.
==== Refs
Mahillon J Chandler M Insertion sequences Microbiol Mol Biol Rev 1998 62 725 774 9729608
Nagy Z Chandler M Regulation of transposition in bacteria Res Microbiol 2004 155 387 398 15207871 10.1016/j.resmic.2004.01.008
Bennett PM Genome plasticity: insertion sequence elements, transposons and integrons, and DNA rearrangement. Methods Mol Biol 2004 266 71 113 15148416
Schneider D Duperchy E Coursange E Lenski RE Blot M Long-term experimental evolution in Escherichia coli. IX. Characterization of insertion sequence-mediated mutations and rearrangements. Genetics 2000 156 477 488 11014799
Kalia A Mukhopadhyay AK Dailide G Ito Y Azuma T Wong BC Berg DE Evolutionary dynamics of insertion sequences in Helicobacter pylori. J Bacteriol 2004 186 7508 7520 15516562 10.1128/JB.186.22.7508-7520.2004
Moran NA Plague GR Genomic changes following host restriction in bacteria. Curr Opin Genet Dev 2004 14 627 633 15531157 10.1016/j.gde.2004.09.003
Zhong S Khodursky AB Dykhuizen D Dean AM Evolutionary genomics of ecological specialization Proc Natl Acad Sci USA 2004 101 11719 11724 15289609 10.1073/pnas.0404397101
van Soolingen D de Haas PE Hermans PW Groenen PM van Embden JD Comparison of various repetitive DNA elements as genetic markers for strain differentiation and epidemiology of Mycobacterium tuberculosis. J Clin Microbiol 1993 31 1987 1995 7690367
Kivi M Liu X Raychaudhuri S Altman RB Small PM Determining the genomic locations of repetitive DNA sequences with a whole-genome microarray: IS6110 in Mycobacterium tuberculosis. J Clin Microbiol 2002 40 2192 2198 12037086 10.1128/JCM.40.6.2192-2198.2002
Green L Miller RD Dykhuizen DE Hartl DL Distribution of DNA insertion element IS5 in natural isolates of Escherichia coli. Proc Natl Acad Sci USA 1984 81 4500 4504 6087327
Ochman H Gerber AS Hartl DL Genetic applications of an inverse polymerase chain reaction. Genetics 1988 120 621 623 2852134
Papadopoulos D Schneider D Meier-Eiss J Arber W Lenski RE Blot M Genomic evolution during a 10,000-generation experiment with bacteria Proc Natl Acad Sci USA 1999 96 3807 3812 10097119 10.1073/pnas.96.7.3807
Zhong S Deam AM Rapid identification and mapping of insertion sequences in Escherichia coli genomes using vectorette PCR BMC Microbiol 2004 4 26 15242519 10.1186/1471-2180-4-26
DeRisi JL Iyer VR Brown PO Exploring the metabolic and genetic control of gene expression on a genomic scale. Science 1997 273 680 686 9381177 10.1126/science.278.5338.680
Pollack JR Perou CM Alizadeh AA Eisen MB Pergamenschikov A Williams CF Jeffrey SS Botstein D Brown PO Genome-wide analysis of DNA copy-number changes using cDNA microarrays. Nat Genet 1999 23 41 46 10471496 10.1038/14385
Horak CE Snyder M ChIP-chip: a genomic approach for identifying transcription factor binding sites. Methods Enzymol 2002 350 469 483 12073330
Iyer V Horak C Scafe C Bostein D Snyder M Brown PO Genomic binding sites of the yeast cell-cycle transcription factors SBF and MBF. Nature 2001 409 533 538 11206552 10.1038/35054095
Gitan RS Shi H Chen CM Yan PS Huang TH Methylation-specific oligonucleotide microarray: a new potential for high-throughput methylation analysis. Genome Res 2002 12 158 164 11779841 10.1101/gr.202801
Raychaudhuri S Stuart JM Liu X Small PM Altman RB Pattern recognition of genomic features with microarrays: site typing of Mycobacterium Tuberculosis strains Proc Int Cont Intell Syst Mol Biol 2000 286 295 10977090
Blattner FR Plunkett G 3rdBloch CA Perna NT Burland V Riley M Collado-Vibes J Glasner JD Rode CK Mayhew GF Gregor J Davis NW Kirkpatrick HA Goeden MA Rose DJ Mau B Shao Y The complete genome sequence of Escherichia coli K-12. Science 1997 277 1453 1462 9278503 10.1126/science.277.5331.1453
Soutourina O Kolb A Krin E Laurent-Winter C Rimsky S Danchin A Bertin PN Multiple control of flagellum biosynthesis in Escherichia coli: role of H-NS protein and the cyclic AMP-catabolite activator protein complex in transcription of the flhDC master operon. J Bacteriol 1999 181 7500 7508 10601207
Schoner B Kahn M The nucleotide sequence of IS5 from Escherichia coli. Gene 1981 14 165 174 6269959 10.1016/0378-1119(81)90112-8
Dombek PE Johnson LK Zimmerley ST Sadowsky MJ Use of repetitive DNA sequences and the PCR To differentiate Escherichia coli isolates from human and animal sources. Appl Environ Microbiol 2000 66 2572 2577 10831440 10.1128/AEM.66.6.2572-2577.2000
Wolfinger RD Gibson G Wolfinger ED Bennett L Hamadeh H Bushel P Afshari C Paules RS Assessing gene significance from cDNA microarray expression data via mixed models. J Comput Biol 2001 8 625 637 11747616 10.1089/106652701753307520
Tusher VG Tibshirani R Chu R Significance analysis of microarrays applied to ionizing radiation response Proc Natl Acad Sci U S A 2001 98 5116 5121 11309499 10.1073/pnas.091062498
Schnetz K Rak B IS5: A mobile enhancer of transcription in Escherichia coli Proc Natl Acad Sci USA 1992 89 1244 1248 1311089
Datsenko KA Wanner BL One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc Natl Acad Sci U S A 2000 97 6640 6645 10829079 10.1073/pnas.120163297
Soutourina OA Bertin PN Regulation cascade of flagellar expression in Gram-negative bacteria. FEMS Microbiol Rev 2003 27 505 523 14550943 10.1016/S0168-6445(03)00064-0
Talaat AM Howard ST Hale W Lyons R Garner H Johnston SA Genomic DNA standards for gene expression profiling in Mycobacterium tuberculosis. Nucleic Acids Res 2002 30 e104 12384606 10.1093/nar/gnf103
Williams BA Gwirtz RM Wold BJ Genomic DNA as a cohybridization standard for mammalian microarray measurements. Nucleic Acids Res 2004 32 e81 15190126 10.1093/nar/gnh078
Oshima T Aiba H Masuda Y Kanaya S Sugiura M Wanner BL Mori H Mizuno T Transcriptome analysis of all two-component regulatory system mutants of Escherichia coli K-12 Mol Microbiol 2002 46 281 291 12366850 10.1046/j.1365-2958.2002.03170.x
Barker CS Prug BM Matsumura P Increased motility of Escherichia coli by insertion sequence element integration into the regulatory region of the flhDC operon J Bacteriol 2004 186 7529 7537 15516564 10.1128/JB.186.22.7529-7537.2004
Dekker J Rippe K Dekker M Kleckner N Capturing chromosome conformation. Science 2002 295 1306 1311 11847345 10.1126/science.1067799
Khodursky AB Bernstein JA Peter BJ Rhodious V Wendisch VF Zimmer DP Escherichia coli spotted double-strand DNA microarrays:RNA extraction, labeling, hybridization, quality control and data management Methods Mol Biol 2003 224 61 78 12710666
Sambrook J Russell DW Molecular Cloning: A Laboratory Manual 2001 Third Edition Cold Spring Harbor, NY, CHSL Press
Alder J Chemotaxis in bacteria Science 1966 153 708 716 4957395
Lonnstedt I Speed T Replicated microarray data. Statistica Sinica 2002 12 31 46
Xie Y Jeong KS Pan W Khodursky AB Carlin B A case study on choosing normalization methods and test statistics for two-channel microarray data Comp Funct Genom 2004 5 432 444 10.1002/cfg.416
Efron B Tibshirani R Storey JD Tusher VG Empirical Bayes analysis of a microarray experiment. Journal of the American Statistical Association 2001 96 1151 1160 10.1198/016214501753382129
Benjamini Y Hochberg Y Storey JD Tibshirani R Controlling the false discovery rate: a practical and powerful approach to multiple testing JRSS-B 1995 57 289 300
Pan W On the use of permutation in and the performance of a class of nonparametric methods to detect differential gene expression. Bioinformatics 2003 19 1333 1340 12874044 10.1093/bioinformatics/btg167
Storey JDTR Statistical significance for genome-wide experiments. Proc Natl Acad Sci 2003 100 9440 9445 12883005 10.1073/pnas.1530509100
| 15929794 | PMC1174868 | CC BY | 2021-01-04 16:39:53 | no | BMC Genomics. 2005 Jun 1; 6:81 | utf-8 | BMC Genomics | 2,005 | 10.1186/1471-2164-6-81 | oa_comm |
==== Front
BMC Infect DisBMC Infectious Diseases1471-2334BioMed Central London 1471-2334-5-421592152310.1186/1471-2334-5-42Case ReportEmbolic stroke complicating Staphylococcus aureus endocarditis circumstantially linked to rectal trauma from foreign body: a first case report Pandey Braj B [email protected] Tuan C [email protected] John F [email protected] University of California, San Diego, USA2005 27 5 2005 5 42 42 27 1 2005 27 5 2005 Copyright © 2005 Pandey et al; licensee BioMed Central Ltd.2005Pandey et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Diagnostic and therapeutic instrumentation of the lower gastrointestinal tract has been reported to result in bacteremia and endocarditis. No such case has been reported in persons with a history of rectal foreign body insertion despite its potential for greater trauma.
Case presentation
A 58-year-old male was admitted with confusion and inability to speak. His past history was notable for hospitalization to extract a retained plastic soda bottle from the rectosigmoid two years prior. On examination, he was febrile, tachycardic and hypotensive. There was an apical pansystolic murmur on cardiac examination. He had a mixed receptive and expressive aphasia, and a right hemiparesis. On rectal examination he had perianal erythema and diminished sphincter tone. Magnetic resonance imaging of the brain showed infarction of the occipital and frontal lobes. Transesophageal Echocardiography of the heart revealed vegetations on the mitral valve. All of his blood culture bottles grew methicillin sensitive Staphylococcus aureus. He was successfully treated for bacterial endocarditis with intravenous nafcillin and gentamicin. The rectum is frequently colonized by Staphylococcus aureus and trauma to its mucosa can lead to bacteremia and endocarditis with this organism.
In the absence of corroborative evidence such as presented here, it is difficult to make a correlation between staphylococcal endocarditis and anorectal foreign body insertion due to patients being less than forthcoming
Conclusion
There is a potential risk of staphylococcal bacteremia and endocarditis with rectal foreign body insertion. Further studies are needed to explore this finding. Detailed sexual history and patient counseling should be made a part of routine primary care.
==== Body
Background
There is a large body of surgical literature reporting anal eroticism resulting in rectal trauma and retained foreign bodies [1,2], but there is no report of bacteremia or endocarditis occurring in these patients. Staphylococcus aureus is an aggressive pathogen and bacteremia with this organism can infect healthy heart valves. The rectal mucosa is a major site of colonization by this organism. We describe a patient with a past history of surgical extraction of a retained plastic soda bottle from the rectosigmoid, who later developed staphylococcal endocarditis resulting in septic embolism and stroke. No such case has been reported in the literature.
Case presentation
A 58-year-old male was brought to the emergency room with confusion and an inability to speak for 1 day. He had a past history of hypertension and hypomania. He was single and he lived alone. On physical examination, his blood pressure was 80/63 mm of Hg, heart rate 126/minute, and temperature 102°F. He was awake but unable to speak due to a mixed receptive and expressive aphasia. He had right homonymous hemianopsia and hemiplegia. Cardiac examination was positive for a pansystolic murmur in the apical area. The abdominal examination was unremarkable. On rectal examination there was perianal erythema and diminished sphincter tone. Complete blood count showed WBC 15,400/mm3 with a left shift, hematocrit 49%, and normal platelets. Serum chemistries showed glucose 84 mg/dl, albumin 2.9 mg/dl, and calcium 8.5 mg/dl. Magnetic Resonance Imaging of the brain, including diffusion weighted imaging, revealed acute hemorrhagic infarction of the left occipital lobe and acute embolic infarctions of the left frontal and right occipital lobes (Fig 1 and 2). Transesophageal Echocardiogram demonstrated mitral regurgitation and large vegetations on the posterior leaf of the mitral valve. Splenic and renal infarcts were visible on Computerized Tomography of the abdomen. All of the blood culture bottles and the urine culture grew methicillin sensitive Staphylococcus aureus. Treatment of the bacterial endocarditis was started with intravenous nafcillin and gentamicin. The patient had a significant recovery of speech and motor function within a few days. When asked about recent dental work, he gave a history of a tooth extraction 2–3 days before the hospitalization but was unable to provide information about his dentist. A dentist from the hospital examined the patient and found no clinical evidence of the extraction. The patient underwent a complete neuropsychiatric evaluation. He displayed confabulation and perseverance (marriage, retirement, hospitalization were reported using the same date which was his birthday). He also had significant executive dysfunction including concrete thinking and poor insight regarding his health and cognitive problems. On hospital day 12 the patient had mitral valve replacement surgery using a bioprosthetic valve. He completed a six-week course of intravenous antibiotic treatment. He also underwent extensive rehabilitation therapy and was sent home after 8 weeks of hospitalization. On subsequent follow-up visits the patient showed complete recovery from the stroke and was back to his baseline. In response to questions about his sexual history the patient indicated having heterosexual relations with multiple partners. However, his answers were inconsistent and seemed unreliable.
Figure 1 Contrast enhancing infarction in occipital lobe (arrow)
A closer examination of his medical records revealed that two years prior to this hospitalization, the patient was admitted with a plastic soda bottle retained in the rectosigmoid for 3 days. The bottle had been filled with warm water and inserted into his anal canal for sexual stimulation. It slipped all the way into the rectum and could not be retrieved. His attempts to extract it at home were unsuccessful. In the emergency room his physical examination was normal except for a palpable mass in the suprapubic area, decreased anal sphincter tone and a dilated rectal vault. On X-ray of the abdomen, the outline of a plastic bottle was visible in the rectosigmoid (Fig 3). The patient was taken to the operating room and under spinal anesthesia the bottle was extracted. He had multiple lacerations of the rectal mucosa but there was no perforation. He went home the next day. On follow-up visits to primary care he was noted to be overall healthy except having mild hypertension. His behavior was indicative of hypomania, but he did not get a formal psychiatric evaluation.
Figure 2 Contrast enhancing infarction in the frontal lobe (arrow).
Figure 3 X-ray of abdomen 2 years prior showing outline of plastic soda bottle in sigmoid colon (arrows).
Foreign bodies in the rectum and methods of their extraction have been amply chronicled in the surgical literature [1]. Sexual stimulation is the reason in a majority of these cases [2]. Local trauma, perforation, and resultant peritonitis are well known complications [3]. An unlimited PUBMED search for articles on bacteremia or endocarditis related to rectal foreign body insertion was unfruitful (We tried MeSH terms anal/rectal/colorectal/foreign bodies/anorectal /sexual deviation for the purpose). Bacteremia and septicemia from barium enema [4], and therapeutic anal dilatation [5] have been published. Procedures like fiberoptic sigmoidoscopy are known to cause endocarditis [6], but septic stroke resulting from endocarditis related to lower gastrointestinal instrumentation has also not been reported.
The patient's history of anorectal insertion of a plastic soda bottle for sexual gratification is consistent with published reports of use of large objects for this purpose [7]. The resultant rectal trauma can easily lead to bacteremia. Rectal carriage of Staphylococcus aureus is well documented and is a potential source of infection [8]. This organism tends to be more abundant on the rectal mucosa than within the feces [9]. In a study of gastrointestinal colonization, Staphylococcus aureus grew from the culture of rectal swabs in 60% cases versus 53% positive culture of nasal swabs taken from the same subjects [10]. This organism is known to cause endocarditis of normal heart valves [11,12]. Neurologic complications of infective endocarditis, particularly embolic events, tend to be higher in cases of endocarditis caused by Staphylococcus aureus [13].
It is known that few patients with rectal foreign bodies will freely admit to transanal introduction [14]. This explains to some extent the paucity of literature linking this practice with bacteremia or endocarditis. We believe our patient was habituated to rectal insertion of foreign bodies and that is evident from his previous history along with the clinical findings of perianal erythema and diminished sphincter tone [14]. In the absence of a reliable history from the patient, the link between endocarditis and rectal trauma in this case is based on circumstantial evidence. A further study of patients with well-documented evidence of rectal foreign body insertion could be the next step to explore this important observation.
Conclusion
The rectum is a frequent site of Staphylococcus aureus carriage. Trauma from foreign objects in the rectum carries a risk of staphylococcal bacteremia that is known to result in endocarditis of both normal and abnormal heart valves. Further studies are needed to explore this finding. It is important to get a detailed sexual history from patients visiting primary care clinics. Patient education and warning may help prevent catastrophic complications of this risky practice.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
BBP carried out the clinical study of the patient, conceived the study, researched the literature, and wrote the article. TCD carried out the clinical study of the patient, researched the literature, and edited the article. JFH provided radiological diagnosis, figure legends and computerized figures.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
Claudia Hall, NP for providing additional clinical information.
A written consent was obtained from the patient for publication of study.
==== Refs
Yaman M Deitel M Burul CJ Shahai B Hadar B Foreign bodies in the rectum Can J Surg 1993 36 173 7 8472230
Cohen JS Sackier JM Management of colorectal foreign bodies J R Coll Surg Edinb 1996 41 312 5 8908954
Barone JE Sohn N Nealon TF Jr Perforations and foreign bodies of the rectum: report of 28 cases Ann Surg 1976 184 601 4 984928
Williams SM Harned RK Recognition and prevention of barium enema complications Curr Probl Diagn Radiol 1991 20 123 51 1889235 10.1016/0363-0188(91)90016-U
Goldman G Zilberman M Werbin N Bacteremia in anal dilatation Dis Colon Rectum 1986 29 304 5 3516600
Norfleet RG Infectious endocarditis after fiberoptic sigmoidoscopy. With a literature review J Clin Gastroenterol 1991 13 448 451 1918852
Subbotin VM Davidov MI Fainshtein AV Abdrashitov RR Rylov IuL Sholin NV Foreign bodies of the rectum Vestn Khir Im I I Grek 2000 159 91 95 10890111
Williams REO Healthy carriage of Staphylococcus aureus: its prevalence and importance Bacteriol Rev 1963 27 56 71 14000926
Kent C Solliday J Comparision of rectal swabs and stool cultures for the detection of gastrointestinal carriage of Staphylococcus aureus Journal of Clinical Microbiology 1980 11 433 434 6989861
Rimland D Roberson B Gastrointestinal carriage of methicillin-resistant staphylococcus aureus Journal of Clinical Microbiology 1986 24 137 138 3722359
Petti CA Fowler VG Jr Staphylococcus aureus bacteremia and endocarditis Cardiol Clin 2003 21 219 233 12874895 10.1016/S0733-8651(03)00030-4
Barrau K Boulamery A Imbert G Casalta JP Habib G Messana T Bonnet JL Rubinstein E Raoult D Causative organisms of infective endocarditis according to host status Clin Microbiol Infect 2004 10 302 308 15059118 10.1111/j.1198-743X.2004.00776.x
Tunkel AR Kaye D Neurologic complications of infective endocarditis Neurol Clin 1993 11 419 40 8316194
Ooi BS Ho YH Eu KW Nyam D Leong A Seow-Choen F Management of anorectal foreign bodies: a cause of obscure anal pain Aust N Z J Surg 1998 68 852 855 9885867 10.1046/j.1440-1622.1998.01463.x
| 15921523 | PMC1174869 | CC BY | 2021-01-04 16:28:16 | no | BMC Infect Dis. 2005 May 27; 5:42 | utf-8 | BMC Infect Dis | 2,005 | 10.1186/1471-2334-5-42 | oa_comm |
==== Front
BMC Mol BiolBMC Molecular Biology1471-2199BioMed Central London 1471-2199-6-161598516310.1186/1471-2199-6-16Research ArticleCloning and characterization of cDNAs encoding putative CTCFs in the mosquitoes, Aedes aegypti and Anopheles gambiae Gray Christine E [email protected] Craig J [email protected] Department of Entomology, Texas A&M University, MS 2475, College Station, TX 77843-2475 USA2 Interdisciplinary Graduate Program in Genetics, Texas A&M University, MS 2475, College Station, TX 77843-2475 USA2005 28 6 2005 6 16 16 7 4 2005 28 6 2005 Copyright © 2005 Gray and Coates; licensee BioMed Central Ltd.2005Gray and Coates; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
One of the many ascribed functions of CCCTC-binding factor (CTCF) in vertebrates is insulation of genes via enhancer-blocking. Insulation allows genes to be shielded from "cross-talk" with neighboring regulatory elements. As such, endogenous insulator sequences would be valuable elements to enable stable transgene expression. Recently, CTCF joined Su(Hw), Zw5, BEAF32 and GAGA factor as a protein associated with insulator activity in the fruitfly, Drosophila melanogaster. To date, no known insulators have been described in mosquitoes.
Results
We have identified and characterized putative CTCF homologs in the medically-important mosquitoes, Aedes aegypti and Anopheles gambiae. These genes encode polypeptides with eleven C2H2 zinc fingers that show significant similarity to those of vertebrate CTCFs, despite at least 500 million years of divergence. The mosquito CTCFs are constitutively expressed and are upregulated in early embryos and in the ovaries of blood-fed females. We have uncovered significant bioinformatics evidence that CTCF is widespread, at least among Drosophila species. Finally, we show that the An. gambiae CTCF binds two known insulator sequences.
Conclusion
Mosquito CTCFs are likely orthologous to the widely-characterized vertebrate CTCFs and potentially also serve an insulating function. As such, CTCF may provide a powerful tool for improving transgene expression in these mosquitoes through the identification of endogenous binding sites.
==== Body
Background
CTCF (CCCTC-binding factor) was originally identified as a transcriptional repressor in studies of the chicken lysozyme silencer [1] and the regulation of the chicken c-myc gene [2]. Since that time, CTCF has been extensively characterized in vertebrates as a ubiquitously-expressed, highly-conserved, multivalent transcription factor that utilizes different zinc finger (ZF) combinations to specifically bind diverse nucleotide sequences, resulting in the repression or activation of target genes, creation of hormone-responsive silencers and the formation of enhancer-blocking boundary elements (reviewed in [3]). Multiple, independent studies have established vertebrate CTCF as a central player in the regulation of gene expression via its association with every known vertebrate insulator [3-5]. Further characterization of these proteins revealed their insulator function to be central in three contexts: (a) constitutive insulation of the chicken β-globin gene at the 5'HS4 site [6,7] and the human apolipoprotein B gene at the 5' boundary [8], (b) imprinted insulation via methylation-sensitive binding to the Igfr2-H19 control locus [6,9-14], the DM1 locus [5] and the DLK1/GTL2 locus [15], and (c) as part of a more complex, multipartite insulator regulated by ligand binding [16]. Most recently, CTCF-dependent insulators have been identified in transitional chromatin, with high levels of H3 acetylation and essentially no CpG methylation, between escape and inactivated genes on both mouse and human inactivated X chromosomes [17]. Finally, Tsix and CTCF have been proposed to comprise a regulated epigenetic switch for X-inactivation in mammals [18]. Clearly, CTCF plays a pivotal role at multiple levels of gene regulation and genome organization in vertebrate organisms.
Long thought to be exclusive to vertebrates, a CTCF orthologue was recently characterized in Drosophila melanogaster with domain structure, binding site specificity and transcriptional repressor activity similar to that of vertebrate CTCF [19]. Significantly, these researchers also demonstrated that a known Drosophila insulator, Fab8, mediates enhancer-blocking via CTCF in both Drosophila and vertebrate cell lines. We have cloned and characterized two mosquito CTCF-like cDNAs encoding polypeptides with significant similarity and insulator binding properties to both the vertebrate and Drosophila CTCFs. Analysis of available genome sequence from numerous invertebrate species yields promising candidates for additional CTCF orthologues. Clearly, this versatile protein has much more ancient roots than once thought.
Results
Cloning of Ae. aegypti and An. gambiae CTCF-like cDNAs
A BLAST search using the human CTCF protein sequence [20] as a query uncovered a cDNA from D. melanogaster [AAL78208], subsequently characterized by Moon et al. [19] as an orthologous CTCF factor. This sequence was then used to query the An. gambiae genome assembly at the Ensembl database [21], resulting in a highly significant hit of the predicted novel gene ENSANGG00000015222 (e-139). These two dipteran sequences were aligned with known vertebrate CTCF sequences from Gallus gallus [22], Mus musculus and Homo sapiens [20], Rattus norvegicus [NP_114012.1] and Xenopus laevis [23] using the ClustalW algorithm (Vector NTI™ Suite 8, InforMax, Inc., 1999). This multiple sequence alignment was used for degenerate PCR primer design. Degenerate PCR amplification, using Ae. aegypti larval cDNA as a template, yielded a single PCR product of 504 base pairs, corresponding to a 168 amino acid polypeptide containing six of the eleven predicted zinc-finger domains. PCR amplification was initially performed with an An. gambiae larval cDNA template and primers corresponding to the 5' and 3' ends of the predicted novel coding sequence. This yielded a single product of 2040 base pairs, corresponding to a translated polypeptide of 680 amino acid residues. Subsequent 5' and 3' RACE (rapid amplification of cDNA ends) in both species yielded putative full-length cDNAs of 2616 and 4544 base pairs for Ae. aegypti (AY935523) and An. gambiae (AY939827), respectively. Alignment of the corresponding polypeptide sequences with both the D. melanogaster and H. sapiens CTCFs revealed significant differences in the N-terminal and C-terminal regions of the protein, however there was 38% identity and 56% similarity across all eleven zinc finger domains (Fig. 1). Furthermore, 68% of the critical binding residues were conserved, despite at least 500 million years of divergence between invertebrate and vertebrate species [24].
CTCF appears widespread in Drosophila species
Available genome sequence for multiple drosophilid species was queried at Flybase [25] using the An. gambiae amino acid sequence and the tBLASTx algorithm. All species searched produced single hits of very high significance, ≤ e-126. Each of these was submitted as a BLASTp query of the non-redundant database at NCBI [26] and confirmed to be a significant match to known CTCFs. Sequences with complete zinc finger regions were trimmed to the zinc-finger region plus five flanking amino acid residues and aligned with the corresponding region of CTCFs from H. sapiens, G. gallus, X. laevis, Danio rerio [NP_001001844], Tetraodon nigroviridis [CAF99566], and Fugu rubripes (Ensembl novel gene SINFRUG00000147322). The corresponding region of zinc finger protein 2 from Caenorhabditis elegans [NP_500033], a protein that contains 11 C2H2 zinc finger domains, a coil-coil region and predicted nuclear localization sequence, was also included in the alignment and used as an outgroup in the subsequent phylogenetic analysis. Two consensus distance-based trees, Neighbor-Joining [27] (Fig. 2) and Fitch-Margoliash [28] (data not shown), were generated with 5000 bootstrap replicates using the Phylip software package [29,30]. Additionally, a maximum-likelihood tree generated by 200,000 iterations of Tree-Puzzle [31] (data not shown) and a Bayesian analysis tree generated by 200,000 cycles of BAMBE [32] with 20,000 cycles of burn-in (data not shown), yielded identical branch topologies.
Mosquito CTCF is expressed constitutively in all developmental stages and is upregulated in early embryos and the ovaries of blood-fed females
Reverse-transcriptase (RT)-PCR amplifications of RNA isolated from embryos, ovaries, larvae, pupae and adults shows CTCF expression across all stages of development and in the ovarian tissues of both Ae. aegypti and D. melanogaster (Fig. 3). Early Ae. aegypti embryos and ovarian tissues from both species clearly show increased expression levels.
Polyclonal antisera raised against An. gambiae CTCF recognizes a single protein band in lysates from An. gambiae Sua4 cultured cells
Immunoblotting of total cell lysate from An. gambiae Sua4 cultured cells with rabbit antisera raised against a c-terminal fragment of An. gambiae CTCF results in identification of a single band migrating at ~84 kD (Fig. 4).
Mosquito CTCF binds in-vitro to both the chicken 5'HS4 and the Drosophila Fab8 insulators
As we were unable to express the full-length mosquito CTCF protein in bacteria, whole cell lysates were prepared from the An. gambiae Sua4 [33] cell line and used in an electrophoretic mobility shift assay (EMSA) to assess whether mosquito CTCF could bind known CTCF-associated insulator sequences (Fig. 5). The intensity of the shifted bands increased with application of greater amounts of protein lysate. The detectable complex was competed by cold, unlabeled probe, indicating that binding was indeed specific. In addition, all reactions contained a 1200-fold excess of cold, non-specific C/G-rich sequences, further illustrating specificity. Finally, the complex could be partially shifted by polyclonal anti-sera generated against the C-terminal region of the An. gambiae CTCF protein.
Discussion
Vertebrate CTCFs, from fish to human, are ≥ 98% identical across the entire zinc finger core of the protein. Comparison of the three dipteran CTCFs reveals 54% identity and 68% similarity within this same region. In addition, amino acid residues considered critical for DNA binding [34] are 89% conserved among these three insect species. This apparent discrepancy can be partially addressed by investigating the molecular substitution rate heterogeneity among vertebrates and invertebrates. Recent maximum likelihood analysis of a set of 50 nuclear genes for vertebrates and dipterans, with Arabidopsis as an outgroup, suggests that the rate of vertebrate molecular evolution slowed considerably with respect to that of dipterans, prior to the origin of the crown-group, Osteichthyes [24]. The much shorter generation times of dipterans have undoubtedly facilitated significant differences in their genome sizes (ranging from 179 Mb in D. melanogaster [35] to 813 Mb in Ae. aegypti [36]) and gene organization patterns, attributable primarily to the number and distribution of repetitive sequences [37]. This would perhaps result in predictions of even greater sequence divergence than is observed in the CTCF genes. It seems likely that at least some of the many attributed vertebrate functions of CTCF are ancestral.
Each of the species examined yielded a single, extremely significant match followed by numerous matches of lesser significance, suggesting a single copy locus. Significant divergence in available N-terminal or C-terminal sequence supports the earlier observation that dipteran genomes have evolved very quickly, and thus these regions may not be critical to the conserved ancestral function(s) of this gene. Additionally, these regions may be more directly involved in protein-protein interactions with other proteins having likewise undergone evolutionary adaptation. High bootstrap support and essentially identical trees generated by four independent methods establishes the tree presented in Fig. 2 as representative of the evolution of this gene sequence. Less bootstrap support in the vertebrate clade is more indicative of the homogeneity of the sequence, rather than uncertainty as to where these species should be located in the tree. Clearly, CTCF is present in vertebrates from fish through mammals and is highly conserved. Of interest is its consistent presence in all Drosophila species queried. The relatedness of the protein sequences mirror the accepted taxonomic relationships among these species as presented at FlyBase [25], likely indicative of a conserved critical function. Significant EST evidence from the flour beetle, Tribolium castaneum, the honey bee, Apis mellifera, and the silkworm moth, Bombyx mori, suggests the presence of CTCF-like genes in multiple insect orders.
The RT-PCR data from both mosquito and fly are consistent with one another, repeatable, and in agreement with both in-situ hybridization data [38] posted for the fly at the Berkeley Drosophila Genome Project website [39] and fly microarray data summarized at Yale University's Drosophila Developmental Gene Expression Timecourse website [40]. In-situ hybridization shows high-levels of Drosophila CTCF transcript ubiquitously distributed throughout stage 1–3 embryos. mRNA levels then decrease until approximately stage 9 where they then increase primarily in the developing nervous and sensory tissues. The neural-specific expression pattern also corresponds to findings in X. laevis where in-situ hybridization with staged embryos revealed weak homogeneous staining prior to stage 14, with subsequent upregulation in neural tissues and the sensory organs of the head [23]. Furthermore, over-expression of CTCF in mice during early embryogenesis resulted in decreased expression of the highly conserved homeobox gene Pax6, causing ocular defects [34]. Microarray data analysis clusters fly CTCF (CG8591) with genes exhibiting a single peak in expression during development, those showing significant expression increases in early embryogenesis, genes with expression changes of at least four-fold across development, and those expressed in the female germline [41]. Taken together, these expression data and the corresponding functional data from vertebrates suggest that CTCF may indeed also be multi-functional in insects. Some possible roles include the regulation of homeobox genes like Pax6, the facilitation of chromatin organization during early development and the establishment and/or maintenance of heterochromatic and euchromatic regions.
The EMSA data support a role for CTCF in endogenous mosquito insulator function and confirm recent findings that the insulator function of CTCF is conserved from invertebrate to vertebrate species [19]. Currently, position effect and position-effect variegation complicate efforts to establish stable transgenic lines in Ae. aegypti and other mosquitoes. Particularly problematic is the highly repetitive nature of much of the intergenic sequence, as well as the compact nature of the genome, which places regulatory elements from neighboring genes in close proximity to one another, where they may inappropriately impact the transgene of interest. The ability to flank transgenes with short, conserved endogenous insulator sequences could significantly improve observed expression levels, and possibly increase the frequency of recovery of transgenic individuals.
Conclusion
We have cloned the cDNAs for two putative mosquito CTCF proteins. We have presented bioinformatics evidence that CTCF is likely present in many arthropod species and that the ancestral portion of the protein is clearly the zinc-finger region. Constitutively expressed in all life stages, mosquito CTCFs are highly upregulated in early embryos and in the ovarian tissues of blood-fed female mosquitoes. Finally, mosquito CTCF specifically binds both the chicken 5'HS4 β-globin and the fly Fab8 insulator sequences. Further characterization of these CTCFs and their binding sites will provide a promising avenue for insulating transgenes in these medically-important mosquito species.
Methods
Isolation of RNA and preparation of cDNA by reverse-transcription
Total RNA was isolated from ~30 mg each of Ae. aegypti and An. gambiae larvae using the RNeasy® Mini Kit (Qiagen, Valencia, CA) followed by DNase I-treatment with DNA-free™ (Ambion, Austin, TX) and was used to synthesize first strand cDNA using the SuperScript II™ reverse transcriptase (Invitrogen, Carlsbad, CA) following the manufacturer's instructions. In order to increase the efficiency of the reverse-transcription reaction, 150 ng/μL of T4 Gene 32 Protein [42] was added to the 1st strand buffer.
Isolation of Ae. aegypti CTCF by degenerate PCR amplification
The amino acid sequences of all known and predicted CTCFs [EAA11339.1, AAL78208, AAG40852, NP_031820, NP_114012, P49711 and Q08705] were identified using the BLAST search algorithm at the National Center for Biotechnology Information (NCBI) website and aligned using the ClustalW algorithm in the Vector NTI™ Suite (InforMax, Inc., 1999). Two completely nested and degenerate PCR primer pairs were designed to a highly-conserved 168 amino acid region using CODEHOP [43,44]. A 504 base pair nested PCR product was obtained from Ae. aegypti larval cDNA using G-1F and K-1R primers in the first PCR reaction, followed by a nested reaction with primers G-2F and J-1R. Each reaction was performed with 2 mM MgCl2, 0.2 μM each primer, 10 mM dNTPs, 0.5 μL cDNA or 1st reaction product and 2.5 units of Taq polymerase (Continental Lab Products, San Diego, CA). The following touchdown PCR conditions were used: 96°C for 4'; 2 cycles of 96°C for 20", 72°C for 1'; 11 cycles of 96°C for 20", 71°C -1.0°C/cycle for 15", 72°C for 45"; 25 cycles of 96°C for 15", 59°C for 15", 72°C for 45"; final extension at 72°C for 2'. Degenerate primers were as follows: G-1F 5' cattccgaggacccgccncayaartg 3', G-2F 5' ggccgctgcagaaccacctiaayacncaya 3', J-1R 5' cgcactgctcgcacctgwancayttytc 3', K-1R 5' ccaggtccagcagctgcykytgickraa 3'.
PCR-amplification and cloning of An. gambiae CTCF
The predicted ORF of An. gambiae CTCF was PCR amplified from ~100 ng of cDNA with 0.2 μM of each primer and 2.5 units of Herculase® Hotstart DNA Polymerase (Stratagene, La Jolla, CA) per the manufacturer's instructions using the following conditions: 95°C for 2'; 5 cycles of 95°C for 30", 55°C for 30", 72°C for 2'45"; 25 cycles of 95°C for 30", 65°C for 30", 72°C for 2'45"; final extension at 72°C for 5'. The primer sequences were: AnophelesCTCFforw 5' caaacgccatatggaggacgtggagctgatat 3' and AnophelesCTCFrev 5' attacctcttgcggccgcttccgtggagaggataaact 3'.
Rapid amplification of cDNA ends (RACE) in Ae. aegypti and An. gambiae
Total RNA was prepared from freshly collected and snap-frozen larvae using the RNeasy® Mini Kit (Qiagen) and immediately DNase I-treated with DNA-free™ (Ambion) according to the manufacturers' instructions. The BD SMART™ RACE cDNA Amplification Kit (Clontech, Palo Alto, CA) was then used to prepare first-strand cDNA and to amplify 5' and 3' RACE products according to the manufacturer's instructions. The gene-specific primers (GSPs) used for each species were: AedesGSP1 5' gtctgtcttgcgcccacatgttg 3', AedesGSP2 5' cgaaagcacgtttacaacttctgg 3', AnophelesGSP1 5' ccacaggtcgtcgggcagagtttgca 3', Anopheles GSP2 5' caatcggagtaagattgtccgaagaaaggtct 3'. GSP1 indicates the primer used for 5' RACE reactions while GSP2 indicates the primer used for 3' RACE reactions. Reaction conditions were as follows: 94°C for 5'; 5 cycles of 94°C for 10", 72°C for 3'; 5 cycles of 94°C for 10", 70°C for 10", 72°C for 3'; 25 cycles of 94°C for 10", 68°C for 10", 72°C for 3'; final extension at 72°C for 8'.
Cloning and sequencing of PCR and RACE products
Products were visualized on a 1% agarose gel, gel purified, cloned into pGEM-T (Promega, Madison, WI) and had their DNA sequence determined using an ABI 3100 capillary sequencer with M13 (-20) and M13 Reverse primers followed by primer walking. At least 3 different clones were analyzed for each PCR or RACE product. The resulting sequences have been deposited in the NCBI GenBank database and have the following accession numbers: [AY935523] (Ae. aegypti) and [AY939827] (An. gambiae).
Phylogenetic analysis
Sequences were trimmed to the 11 ZF region plus five flanking amino acid residues and aligned using MultAlin [45] with the Blosum62 model, a gap opening penalty of 35, a gap extension penalty of 0.5 and no end gap penalty. The resulting alignment was analyzed using the Phylip software package [29]: bootstrapped (5000 replicates) with Seqboot, a distance matrix computed using Protdist (5000 datasets), the matrix submitted to Neighbor or Fitch (5000 trees), a consensus tree determined using Consense and the tree drawn using Drawgram. The MultAlin alignment was also submitted to Tree-Puzzle [31] with 200,000 replicates and to BAMBE [32] with 200,000 cycles and 20,000 burn-in.
Reverse-Transcriptase (RT)-PCR analysis/developmental profile
Total RNA was prepared from freshly collected and snap-frozen samples, DNase treated and the reverse-transcription reaction performed as described above. PCR reactions were assembled with 100 ng cDNA template, 10X buffer, 1.5 μL 10 mM dNTPs, 0.2 μM each primer (Table 1) and 1 μL Advantage2 Taq Polymerase (Clontech) in a total volume of 50 μL. Reaction conditions were as follows: 95°C for 5'; 20, 25 or 30 cycles (see Fig. 3) of 95°C for 15", 55°C for 15", 72°C for 30"; final extension at 72°C for 2'. Products were electrophoresed on a 2% agarose gel, stained with ethidium bromide, destained with ddH2O and imaged. The constitutively expressed D. melanogaster Rp49 gene (153 bp product) and Ae. aegypti S17 gene (200 bp product) were used as controls. Primers were as follows: AedesRT-Forw 5' gtgtttcattgcgagctttgcc 3', AedesRT-Rev 5' tgtctcgatcctccggaatg 3', S17RT-Forw 5' cgaagcccctgcgcaacaagat 3', S17RT-Rev 5' cagctgcttcaacatctccttg 3', DrosophilaRT-Forw 5' atggagactcacgatgattcgg 3', DrosophilaRT-Rev 5' ctcgtcgccattaaccagct 3', Rp49RT-Forw 5' gcgcaccaaggacttcatc 3', Rp49RT-Rev 5' gaccgactctgttgtcgatacc 3'.
Generation of polyclonal antisera against An. gambiae CTCF
The coding sequence for a C-terminal region (amino acid residues 444–680) was PCR amplified and cloned into the pET-30 plasmid (Novagen, VWR International, Bristol, CT), expressed in E. coli (BL21-DE3) and His-tag purified on a Ni-NTA column (Novagen). The purified protein was used to immunize two New Zealand white rabbits following standard procedures.
Immunoblotting
Sua4 cells were lysed in ice-cold lysis buffer (50 mM Tris, pH 7.8; 150 mM NaCl; 1% IGEPAL CA360 (Sigma, St. Louis, MO)) with Complete Protease Inhibitor Cocktail (Roche, Indianapolis, IN) and 1 mM PMSF. Total cell lysate protein was quantitated using the BCA Protein Assay (Pierce, Rockford, IL), aliquoted and frozen at -20°C. Total cell lysate was separated on 8% SDS-PAGE gel and electroblotted to a PVDF membrane in 1X Towbin buffer according to standard protocols. Upon completion of the protein transfer, the gel was washed twice for 10 minutes in 1X TBS buffer (10 mM Tris-HCl, pH 7.5; 150 mM NaCl). It was then blocked in blocking buffer (1.5% non-fat dry milk (NFDM), 1.5% fraction V Bovine Serum Albumin (BSA), 1X TBS, 0.05% Tween-20) with 20% 5X casein (Novagen), in a sealed bag overnight at 4°C. The blot was then washed twice for 10 minutes in 1X TBSTT and once for 10 minutes in 1X TBS and was incubated for 1 hour at room temperature on an orbital shaker with CTCF polyclonal antisera diluted 1:250 in blocking buffer without casein. After antibody binding, the blot was washed twice in 1X TBSTT (1X TBS, 0.05% Tween-20, 0.2% Triton X-100) for 10 minutes and once in 1X TBS for 10 minutes. Anti-Rabbit IgG (Fc) AP conjugate (Promega, Madison, WI 53711) was diluted 1:7500 in blocking buffer without casein and incubated with the blot for 1 hour at room temperature on an orbital shaker. The blot was then washed for 10 minutes five times in 1X TBSTT. Finally, it was developed for 1–10 minutes in Sigma-FAST™ (Sigma Aldrich Chemical Company, St. Louis, MO 63178) according to the manufacturer's instructions.
Electromobility Shift Assay (EMSA)
Sua4 cell lysates were prepared and the total protein quantitated as described above. Probes for EMSA were amplified and simultaneously labelled with α-32P (Amersham) by PCR using the following primers: 5'HS4Forw 5' gagctcacggggacagcccccc 3', 5'HS4Rev 5' aagctttttccccgtatccccc 3', Fab8Forw 5' ggcacaatcaagttaatgttgg 3', Fab8Rev 5' gcaagcgaagagttccattc 3'. The chicken 5'HS4 fragment (250 bp) was amplified from pJC13-1 [46] and the Drosophila Fab8 fragment (309 bp) was amplified from Drosophila genomic DNA. The binding reaction protocol was adapted from Filippova et al. [20]. Approximately 10 fmol of labelled probe was incubated for 15 minutes on ice with 0, 1.5, 7.5 or 15 μg of total cell protein in binding buffer (1X PBS with 5 mM MgCl2, 0.1 mM ZnSO4, 1 mM DTT, 0.1% IGEPAL CA360 (Sigma), 10% glycerol) in the presence of a mixture of non-specific, cold, double-stranded competitor DNAs (500 ng polydI· polydC, 500 ng polydG· polydC, 500 ng SpI oligos, 500 ng Egr1 oligos). The SpI and Egr1 ds oligos contain strong, C/G-rich binding sites for the zinc-finger proteins SpI and Egr1 respectively. Sample 5 contained 150-fold excess unlabeled specific competitor. For the supershift, anti-sera against the An. gambiae CTCF was then added and the reactions incubated an additional 15 minutes on ice. Complexes were separated from the free probe on a 5% native PAGE gel in 0.5X TBE. The gel was run for 3.5 hours at 4°C at 10 V/cm.
Authors' contributions
CEG carried out the studies described in this paper and drafted the manuscript. CJC participated in the design and planning of this study and edited the manuscript. Both authors conceived of the study and have read and approved the final manuscript.
Acknowledgements
We gratefully acknowledge Hans-Michael Muller for the Sua4 cell line, Gary Felsenfeld for the plasmid containing 5'HS4, and Andrea Taylor at the LARR facility at TAMU for antibody generation. This work was supported by NIH grant RO1 AI46432 to CJC.
Figures and Tables
Figure 1 The zinc-finger (ZF) domain is highly conserved between humans and the dipteran insects, Ae. aegypti, An. gambiae and D. melanogaster. Each of the eleven ZFs were aligned using the ClustalW algorithm. Identical and highly conserved residues are highlighted in gray. Weakly conserved residues, the zinc-coordinating residues, and the amino acids with identical binding site recognition properties are indicated in gray, red and blue font respectively.
Figure 2 Phylogenetic analysis of CTCF-like candidates in multiple species. Dendrogram of a neighbor-joining consensus tree of 5000 bootstrap replicates for an alignment of the 11 ZF region of known and predicted CTCFs. The tree topology is consistent with the taxonomic classification of all Drosophila species.
Figure 3 Developmental expression profile of CTCF protein in Ae. aegypti and D. melanogaster. The expression of CTCF was analyzed using RNA isolated from multiple individuals at each of the indicated stages: E1 and E24 (embryos ≤ 1 hr and 24 hrs post-oviposition respectively), Lv (larvae), Pf (female pupae), Pm (male pupae), Pu (pupae), Af (adult females), Am (adult males), Ov- and Ov+ (ovaries from non-blood-fed and blood-fed females respectively). -RT, no reverse-transcriptase. A. and B.) Ae. aegypti CTCF, 20 cycles and 30 cycles respectively. C.) Ae. aegypti S17, 25 cycles. D.) D. melanogaster CTCF, 25 cycles. E.) D. melanogaster Rp49, 20 cycles.
Figure 4 An. gambiae CTCF polyclonal antisera recognizes a distinct band migrating ~84 kD in SDS-PAGE. Lysates from An. gambiae Sua4 cultured cells were separated by 8% SDS-PAGE and immunoblotted with CTCF rabbit antisera. The arrow indicates the position of the 81.1 kD marker. The bounding box marks the edges of the gel.
Figure 5 An. gambiae CTCF specifically binds the chicken 5'HS4 and Drosophila Fab8 insulator sequences. Sua4 cells were lysed and increasing amounts of total cell protein (1.5, 7.5, 15 μg represented as solid triangle) were incubated with radiolabeled insulator sequences as follows: A.) Drosophila Fab8 insulator sequence [19]; B.) chicken β-globin FII insulator sequence [6]. The complex was competed (Cp) with ~150-fold excess of cold, unlabeled probe DNA and supershifted (Ab) with polyclonal antibody sera raised against the C-terminal fragment of An. gambiae CTCF. The probe only lane is indicated by P.
==== Refs
Baniahmad A Steiner C Kohne AC Renkawitz R Modular structure of a chicken lysozyme silencer: Involvement of an unusual thyroid hormone receptor binding site Cell 1990 61 505 514 2159385 10.1016/0092-8674(90)90532-J
Lobanenkov VV Nicolas RH Adler VV Paterson H Klenova EM Polotskaja AV Goodwin GH A novel sequence-specific DNA binding protein which interacts with three regularly spaced direct repeats of the CCCTC-motif in the 5'-flanking sequence of the chicken c-myc gene Oncogene 1990 5 1743 1753 2284094
Ohlsson R Renkawitz R Lobanenkov V CTCF is a uniquely versatile transcription regulator linked to epigenetics and disease Trends Genet 2001 17 520 527 11525835 10.1016/S0168-9525(01)02366-6
West AG Gaszner M Felsenfeld G Insulators: many functions, many mechanisms Genes Dev 2002 16 271 288 11825869 10.1101/gad.954702
Filippova GN Thienes CP Penn BH Cho DH Hu YJ Moore JM Klesert TR Lobanenkov VV Tapscott SJ CTCF-binding sites flank CTG/CAG repeats and form a methylation-sensitive insulator at the DM1 locus Nat Genet 2001 28 335 343 11479593 10.1038/ng570
Bell AC West AG Felsenfeld G The protein CTCF is required for the enhancer blocking activity of vertebrate insulators Cell 1999 98 387 396 10458613 10.1016/S0092-8674(00)81967-4
Recillas-Targa F Pikaart MJ Burgess-Beusse B Bell AC Litt MD West AG Gaszner M Felsenfeld G Position-effect protection and enhancer blocking by the chicken b-globin insulator are separable activities Proc Natl Acad Sci U S A 2002 99 6883 6888 12011446 10.1073/pnas.102179399
Antes TJ Namciu SJ Fournier REK Levy-Wilson B The 5' Boundary of the Human Apolipoprotein B Chromatin Domain in Intestinal Cells Biochem 2001 40 6731 6742 11389587 10.1021/bi0100743
Hark AT Schoenherr CJ Katz DJ Ingram RS Levorse JM Tilghman SM CTCF mediates methylation-sensitive enhancer-blocking activity at the H19/Igf2 locus Nature 2000 405 486 489 10839547 10.1038/35013106
Ishihara K Sasaki H An evolutionarily conserved putative insulator element near the 3' boundary of the imprinted Igf2/H19 domain Human Mol Genet 2002 11 1627 1636 12075007 10.1093/hmg/11.14.1627
Fedoriw AM Stein P Svoboda P Schultz RM Bartolomei MS Transgenic RNAi Reveals Essential Function for CTCF in H19 Gene Imprinting Science 2003 303 238 240 14716017 10.1126/science.1090934
Du M Beatty LG Winjing Z Lew J Schoenherr C Weksberg R Sadowski PD Insulator and silencer sequences in the imprinted region of human chromosome 11p15.5 Human Mol Genet 2003 12 1927 1939 12874112 10.1093/hmg/ddg194
Schoenherr CJ Levorse JM Tilghman SM CTCF maintains differential methylation at the Igf2/H19 locus Nat Genet 2003 33 66 69 12461525 10.1038/ng1057
Pant V Mariano P Kanduri C Mattsson A Lobanenkov V Heuchel R Ohlsson R The nucleotides responsible for the direct physical contact between the chromatin insulator protein CTCF and the H19 imprinting control region manifest parent of origin-specific long-distance insulation and methylation-free domains Genes Dev 2003 17 586 590 12629040 10.1101/gad.254903
Wylie AA Murphy SK Orton TC Jirtle RL Novel imprinted DLK1/GTL2 domain on human chromosome 14 contains motifs that mimic those implicated in IGF2/H19 regulation Genome Res 2000 10 1711 1718 11076856 10.1101/gr.161600
Lutz M Burke LJ LeFevre P Myers FA Thorne AW Crane-Robinson C Bonifer C Filippova GN Lobanenkov V Renkawitz R Thyroid hormone-regulated enhancer blocking: cooperation of CTCF and thyroid hormone receptor Embo J 2003 22 1579 1587 12660164 10.1093/emboj/cdg147
Filippova GN Cheng MK Moore JM Truong JP Hu YJ Nguyen DK Tsuchiya KD Disteche CM Boundaries between Chromosomal Domains of X Inactivation and Escape Bind CTCF and Lack CpG Methylation during Early Development Devel Cell 2005 8 31 42 15669143 10.1016/j.devcel.2004.10.018
Chao W Huynh KD Spencer RJ Davidow LS Lee JT CTCF, a candidate trans-acting factor for X-inactivation choice Science 2002 295 345 347 11743158 10.1126/science.1065982
Moon H Filippova GN Loukinov DI Pugacheva E Chen Q Smith ST Munhall A Grewe B Bartkuhn M Arnold R Burke LJ Renkawitz-Pohl R Ohlsson R Zhou J Renkawitz R Lobanenkov V CTCF is conserved from Drosophila to humans and confers enhancer blocking of the Fab-8 insulator EMBO Reports 2005 6 165 170 15678159 10.1038/sj.embor.7400334
Filippova GN Fagerlie S Klenova EM Myers C Dehner Y Goodwin G Neiman PE Collins SJ Lobanenkov VV An exceptionally conserved transcriptional repressor, CTCF, employs different combinations of zinc fingers to bind diverged promoter sequences of avian and mammalian c-myc oncogenes Mol Cell Biol 1996 16 2802 2813 8649389
Ensembl Mosquito Genome Browser
Klenova EM Nicolas RH Paterson HF Carne AF Heath CM Goodwin GH Neiman PE Lobanenkov VV CTCF, a conserved nuclear factor required for chicken c-myc gene, is an 11-Zn-finger protein differentially expressed in multiple forms Mol Cell Biol 1993 13 7612 7624 8246978
Burke LJ Hollemann T Pieler T Renkawitz R Molecular cloning and expression of the chromatin insulator protein CTCF in Xenopus laevis Mech Dev 2002 113 95 98 11900981 10.1016/S0925-4773(02)00005-9
Peterson KJ Lyons JB Nowak KS Takacs CM Wargo MJ McPeek MA Estimating metazoan divergence times with a molecular clock Proc Natl Acad Sci U S A 2004 101 6536 6541 15084738 10.1073/pnas.0401670101
Flybase
National Center for Biotechnology Information (NCBI)
Saitou N Nei M The neighbor-joining method: a new method for reconstructing phylogenetic trees Mol Biol Evol 1987 4 406 425 3447015
Fitch WM Margoliash E Construction of phylogenetic trees Science 1967 155 279 284 5334057
Felsenstein J PHYLIP-Phylogeny Interference Package (Version 3.2) Cladistics 1989 5 164 166
Felsenstein J Inferring phylogenies from protein sequences by parsimony, distance, and likelihood methods Methods Enzymol 1996 266 418 427 8743697
Strimmer K von Haeseler A Quartet puzzling: A quartet maximum likelihood method for reconstructing tree topologies Mol Biol Evol 1996 13 964 969
Larget B Simon D Markov chain Monte Carlo algorithms for the Bayesian analysis of phylogenetics trees Mol Biol Evol 1999 16 750 759
Muller JM Dimopoulos G Blass C Kafatos FC A Hemocyte-like Cell Line Established from the Malaria Vector Anopheles gambiae Expresses Six Prophenoloxidase Genes J Biol Chem 1999 274 11727 11735 10206988 10.1074/jbc.274.17.11727
Suzuki M Gerstein M Yagi N Sterochemical basis of DNA recognition by Zn fingers Nucl Acids Res 1994 22 3397 3405 8078776
Holt RA Subramanian GM Halpern A Sutton GG Charlab R The genome sequence of the malaria mosquito Anopheles gambiae Science 2002 298 129 149 12364791 10.1126/science.1076181
Warren AM Crampton JM The Aedes aegypti genome: complexity and organization Genet Res 1991 58 225 232 1802804
Severson DW Knudson DL Soares MB Loftus BJ Aedes aegypti genomics Insect Biochem Mol Biol 2004 34 715 721 15242713 10.1016/j.ibmb.2004.03.024
Tomancak P Beaton A Weiszmann R Kwan E Shu SQ Lewis SE Richards S Ashburner M Hartenstein V Celniker SE Rubin GM Systematic determination of patterns of gene expression during Drosophila embryogenesis Genome Biol 2002 3 research0088.1 88.14 12537577 10.1186/gb-2002-3-12-research0088
Berkeley Drosophila Genome Project
Drosophila Developmental Gene Expression Timecourse
Arbeitman MN Furlong EEM Imam F Johnson E Null BH Baker BS Krasnow MA Scott MP Davis RW White KP Gene Expression During the Life Cycle of Drosophila melanogaster Science 2002 297 2270 2275 12351791 10.1126/science.1072152
Villalva C Touriol C Seurat P Trempat P Delsol G Brousset P Increased Yield of PCR Products by Addition of T4 Gene 32 Protein to the SMART™ PCR cDNA Synthesis System Biotechniques 2001 31 81 86 11464524
Rose TM Schultz ER Henikoff JG Pietrokovski S McCallum CM Henikoff S Consensus-degenerate hybrid oligonucleotide primers for amplification of distantly related sequences Nucl Acids Res 1998 26 1628 1635 9512532 10.1093/nar/26.7.1628
Rose T Henikoff J Henikoff S CODEHOP (COnsensus-DEgenerate Hybrid Oligonucleotide Primer) PCR primer design Nucl Acids Res 2003 31 3763 3766 12824413 10.1093/nar/gkg524
Corpet F Multiple sequence alignment with hierarchical clustering Nucl Acids Res 1988 16 10881 10890 2849754
Chung JH Whiteley M Felsenfeld G A 5' element of the chicken beta-globin domain serves as an insulator in human erythroid cells and protects against postion effect in Drosophila Cell 1993 74 505 514 8348617 10.1016/0092-8674(93)80052-G
| 15985163 | PMC1174870 | CC BY | 2021-01-04 16:22:25 | no | BMC Mol Biol. 2005 Jun 28; 6:16 | utf-8 | BMC Mol Biol | 2,005 | 10.1186/1471-2199-6-16 | oa_comm |
==== Front
BMC Med GenetBMC Medical Genetics1471-2350BioMed Central London 1471-2350-6-211590450610.1186/1471-2350-6-21Research ArticleSubtelomeric study of 132 patients with mental retardation reveals 9 chromosomal anomalies and contributes to the delineation of submicroscopic deletions of 1pter, 2qter, 4pter, 5qter and 9qter Sogaard Marie [email protected]ümer Zeynep [email protected] Helle [email protected] Johanne [email protected] Birgitte [email protected] Paal [email protected] Vibeke Faurholt [email protected] Peter [email protected] Niels [email protected]öz Sultan [email protected] Morten [email protected] Karen [email protected] The John F. Kennedy Institute, Glostrup, Denmark2 Wilhelm Johannsen Centre for Functional Genome Research, IMBG, The Panum Institute, University of Copenhagen, Denmark3 Department of Paediatrics, Glostrup Hospital, Denmark4 Department of Paediatrics, Roskilde Hospital, Denmark5 Department of Paediatrics, Sonderborg Hospital, Denmark6 Department of Medical Biology and Genetics, Dokuz Eylul University, Faculty of Medicine, Izmir, Turkey7 Department of Clinical Genetics, Rigshospitalet, Denmark2005 17 5 2005 6 21 21 1 11 2004 17 5 2005 Copyright © 2005 Sogaard et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Cryptic chromosome imbalances are increasingly acknowledged as a cause for mental retardation and learning disability. New phenotypes associated with specific rearrangements are also being recognized. Techniques for screening for subtelomeric rearrangements are commercially available, allowing the implementation in a diagnostic service laboratory. We report the diagnostic yield in a series of 132 subjects with mental retardation, and the associated clinical phenotypes.
Methods
We applied commercially available subtelomeric fluorescence in situ hybridization (FISH). All patients referred for subtelomeric screening in a 5-year period were reviewed and abnormal cases were further characterized clinically and if possible molecularly.
Results
We identified nine chromosomal rearrangements (two of which were in sisters) corresponding to a diagnostic yield of approx. 7%. All had dysmorphic features. Five had imbalances leading to recognizable phenotypes.
Conclusion
Subtelomeric screening is a useful adjunct to conventional cytogenetic analyses, and should be considered in mentally retarded subjects with dysmorphic features and unknown cause.
==== Body
Background
Mental retardation (MR) is a common disorder affecting 1–3% of the population, and yet the pathogenesis is only partly understood. Specific etiological factors are found only in about half of the patients, despite thorough clinical and laboratory investigations [1]. It is reasonable to believe that genetic factors are involved in many of the undiagnosed cases, since there is a generally increased recurrence risk for siblings [2], and a large number of different gene mutations are known to be associated with mental retardation. At present, using mental retardation as search word, at least 1000 items will turn out to be associated with mental retardation in the Online Mendelian Inheritance in Man (OMIM). In cases with prenatal onset of symptoms, growth retardation, malformations and dysmorphic features a chromosomal imbalance may play a role. The most common demonstrable genetic causes of global development delay involve chromosomal imbalance, the fragile X syndrome and Rett syndrome [3]. The available data indicate that chromosome abnormalities are found in 4–28% of individuals with mental retardation, and that severity of MR and the presence of congenital anomalies increase the diagnostic yield of chromosome abnormalities [4].
The subtelomeric regions are gene-rich and are often involved in chromosomal rearrangements [5]. Since 1995 it has been recognized that subtle rearrangements at the telomere regions may account for a proportion of cases with unexplained mental retardation [1]. A number of techniques can be applied for subtelomeric screening, for instance FISH with subtelomeric probes, analysis with microsatellite markers or high resolution comparative genome hybridization (HR-CGH) [6]. Recently, the use of MLPA (multiplex ligation dependent probe amplification) [7], and microarrays [8] were described. A recent review of over 2500 tested and reported subjects with mental retardation revealed the presence of subtelomeric rearrangements in approximately 5 % of the cases. [9].
In this study we have investigated 132 mentally retarded patients for subtelomeric rearrangements by FISH analysis using a commercially available system (Cytocell R). We found 9 rearrangements (two were siblings) among 113 cases with both MR and dysmorphic features. Three of the rearrangements were characterized previously by molecular means [10], and in this study we have delineated the size of the deletion in one case with a 2q telomere abnormality.
Methods
Patients
A total of 132 cases from 131 families were referred to the John F. Kennedy Institute for subtelomeric screening. They were all examined in a period between January 1998 and February 2003. A previous cytogenetic analysis at the 500–550 band stage was normal in 130 cases, and in two cases an unresolved subtle chromosome anomaly was suspected. Of these patients 113 cases had mental retardation together with dysmorphic features. Sixteen of these cases had a positive family history of MR and dysmorphic features. In 19 patients mental retardation was present without apparent dysmorphic features. Four of these cases had a positive family history of MR. In all cases where a subtelomeric abnormality was detected the parents were investigated if available.
Subtelomeric screening, FISH analysis, and quantitative PCR
Subtelomeric regions were screened with the Chromoprobe Multiprobe T system (Cytocell). The analysis was performed as instructed by he manufacturer. With this system a simultaneous analysis of 41 chromosome arms, except for the p arms of the acrocentric chromosomes is possible. The signals were detected and analyzed in a fluorescence microscope. All abnormal or ambiguous results were reanalyzed using a specific subtelomeric probe for the suggested abnormality (Cytocell aqua or TelVysion (Vysis)).
To establish the size of the deletion in patient 3, FISH analyses were carried out using 19 BAC clones from the RP11 library mapping to 2q37.2-q37.3 (Human Genome Browser). FISH analysis was carried out using 200 ng BAC DNA using standard protocols. Probe DNA was labeled with biotin-11-dUTP (Boehringer Mannheim) and signals were visualized using avidin-FITC detection system.
To establish if the deletion in case 8 included the NSD1 gene real-time quantitative PCR was carried out: A reference PCR amplifying a 51 bp genomic fragment of the GAPDH gene and a test PCR allowing an amplification of a segment in exon 5 of NSD1, was developed according to the Primer Express/SDS7000 guidelines and evaluated on a SDS7000 real-time PCR utility (Applied Biosystems). DNA from patient 8 and two controls were isolated and diluted (approx 5 ng/ml) to obtain nearly exact cycle threshold values (CT) for the reference PCR. The comparative CT method was used to calculate the relative quantitative relation between the two PCR reactions, and showed a steady 1:0.5 relation in the patient sample (consistent with a deletion of the NSD1 gene)
Case Reports
Patient 1
Patient 1 was a female born to unrelated healthy parents at 40 weeks of gestation. She had a healthy older brother. Birth weight was 2600 g, birth length 49 cm, and head circumference was 31 cm. Apgar scores were 7/1, 9/5, 10/10. The patient showed prenatal- as well as postnatal growth retardation (-3/-3 1/2 SD). She was hypotonic and severely delayed in psychomotor functions. The patient died of pneumonia 3 years old. At that time she was only able to sit with support and had no language except from babbling. She had dysmorphic features including prominent forehead, high-arched palate, flat mid-face with small nose, small palpebral fissures, long philtrum, small mouth with thin lower lip, small hands and feet and vision abnormalities in the form of markedly delayed visual maturation. In addition she had severe skin problems with suppurate eczema in periods. No other cases of mental retardation were known in the family. Clinical features are shown in fig. 1. An autopsy did not reveal any organ malformations.
Her karyotype was 46,XX.ish del(1)(p36.3). Her parents had normal karyotypes without subtelomeric rearrangements. Hence, her deletion at 1p36 was de novo. The size of deletion was characterized at the molecular level in a previous study [10], comprising 8 Mb.
Patient 2
The boy was born as the first child to unrelated healthy parents. The father had two apparently healthy children from a previous relationship. During pregnancy the fetus showed growth retardation. He was born at 39 weeks of gestation with a birth weight of 2125 g. Apgar scores were 8/1, 10/5. He was born with anal atresia, subsequently operated with good results. He had MR and facial dysmorphic features. He showed growth retardation and was found to have a mild delay of motor function at 21 months. He could walk at the age of 2 1/2 years but had no spoken language; instead he used sign language. He had congenital hearing- and vision-impairment with hypermetropia, which were corrected with hearing aid and glasses respectively. Ophthalmological examination revealed optic atrophy. No other cases of mental retardation were known in the family. Cytogenetic analysis revealed additional material on the long arm of chromosome 2. Whole chromosome painting with a chromosome 2 probe did not paint the additional material Subtelomeric screening showed subtelomeric signals on both chromosomes 2 and furthermore, a signal from the subtelomeric probe from chromosome 22q was localized to the additional material on chromosome 2 demonstrating that the extra material derived from 22qter. (Normal subtelomeric signals were also present on 22q). The mother had a normal karyotype without subtelomeric rearrangements, and the father was not available for examination. The karyotype was 46,XY.ish der(2)t(2;22)(q37.2;q1?). Hence, the patient had an unbalanced karyotype with a partial trisomy for the long arm of chromosome 22.
Patient 3
The patient was born as number 2 of 2. She was delivered by Cesarean section at 41 weeks of gestation because of her size and previous section. Birth weight was 5500 g and birth length 58 cm. She had MR with mild delay of motor function and walked alone 16 months old. She had recurrent airway infections and bronchitis. Growth was normal, but her head was large (+3SD). MRI of the brain was normal. She had recurrent fractures and she was diagnosed with osteoporosis using DEXA scanning 6 years old. Her height was 119 cm and weight 25 kg at age 7 years. She had normal levels of calcium, phosphate, magnesium and basic phosphatase. Her thyroid status was normal. Dysmorphic features included frontal bossing, narrow and low dorsum of nose and hypertelorism. No other cases of MR were known in the family.
Her karyotype by subtelomeric screening was 46,XX.ish del(2)(q37.2). Her mother showed the same subtelomeric abnormality which is a known polymorphism of D2S9886. However, further FISH analyses were carried out as described.
Patient 4
The patient was born to healthy, unrelated parents. She was born at 40 weeks of gestation with a birth weight of 2500 g, birth length of 46 cm and head circumference of 28.6 cm. Apgar scores were 10/1, 10/5. One older sibling was healthy. She thrived poorly and was found to be generally hypotonic 2 months old. Her development was retarded and she was admitted to hospital several times with infections and febrile convulsions. A chromosome analysis and metabolic screening at the time was normal. She was treated for epilepsy from the age of 2 years. Feeding problems were pronounced, and she was considered to have an autistic disorder. Her development corresponded to approximately 11–15 months at the age of 2 1/2 years. As she grew up she became fond of eating. MRI of the brain at the age of 9 years was normal. Menarche occurred at the age of 11 years. She had well-developed gross motor and some language skills at the age of 15 years. She was referred to subtelomere chromosome analysis at the age of 12 years because of MR, dwarfism with growth corresponding to -3SD (146 cm and 43 kg), hypotonia and dysmorphic features including microcephalia, micrognathia and protrusion of the eyeballs. See figure 2 for clinical features.
Her karyotype by subtelomeric screening was 46,XX.ish del(4)(p16.1). The parents had normal karyotypes without subtelomeric rearrangements. Hence, she had an unbalanced karyotype with monosomy 4p de novo.
Patient 5
This patient was 27 years old at the time of reinvestigation. Pregnancy, birth and neonatal period were reported as normal. At age 3 years he was referred for chromosome analysis because of mild MR and delay of motor function. Slight dysmorphic features (round facies, small head) were noted. The chromosome analysis (performed both in peripheral blood lymphocytes and cultured skin fibroblasts) revealed additional material on the short arm of chromosome 22 in about 25 % of the metaphases analyzed. At that time it was not possible to reveal the origin of this extra material. At age 27 years he was mildly mentally retarded and worked in a sheltered workshop. His height and weight were normal. Subtelomeric screening showed a signal from the subtelomeric probe12p (in addition to the normal signals on chromosome 12p) localized to the short arm of chromosome 22. His karyotype by subtelomeric screening was 46,XY.ish der(22)t(12;22)(p13;p?). Furthermore, Multicolor FISH confirmed the presence of chromosome 12 material on 22p in approx. 25% of the metaphases. Hence, he had an unbalanced mosaic karyotype with a partial trisomy for chromosome12pter.
Patient 6a, b (and c)
The patients are two sisters (6a,b), who are distantly related to a girl with a similar phenotype (6c) (figure 3). The two sisters were born in 1966 (6a) and 1976 (6b) to healthy, unrelated parents. Patient 6a was born at term with a birth weight of 2900 g and birth length of 48 cm after an uncomplicated pregnancy. Apgar scores were unknown, but amnion fluid was green and the baby was placed in an incubator. Feeding was poor in the neonatal period. One year old her weight was 8700 g and the height was 75 cm. Today (37 years old) her weight is 48 kg and the height is 154 cm. She has scoliosis, small head circumference and spastic gait. She has severe MR and needs help with almost everything. The fine and gross motor skills are poor. (According to the parents she likes being with other people and is generally a happy person).
The younger sister (6b) was also born at term after normal pregnancy and delivery. The birth weight was 3470 g and birth length 48 cm. She had congenital dislocation of the hip and one week old a brace was applied. According to the parents she resembled her older sister, but was much stronger and was able to suckle by herself. One year old her weight was 7300 g and height was 73 cm. She was hyperactive for a number of years with a very little sleep demand. Today (27 years old) she weighs 44 kg and the height is 144 cm. She has severe MR and is dependent on constant assistance. Both sisters have good health in general.
The third girl (6c) within this family was born in 1996 as the first child to healthy parents. Birth weight 1935 g, length 44 cm, head circumference 29.6 cm. She has moderate/severe mental retardation without language. She is growth retarded with microcephalia (head circumference below -3 standard deviations)
Subtelomeric screening revealed an unbalanced translocation in the sisters, and a balanced translocation in their mother. The unbalanced karyotype was 46,XX.ish der(13)t(5;13)(q35.2;q34). Thus the patients had partial monosomy 13qter as well as partial trisomy 5qter. Case 6c had the same unbalanced karyotype.
Patient 7
Patient 7 was a girl born at 42 weeks gestation as the first child to healthy, unrelated parents. The mother previously had two spontaneous abortions. Birth weight was 2770 g, birth length 47 cm and head circumference 30.5 cm. She had a hemangioma on the forehead and a neonatal tooth. She was severely delayed in development with hypotonia and postnatal growth retardation and developed epilepsy. She had dysmorphic features including hypertelorism, narrow eye fissures, broad nasal bridge, large philtrum, abundant head- and body hair, clubfoot and atrial septal defect (ASD). She was severely retarded and died 3 years old from pneumonia.
Subtelomeric screening showed an unbalanced translocation inherited from a balanced translocation in the father. Her karyotype was 46,XX.ish der(9)t(9;22)(q34.2;q13.3)pat.
Patient 8
The patient was a girl born to healthy, unrelated parents. Two older maternal half-sibs were healthy. The pregnancy was complicated by polyhydramnios and she was delivered at 40 weeks of gestation by Cesarean section due to slow progress in labor. Forceps was used through the uterotomia because of macrocephaly. Birth weight was 3955 g, birth length 55 cm, and head circumference 40 cm. The girl was asphyctic with apgar scores 2/1, 7/3, 9/5, and 10/10. Edema of hands and feet were noticed at birth and the neonatal period was complicated by hypoglycaemia. An ultrasound scan of the brain at age 3 days showed signs of haemorrhage in the lateral ventricles, small periventricular cavities, and suggested corpus callosum hypoplasia. The latter was verified at MRI of the brain at 4 weeks of age and in addition enlarged lateral ventricles were present. At 11 months of age renewed MRI in addition showed partial agenesis of gyrus cinguli and periventricular leukomalacies. At 12 months of age significant dysmorphic features, as macrocephaly, dolicocephaly, frontal bossing, receding frontal hairline, deep-set eyes with epicanthus, nystagmus, depressed nasal bridge, and a protruding tongue were noticed. The trunk of her body was long compared to the extremities. Finger- and toenails were thin, brittle and deep set with periungual edema and a tendency to develop paronychion. Skin of the palms and foot soles were thickened, fingertip pads were prominent, and the thumbs were broad and adducted (fig. 4). She was hypotonic. Echocardiography was normal. X-ray examinations showed that metatarsal bones were short and metacarpal bones were short and broad. Furthermore, the bone age at 23 months was dissociated with phalangeal bone age corresponding well to chronological age, but carpal bone age corresponding to 6 months. Her motor and mental milestones as well as expressive language were delayed. Growth at 26 months of age showed: weight +1 SD, height +1,5 SD, and head circumference +3–4 SD. During the first two years she had recurrent urinary and upper respiratory tract infections. At 29 months age hypermobility of joints, redundant skin, a small umbilical hernia, mild kyphoscoliosis and contractures of knees were observed. Due to the phenotype the girl was suspected of having Sotos syndrome.
Subtelomeric screening revealed the karyotype 46,XX.ish del(5)(q35). Quantitative PCR revealed that the NSD1 gene was deleted, confirming the diagnosis of Sotos syndrome.
Results
In this study we analyzed 132 patients (clinical features summarized in table I) by subtelomeric screening, where conventional chromosome analyses were normal in 130 cases. Table II summarizes the results. Seven cases (including two sisters) had cryptic aberrations not visible by conventional cytogenetic analysis. Both novel as well as previously described subtelomeric aberrations were identified. Furthermore, in two cases (patient 2 and patient 5) an unresolved visible small structural abnormality of chromosome 2 and 22, respectively, was suspected and resolved by subsequent subtelomere testing. In patient 2 the extra material was shown to be due to partial trisomy for the long arm of chromosome 22. Apparently no specific phenotype is associated with cases with this karyotype. However, anal atresia (observed in our patient) is present in patients with cat eye syndrome due to partial tetrasomy of the 22q11 region [11]. Patient 5 who was mildly retarded with few dysmorphic features was found to have a mosaic karyotype with partial trisomy 12p in about 25% of cells. More than two dozen patients with duplication of 12p have been described, mostly with severe mental retardation. Pallister-Killian syndrome is associated with tetrasomy (12p) mosaicism in fibroblasts [11]. Our patient did not resemble these phenotypes.
Patients 6a, 6b and 6c had both inherited the same submicroscopic chromosomal imbalance, I e partial monosomy 13qter and partial trisomy 5qter, due to an unbalanced translocation which segregated in balanced form through several generations in the family. The 3 affected relatives were severely retarded with microcephaly.
For patient 3 the extent of the deletion at 2q37 was delineated further with detailed FISH, prompted by the finding that the normal mother had apparently the same deletion with the probe set from Cytocell(R) which is recognized as a wellknown polymorphism [12]. In the child the most distal BAC clone, which was present on the abnormal chromosome 2 was RP11-1006P17, while the distal overlapping BAC clone RP11-473L20 was deleted. The distal breakpoint of the deletion was thus mapped within a 100 kb region (Human Genome Browser chromosome position 236,195,870-236,292,756) at cytogenetic band 2q37.2 and the deletion extended further to the telomere. The size of the deletion was thus approximately 6.8 Mb. Monosomy of 2q37 has been reported in more than 60 patients and recently deletion mapping of 20 cases has been published [13]. Monosomy 2q37 patients show significant clinical variability mainly dependent of the size of the deletion, though some degree of mental retardation and facial dysmorphism have been recognized in all patients, as it is also valid for the present case. In the present case the deletion breakpoint is mapped within or at the promoter region of CENTG2 gene between the microsatellite markers D2S336 and D2S338, which have been used in the study described by Aldred et al. [13]. In this study a critical interval for brachymetaphalangism, which is the main symptom of Albright hereditary osteodystrophy (AHO)-like brachymetaphalangism, has been assigned to the 3 Mb region from HDAC4 gene to the telomere. This region is deleted in the present patient, who does not present this feature. Brachymetaphalangism was suggested to be partially penetrant and some patients deleted for this region show other severe skeletal abnormalities. Our patient suffers osteoporosis, which also might be due to the mutations of the same gene leading to different degrees of symptoms. Recently Giardino et al. [14] published a patient with AHO-like syndrome where the deletion breakpoint of the patient was within BAC clone RP11-585E12, approximately 1.6–18 Mb distal to the breakpoint of the present case. This region includes five genes (CENTG2; GBX2; ASB18; CMKOR1; FLJ22527), which are deleted from the present case but not from the case described by Giardino et al. [14]. However at present it is difficult to predict the effects of the protein products on the different phenotypes observed in these patients. The phenotypically normal mother only had a deletion with probe D2S2986, but was not deleted with other probes. .
Discussion
Several conclusions can be drawn from the present study:
In 113 patients with both mental retardation and dysmorphic features we have identified subtelomeric abnormalities in 9 patients, and this result corresponds well with other studies. In a recent review of 20 studies the mean detection rate was found to be 4.8%, ranging between 0%-23% [9]. In our study 19 subjects with apparently non-syndromic idiopathic mental retardation, with or without positive family history were investigated. None of these patients showed rearrangements. This is in accordance with the study of Joyce et al. [15] who showed that cryptic telomeric rearrangements were not a significant cause of idiopathic mental retardation. It can be speculated that especially in familial cases of idiopathic mental retardation mendelian or multifactorial inheritance is more often the cause.
Our series represent referrals to a diagnostic laboratory for various reasons but undiagnosed mental retardation with dysmorphic features suggesting a "chromosomal phenotype" account for the majority. It is reasonable to assume that different detection rates are due to selection criteria, and it is likely that stringent selection criteria like the 5-item checklist (family history of MR, prenatal onset growth retardation, postnatal growth abnormalities, facial dysmorphic features, non-facial dysmorphism and/or congenital abnormalities) provided by de Vries et al. [16] will increase the detection yield of chromosome abnormalities. This can also be seen in the study reported by Walter et al. [17], where the authors have performed a subtelomere FISH study of 50 unrelated children ascertained by a checklist that evaluates MR or development delay, dysmorphism, growth defect, and abnormal pedigree and found ten causal rearrangements (detection rate of 20%).
The second conclusion is that some abnormalities are recurrent and not very rare.
For most of the cases a genotype-phenotype correlation was present. Patient 1, with deletion at 1pter had a phenotype correlating well with the 1p- syndrome, a now well-described syndrome and probably the most common terminal deletion syndrome [18-20]. However, it should be kept in mind that phenotype of these cases are partly dependent on the deletion size. The same conclusion can also be drawn for patient 3 with deletion at 2qter.
Patient 4 had deletion of 4pter including the Wolf-Hirshhorn syndrome region. This patient's phenotype was relatively mild and correlated with some patients previously described with similar 4p16.3 microdeletions [21,22]. Some patients have been designated Pitt syndrome (Pitt-Rogers-Danks syndrome, PRDS), but recently it was argued that Pitt and Wolf-Hirshhorn syndromes represent phenotypic variations of the same microdeletion [23].
Patient 7 had the karyotype, 46,XX.ish der(9)t(9;22)(q34.2;q13.2q13.3)pat. This patient was monosomic for distal 9q and trisomic for distal 22qter. An emerging phenotype of patients with distal 9q-deletions has been suggested, characterized by mental retardation, hypotonia, microcephaly, synophrys, short nose with anteverted nares, midface hypoplasia, a tented upper lip with a large, protruding tongue and sometimes neonatal teeth. Congenital heart disease and seizures are common complications [24]. These features were overlapping with those of the present case. It has also been suggested that 9qter deletions can cause syndromic obesity in children [24,25], but this was not the case for patient 7. Patient 8 had deletion of 5qter, which was in line with the clinical suspicion of Sotos syndrome. The NSD1 gene defective in Sotos syndrome is mapped at 5q35, 4.2 Mb from the telomere. Real-time quantitative PCR analysis demonstrated subsequently that the deletion in the patient comprised exon 5 of NSD1, confirming the clinical diagnosis. The deletion in our patient encompasses the subtelomeric region, the NSD1 gene and potentially further chromosomal material towards the centromere. Nagai et al. [26] observed that patients with a deletion involving the entire NSD1 gene and several other genes exhibited abnormal features of the CNS, the cardiovascular and the urinary systems, while these features were absent in patients with NSD1 point mutations. In addition to the features of Sotos syndrome, our patient showed several features consistent with Weaver syndrome [27]. A clear clinical distinction between these two syndromes is however difficult, and it is still discussed whether they result from allelic heterogeneity or are indeed distinct syndromes [28,29]. Recently, NSD1 mutations have been found in some cases of Weaver syndrome patients arguing for allelic heterogeneity [30].
A third conclusion is that a significant proportion of subtelomeric abnormalities are due to familial translocations, which is in accordance with previous studies [31,32]. This is important for genetic counseling.
In summary our study confirms the diagnostic value of subtelomeric screening, especially in mentally retarded subjects with dysmorphic features. A family history is of additional significance. Finally our study can contribute to the delineation of syndromes like the 1p- syndrome, the 2q- syndrome, the 4p- syndrome, the 9q- syndrome and the 5q- Sotos syndrome.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MS carried out genotype-phenotype analyses, participated in the design and coordination of the study and drafted the manuscript. ZT, SC and NT carried out the molecular analyses and FISH mapping of patient 3. JH supervised the subtelomeric and cytogenetic analyses. HH, BF, PL, VFP and PB contributed detailed clinical data and photos of patients. MD performed real-time quantitative PCR. KBN conceived of the study, and participated in its design and coordination. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We want to thank Jette Rasmussen for photographic assistance and Winni Petersen and Alla Jacobsen for expert technical assistance. The BAC clones were kindly provided by Vera Kalscheuer, Max-Planck-Institute for Molecular Genetics.
Figures and Tables
Figure 1 Facial features of patient 1 with 1p deletion (1 year old).
Figure 2 a,b,c,d. Patient 4 aged 3 months (a,b) and 13 years (c,d) with 4p deletion.
Figure 3 Pedigree for patients 6a & 6b & 6c illustrating segregation of t(5;13) translocation. Filled symbols are mentally retarded individuals, dotted symbols are carriers of the translocation. Arrows point to the two probands.
Figure 4 Patient 8 with Sotos syndrome (9 months old).
Table 1 Clinical features for the nine patients with rearrangements.
Case Dysmorphic facial features Other dysmorphic features/ abnormalities OFC (cm) BW (g) GA (weeks)
1 Yes Small hands and feet 31 2600 40
2 Yes Anal atresia - 2125 39
3 Yes No - 5500 41
4 Yes Dwarfism 28.6 2500 40
5 Yes No - - -
6 a Yes Scoliosis - 2900 40
6 b Yes Congenital dislocation of the hip - 3470 40
7 Yes Clubfoot, atrial septum defect, abundant head- and bodyhair 30.5 2770 42
8 Yes Adducted and broad thumbs, thickened skin in sole of foots and palms 40 3995 40
OFC, occiput frontal circumference; BW, birth weight; GA, gestational age
Table 2 Results of subtelomeric FISH for the nine patients with rearrangements.
Case Family history of MR Karyotype Deletion/duplication size Parents karyotype
1 No 46,XX.ish del(1)(p36.3) 8 Mb (a) Normal
2 No 46,XY.ish der(2)t(2;22)(q37.2;q1?) NA Mother normal / father NA
3 No 46,XX.ish del(2)(q37.2) 6.8 Mb Mother normal (e) / father NA
4 No 46,XX.ish del(4)(p16.1) Approx 4 Mb (d) Normal
5 No 46,XY/46,XY.ish der(22)t(12;22)(p13;p?) NA Normal
6a,b Yes 46,XX.ish der(13)t(5;13)(q35.2;q34) 3.9 Mb (13qter) and 6.5 Mb (5qter) (b) Mother: 46,XX,t(5;13)(q35.2;q34)
7 Yes 46,XX.ish der(9)t(9;22)(q34.2;q13.3)nat 4.1 Mb (9qter) and 5.7 Mb (22qter) (c) Father: 46,XY,t(9;22)(q34.2;q13.31)
8 No 46,XX.ish del(5)(q35) 4–7 Mb (f) Normal
(a),(b),(c), Cases 15, 16, 14, respectively, described by Schoumans et al. (2004); (d) also deleted for Wolf-Hirshhorn probe (Vysis ®); (e), Deletion D2S2986 polymorphism (see text for detail); (f), see text for detail; MR, mental retardation; NA, not analyzed.
==== Refs
Flint J Wilkie AO The genetics of mental retardation Br Med Bull 1996 52 453 464 8949249
Crow YJ Tolmie JL Recurrence risks in mental retardation J Med Genet 1998 35 177 182 9541099
Shevell M Ashwal S Donley D Flint J Gingold M Hirtz D Majnemeer A Noetzel M Sheth RD Quality standards subcommittee of the American Academy of Neurology Practice parameter: evaluation of the child with global developmental delay: report of the Quality Standards Subcommittee of the American Academy of Neurology and The Practice Committee of the Child Neurology Society Neurology 2003 60 367 380 12578916
Curry CJ Stevenson RE Aughton D Byrne J Carey JC Cassidy S Cunniff C Graham JM jrJones MC Kaback MM Moeschler J Schaffer GB Schwartz S Tarleton J Opitz J Evaluation of mental retardation: recommendations of a Consensus Conference: American College of Medical Genetics Am J Med Genet 1997 72 468 477 9375733 10.1002/(SICI)1096-8628(19971112)72:4<468::AID-AJMG18>3.0.CO;2-P
Saccone S De Sario A Della VG Bernardi G The highest gene concentrations in the human genome are in telomeric bands of metaphase chromosomes Proc Natl Acad Sci U S A 1992 89 4913 4917 1594593
van Karnebeek CD Koevoets C Sluijter S Bijlsma EK Smeets DF Redeker EJ Hennekam RC Hoovers JM Prospective screening for subtelomeric rearrangements in children with mental retardation of unknown aetiology: the Amsterdam experience J Med Genet 2002 39 546 553 12161591 10.1136/jmg.39.8.546
Koolen DA Nillesen WM Versteeg MH Merkx GF Knoers NV Kets M Vermeer S van Ravenswaaij CM de Kovel CG Brunner H Smeets D de Vries BB Sistermans EA Screening for subtelomeric rearrangements in 210 patients with unexplained mental retardation using multiplex ligation dependent probe amplification (MLPA) J Med Genet 2004 41 892 899 15591274 10.1136/jmg.2004.023671
Le Caignec C Boceno M Saugier-Veber P Jacquemont S Joubert M David A Frebourg T Rival JM Detection of genomic imbalances by array based comparative genomic hybridisation in fetuses with multiple malformations J Med Genet 2005 42 121 128 15689449 10.1136/jmg.2004.025478
de Vries BB Winter R Schinzel A Ravenswaaij-Arts C Telomeres: a diagnosis at the end of the chromosomes J Med Genet 2003 40 385 398 12807958 10.1136/jmg.40.6.385
Schoumans J Nielsen K Jeppesen I Anderlid BM Blennow E Brondum-Nielsen K Nordenskjold M A comparison of different metaphase CGH methods for the detection of cryptic chromosome aberrations of defined size Eur J Hum Genet 2004 12 447 454 15026784 10.1038/sj.ejhg.5201175
Schinzel A de Gruyter Catalogue of unbalanced chromosome aberrations in man. Second edition 2001 Berlin-New York
Knight SJ Flint J Perfect endings: a review of subtelomeric probes and their use in clinical diagnosis J Med Genet 2000 37 401 409 10851249 10.1136/jmg.37.6.401
Aldred MA Sanford RO Thomas NS Barrow MA Wilson LC Brueton LA Bohaglia MC Hennekam RC Eng C Dennis NR Trembath RC Molecular analysis of 20 patients with 2q37.3 monosomy: definition of minimum deletion intervals for key phenotypes J Med Genet 2004 41 433 439 15173228 10.1136/jmg.2003.017202
Giardino D Finelli P Gottardi G De Canal G Della MM Lonardo F Scarano G Larizza L Narrowing the candidate region of Albright hereditary osteodystrophy-like syndrome by deletion mapping in a patient with an unbalanced cryptic translocation t(2;6)(q37.3;q26) Am J Med Genet 2003 122A 261 265 10.1002/ajmg.a.20287
Joyce CA Dennis NR Cooper S Browne CE. Subtelomeric rearrangements: results of selected and unselected probands with idiopathic mental retardation and control individuals by using high-resolution G-banding and FISH Hum Genet 2001 109 440 451 11702226 10.1007/s004390100588
de Vries BB White SM Knight SJ Regan R Homfray T Young ID Super M McKeown C Splitt M Quarrell OW Trainer AH Niermeijer MF Malcolm S Flint J Hurst JA Winter RM Clinical studies on submicroscopic subtelomeric rearrangements: a checklist J Med Genet 2001 38 145 150 11238680 10.1136/jmg.38.3.145
Walter S Sandig K Hinkel GK Mitulla B Ounap K Sims G Sitska M Utermann B Viertel P Kalscheuer V Bartsch O Subtelomere FISH in 50 children with mental retardation and minor anomalies, identified by a checklist, detects 10 rearrangements including a de novo balanced translocation of chromosomes 17p13.3 and 20q13.33 Am J Med Genet 2004 128A 364 373 10.1002/ajmg.a.30160
Reish O Berry SA Hirsch B Partial monosomy of chromosome 1p36.3: characterization of the critical region and delineation of a syndrome Am J Med Genet 1995 59 467 475 8585567 10.1002/ajmg.1320590413
Rio M Molinari F Heuertz S Ozilou C Gosset P Raoul O Cormier-Daire V Amiel J Lyonnet S Merrer M Turleau C de Blois MC Prieur M Romana S Vekemans M Munnich A Colleaux L Automated fluorescent genotyping detects 10% of cryptic subtelomeric rearrangements in idiopathic syndromic mental retardation J Med Genet 2002 39 266 270 11950856 10.1136/jmg.39.4.266
Heilstedt HA Ballif BC Howard LA Lewis RA Stal S Kashork CD Bacino CA Shapira SK Shaffer LG Physical map of 1p36, placement of breakpoints in monosomy 1p36, and clinical characterization of the syndrome Am J Hum Genet 2003 72 1200 1212 12687501 10.1086/375179
Zollino M Lecce R Fischetto R Murdolo M Faravelli F Selicorni A Butte C Memo L Capovilla G Neri G Mapping the Wolf-Hirschhorn syndrome phenotype outside the currently accepted WHS critical region and defining a new critical region, WHSCR-2 Am J Hum Genet 2003 72 590 597 12563561 10.1086/367925
Anderlid BM Schoumans J Anneren G Sahlen S Kyllerman M Vujic M Hagberg B Blennow E Nordenskjold M Subtelomeric rearrangements detected in patients with idiopathic mental retardation Am J Med Genet 2002 107 275 284 11840483 10.1002/ajmg.10029
Zollino M Wright TJ Di Stefano C Tosolini A Battaglia A Altherr MR Neri G "Tandem" duplication of 4p16.1p16.3 chromosome region associated with 4p16.3pter molecular deletion resulting in Wolf-Hirschhorn syndrome phenotype Am J Med Genet 1999 82 371 375 10069706 10.1002/(SICI)1096-8628(19990219)82:5<371::AID-AJMG3>3.0.CO;2-J
Stewart DR Huang A Faravelli F Anderlid BM Medne L Ciprero K Kaur M Rossi E Tenconi R Nordenskjold M Gripp KW Nicholson L Meschino WS Capua E Quarrell OW Flint J Irons M Giampietro PF Schowalter DB Zaleski CA Malacarne M Zackai EH Spinner NB Krantz ID Subtelomeric deletions of chromosome 9q: a novel microdeletion syndrome Am J Med Genet 2004 128A 340 351 10.1002/ajmg.a.30136
Cormier-Daire V Molinari F Rio M Raoul O de Blois MC Romana S Vekemans M Munnich A Colleaux L Cryptic terminal deletion of chromosome 9q34: a novel cause of syndromic obesity in childhood? J Med Genet 2003 40 300 303 12676904 10.1136/jmg.40.4.300
Nagai T Matsumoto N Kurotaki N Harada N Niikawa N Ogata T Imaizumi K Kurosawa K Kondoh T Ohashi H Tsukahara M Makita Y Sugimoto T Sonoda T Yokoyama T Uetake K Sakazume S Fukushima Y Naritomi K Sotos syndrome and haploinsufficiency of NSD1: clinical features of intragenic mutations and submicroscopic deletions J Med Genet 2003 40 285 289 12676901 10.1136/jmg.40.4.285
Weaver DD Graham CB Thomas IT Smith DW A new overgrowth syndrome with accelerated skeletal maturation, unusual facies, and camptodactyly J Pediatr 1974 84 547 552 4366187
Cole T Growing interest in overgrowth Arch Dis Child 1998 78 200 204 9613346
Opitz JM Weaver DW Reynolds JF Jr The syndromes of Sotos and Weaver: reports and review Am J Med Genet 1998 79 294 304 9781911 10.1002/(SICI)1096-8628(19981002)79:4<294::AID-AJMG12>3.0.CO;2-M
Douglas J Hanks S Temple IK Davies S Murray A Upadhyaya M Tomkins S Hughes HE Cole TR Rahman N NSD1 mutations are the major cause of Sotos syndrome and occur in some cases of Weaver syndrome but are rare in other overgrowth phenotypes Am J Hum Genet 2003 72 132 143 12464997 10.1086/345647
Knight SJ Regan R Nicod A Horsley SW Kearney L Homfray T Winter RM Bolton P Flint J Subtle chromosomal rearrangements in children with unexplained mental retardation Lancet 1999 354 1676 1681 10568569 10.1016/S0140-6736(99)03070-6
Baker E Hinton L Callen DF Altree M Dobbie A Eyre HJ Sutherland GR Thompson E Thompson P Woollatt E Haan E Study of 250 children with idiopathic mental retardation reveals nine cryptic and diverse subtelomeric chromosome anomalies Am J Med Genet 2002 107 285 293 11840484 10.1002/ajmg.10159
| 15904506 | PMC1174871 | CC BY | 2021-01-04 16:03:32 | no | BMC Med Genet. 2005 May 17; 6:21 | utf-8 | BMC Med Genet | 2,005 | 10.1186/1471-2350-6-21 | oa_comm |
==== Front
BMC Med GenetBMC Medical Genetics1471-2350BioMed Central London 1471-2350-6-221591068610.1186/1471-2350-6-22Research ArticleSMN1 dosage analysis in spinal muscular atrophy from India Kesari Akanchha [email protected] Hanna [email protected] Debra GB [email protected] Balraj [email protected] Department of Genetics, Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow-14, U.P, India2 Department of Pathology and Laboratory Medicine, University of Pennsylvania, Philadelphia USA3 Current Address- Center for Genetic Medicine, Children's National Medical Center Washington- DC. USA4 Department of Pathology and Laboratory Medicine, Newyork Presbyterian Hospital, Cornell Campus, Newyork USA2005 23 5 2005 6 22 22 18 9 2004 23 5 2005 Copyright © 2005 Kesari et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Spinal muscular atrophy (SMA) represents the second most common fatal autosomal recessive disorder after cystic fibrosis. Due to the high carrier frequency, the burden of this genetic disorder is very heavy in developing countries like India. As there is no cure or effective treatment, genetic counseling becomes very important in disease management. SMN1 dosage analysis results can be utilized for identifying carriers before offering prenatal diagnosis in the context of genetic counseling.
Methods
In the present study we analyzed the carrier status of parents and sibs of proven SMA patients. In addition, SMN1 copy number was determined in suspected SMA patients and parents of children with a clinical diagnosis of SMA.
Results
wenty nine DNA samples were analyzed by quantitative PCR to determine the number of SMN1 gene copies present, and 17 of these were found to have one SMN1 gene copy. The parents of confirmed SMA patients were found to be obligate carriers of the disease. Dosage analysis was useful in ruling out clinical suspicion of SMA in four patients. In a family with history of a deceased floppy infant and two abortions, both parents were found to be carriers of SMA and prenatal diagnosis could be offered in future pregnancies.
Conclusion
SMN1 copy number analysis is an important parameter for identification of couples at risk for having a child affected with SMA and reduces unwarranted prenatal diagnosis for SMA. The dosage analysis is also useful for the counseling of clinically suspected SMA with a negative diagnostic SMA test.
==== Body
Background
With a prevalence of 1 in 10,000 live births and a carrier frequency of approximately 1 in 50 [1], proximal spinal muscular atrophy (SMA) represents the second most common fatal autosomal recessive disorder after cystic fibrosis [2]. SMA is characterized by the degeneration of anterior horn cells of the spinal cord, resulting in progressive weakness. The condition is clinically heterogeneous and has been divided into four subtypes according to age of onset and clinical severity [3]. Molecular genetic analysis has mapped all four forms of childhood and adult SMA to chromosome 5q11.2-q13.3, suggesting that they are allelic disorders.
In a majority of normal individuals in the population, survival motor neuron (SMN) genes are present in at least one telomeric (SMN1) and one centromeric (SMN2) copy per chromosome. However, 26% of all normal chromosomes 5 lack SMN2 copy of the gene. The two SMN genes are highly homologous but a single nucleotide variation in exon 7 of SMN1 and SMN2 genes is responsible for functional differences [4]. The majority of SMA patients, irrespective of their clinical types, have homozygous deletion of the SMN1 gene [5]. In addition intragenic mutations have been identified in most of the patients who have only one copy of SMN1 gene, confirming the involvement of SMN1 in the pathogenesis of SMA [6]. Normal individuals with one copy of the SMN1 gene are carriers for this autosomal recessive disorder.
The single nucleotide differences in exons 7 and 8 are used to distinguish SMN1 and SMN2 in diagnostic and prenatal testing for SMA [7,8]. Although this methodology can detect homozygous absence of SMN1, it cannot differentiate the presence of one copy from two or more copies of SMN1. In recent years, molecular diagnostic testing for SMN1 copy number by dosage analysis has been developed [9,10], and is used to determine the carrier status for SMA in the majority of cases. However, in 3.7% of carriers, referred as "2+0" carriers, two copies of the SMN1 gene are present on one chromosome 5 and a deletion of SMN1 allele is present on the other chromosome 5 [11]. This situation cannot be distinguished by dosage analysis alone. In such cases dosage analysis in combination with linkage analysis for extended family members may be required to unequivocally determine whether an individual is "2+0" (a carrier) or "1+1" (not a carrier) [10].
Due to high carrier frequency, the burden of this genetic disorder is very heavy and genetic counseling is an active component in the disease management. In many cases the index patient dies without proper clinical diagnosis due to limited facilities and lack of awareness in primary care providers in India [12]. Therefore, carrier testing is particularly useful to direct genetic counseling for prenatal diagnosis. This is the first study from India using testing for the carrier status of parents and sibs of patients with a proven diagnosis of SMA. Additionally, SMA carrier testing has been performed for suspected SMA patients and parents of children clinically suspected to have SMA.
Methods
Subjects
Detailed clinical history and pedigree was drawn for all the cases and their families. Clinical examinations of patients (if alive) and family members were carried out in the Genetics Outpatient Department at Sanjay Gandhi Post Graduate Institute of Medical Sciences. Informed consents were obtained from all study members following genetic counseling. Twenty-nine individuals included in the study were divided into three categories. Group I comprised 14 parents and sibs of SMN1 gene deletion positive cases. Group II comprised 10 individuals, four couples, and two sibs whose deceased child (or children) or sibs were suspected to have SMA. Group III comprised 5 patients with a strong clinical suspicion of SMA but did not show homozygous deletions of exons 7 and 8 of the SMN1 gene. Five ml blood was collected in EDTA tubes from all individuals enrolled in the study.
DNA isolation and PCR
DNA was extracted from blood samples using DNA extraction kit (Qiagen, Germany) according to the manufacturer's instructions. Patients were tested for homozygous deletion of exons 7 and 8 of SMN1 gene by a Polymerase Chain Reaction-Restriction Fragment Length Polymorphism (PCR-RFLP) method as described previously [13].
SMN1 copy number assay
The SMA carrier test determines the SMN1 gene copy number by comparing the levels of amplified products generated from exon 7 of SMN1 and a two-copy control gene following PCR amplification. Multiplex PCR of exon 7 SMN1 genomic sequences, CFTR exon 4 genomic sequences, and modified SMN1 and CFTR internal standard (IS) was performed using hexachloroflorescein labeled primers. Primer sequences and PCR amplification conditions were according to McAndrew et al (1997) [9] and Chen et al (1999) [10]. Briefly, PCR was performed in a 50 μl reaction mix containing 10 mM Tris-HCl (pH-8.3) PCR buffer, 1.5 mM MgCl2, 50 mM KCl, 0.01% (w/v) gelatin (Applied BioSystems Inc., (ABI) CA, USA), 1 mM spermidine, 0.05 mM each primer, 400 mM each dNTP, 1 unit Ampli Taq Polymerase (ABI) and 200 ng of genomic DNA. Internal Standards at approximately equimolar concentration to the genomic templates (6.4 × 104 copies) were added to each reaction tube. PCR amplification conditions were as follows: initial denaturation at 95°C for 5 min followed by 22 cycles of denaturation for 30 sec at 95°C, annealing for 30 s at 55°C, and elongation for 2 min at 72°C, followed by a final extension step for 10 min at 72°C. Each patient sample was analyzed in duplicate, and controls included 5 known two-copy, a one-copy and a homozygous deleted DNA samples, as well as water (no DNA template control).
After PCR, 4 μl of amplified DNA from each reaction was digested in a 6 μl reaction volume with 1 unit of Dra I (New England Biolabs, USA) for 3 hours at 37°C. Two μl of the Dra I digestion were mixed with 4 μl of formamide, 0.5 μl of Gene Scan-500 Rox (ABI) and 0.5 μl of loading buffer, heated at 95°C for 5 min and then placed on ice for 2 min. Digested PCR product mixtures were analyzed on an ABI 373 Genetic Analyzer instrument (ABI). Peak areas of the various products were determined using ABI Gene Scan 672 Software (ABI).
The SMN1 copy number was calculated using peak areas as follows: (SMN1 genomic/SMN1-IS) / (CFTR genomic/ CFTR-IS). Ratios were then normalized to the mean of five control samples with two copies of SMN1 gene.
Results
Fig 1 shows the pedigree for Case 3 (Fig 1A), results of PCR-RFLP for exons 7 (Fig 1B) and 8 (Fig 1C) of SMN1 gene, and SMN1 dosage analysis gel scans (Fig 1D). The index patient had homozygous deletion of both exons 7 and 8 of SMN1 gene as shown by the absence of 188 bp and 187 bp bands, respectively (Fig. 1B and 1C, lanes 2 and 2, respectively). Dosage analysis gel scans of the father, mother, and sib are shown with band sizes (in base pairs) and peak areas under each peak (Fig 1D). The number of SMN1 gene copies was calculated as described in Methods. Normalized results (the mean of five control samples with two copies of the SMN1 gene) were consistently within the ranges of 0.8–1.2 for normal controls with two copies of SMN1 gene.
Fourteen samples in Group I (Cases 1–6) were from the parents and sibs of SMA patients with homozygous deletion of both exons 7 and 8 of SMN1 gene. Out of these, all parents and three of the four sibs had one SMN1 gene copy and one sib had two SMN1 gene copies (Table 1). In Group II (Cases 7–10), carrier analysis was performed on the basis of family history of SMA. In this group, five parents had two SMN1 copies (Cases 7, 8, 9) and three parents had one SMN1 gene copy (Case 9 and 10). Case 10 had two children and they showed two copies of SMN1 gene, respectively (Table 1). Five samples in Group III were analyzed to determine the copy number of SMN1 gene as these patients had strong clinical features consistent with SMA but did not have homozygous deletion of exons 7 and 8 of the SMN1 gene. Out of these, only a single patient had one copy of SMN1 gene, and the remaining four cases had two copies of SMN1 gene (Data not shown).
Discussion
Twenty-nine samples were analyzed by quantitative PCR to determine the number of SMN1 gene copies present, and 17 of these were found to have one SMN1 gene copy. In the previous study, it was reported that 94.3% of normal individuals had two SMN1 copies and 2.1%, 0.7% and 2.9% had three, four and one copy, respectively [14]. Only one SMN1 gene copy is sufficient for normal functioning in an individual, as all parents with one copy of the SMN1 gene are asymptomatic.
From our small group of SMA cases, parents of confirmed SMA patients were obligate carriers of the disease and this was confirmed by SMA carrier testing. However, parents of children with SMA may not always be carriers as de-novo deletions of the SMN1 gene occurs in more than 2% of patients, with SMA [10,15]. Presence of de-novo deletion in the family lowers the recurrence risk for the couples from 25% to the risk of a second de-novo mutation which is very low. Knowledge of the carrier status of parents of affected children is useful for determining if a de-novo mutation has occurred and establishing the couple's future risk of having an affected child. If the parents are found to be carriers, then carrier testing can be offered to the siblings of the parents, who have a 50% risk of also being a carrier for SMA [1].
It has been reported that a small proportion of parents may have a "2+0" genotype in which there are two SMN1 gene copies on one chromosome and none on the other. In such cases a normal dosage analysis should be followed by linkage analysis of the family in order to try to distinguish between individuals who carry one SMN1 gene on each chromosome and those with a two-copy allele [1]. The recurrence risk for the family in which the "2+0" genotype is present is 25%.
Copy number analysis is also useful for testing of patients with a clinical diagnosis of SMA who are negative by a SMA diagnostic test that looks for homozygous deletion of exon 7 and 8 of the SMN1 gene. SMA is one of a wide spectrum of muscle and nerve disorders such as Becker muscular dystrophy, myotonic dystrophy, and Charcot-Marie tooth disease (Type IA) [16] that affect infants and young children. Clinical symptoms among these disorders are overlapping and may not be sufficiently specific to make a reliable clinical diagnosis of SMA. Patients that are referred for molecular genetic testing generally have symptoms, like hypotonia, floppiness, proximal muscle weakness, and loss of ambulation. These symptoms are not specific to 5q-linked SMA and in infants there are additional complications. In such circumstances molecular genetic testing for SMA is the best method to confirm the clinical diagnosis [17,18]. Muscle testing (EMG examination) at times may be difficult to perform in a neonate.
Homozygous deletion of the SMN1 gene confirms the diagnosis of 5q-linked SMA. If no homozygous deletion is detected, then SMN1 copy number analysis may be used if there is a strong clinical suspicion of SMA, since there is a 1 in 50 carrier frequency in the general population. An individual with signs and symptoms suggestive of SMA who does not lack at least one copy of the SMN1 gene is less likely to have 5q-linked SMA. For example, in the Group III patients, four of five had two copies of SMN1 gene and therefore were unlikely to have 5q-linked SMA. In such situations, the possibility of related disorders other then 5q-linked SMA should be considered. However, in patients with one copy of the SMN1 gene, the chances of a non-deletion type of mutation in the other allele may be explored [1,6].
The most severe form of SMA occurs at birth or in early infancy and may be difficult for primary care providers in India to diagnose. No data are available on the population prevalence of SMA and the status of diagnosis of SMA from India, due to the limited number of centers and the high cost and complexity of the molecular genetic test. So in many cases the child usually expires before the diagnosis is confirmed and the parents approach with a history of a previous child's death with symptoms consistent with SMA. In absence of a sample for molecular genetic testing for SMA, the information obtained by SMN1 copy number analysis for the parents can be utilized to confirm the diagnosis for the deceased child and to offer prenatal diagnosis for future pregnancies. The presence of one copy of the SMN1 gene in the parents will confirm their carrier status and prenatal diagnosis can be clearly advised in subsequent pregnancies. The family in case 10 had a history of a deceased floppy infant and two abortions. No surviving affected child was available. After SMN1 copy number analysis, both parents were found to have one SMN1 copy of the gene. Prenatal diagnosis was offered to the family and the fetus was found to be normal. Another child was born in the family and he was normal and dosage analysis showed two copies of SMN1 gene.
If both parents have two or more copies of SMN1 gene, as for the parents of Cases 7 and 8, then prenatal diagnosis for SMA will not be useful for future pregnancies. However, if only one of the parents has one SMN1 copy, as was seen in Case 9, then additional testing could be performed to clarify whether the two-copy parent is 1+1 or 2+0 using carrier testing and linkage analysis of additional family members. If the two-copy parent is 2+0, then prenatal diagnosis using the diagnostic test is still useful. If the two-copy parent is 1+1, then the affected child will usually have one SMN1 copy, and dosage analysis can be used for prenatal diagnosis, with a two-copy result indicating the fetus will not be affected. If the fetus has one SMN1 copy, then linkage analysis can be used, if DNA is available from an affected child. If only one parent has one copy of SMN1 gene, the situation may be clearly discussed with the parents before any prenatal testing is considered.
Conclusion
Our results confirm that SMN1 copy number analysis is an important parameter for identification of couples at risk for having a child affected with SMA and reduces unwarranted prenatal diagnosis for SMA. Copy number analysis is also useful in the setting of clinically suspected SMA with a negative diagnostic SMA test.
Competing interests
The author(s) declare that they have no competing interests
Authors' contributions
AK and HR carried out the molecular genetic studies, participated in analyzing the data & drafted the manuscript. DGBL helped in analyzing the data and gave valuable suggestions in preparing the manuscript. BM participated in the designing of the study & manuscript. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
The authors are thankful to Dr Shubha Phadke, Department of Genetics, and Drs U. K. Mishra and S. Pradhan, Department of Neurology, SGPGIMS Lucknow for referring the patients for the study.
Figures and Tables
Figure 1 SMN1 copy number analysis for the family of Case 3. 1A The pedigree of the family of Case 3. 1B: Exon 7 PCR-RFLP Polyacrylamide gels showing the undigested and digested products of SMN exon 7 after Dra I digestion. Lane 1, undigested product of patient; Lane 2, digested product of patient with SMN1 deletion; Lane 3, undigested product of sibling; Lane 4, digested product of sibling with no SMN1 deletion. 1C: Exon 8 PCR-RFLP after Dde I digestion. Lanes 1 and 3, undigested PCR product; Lane 2, digested product of patient with homozygous deletion of exon 8 of the SMN1 gene; Lane 4, digested product of sibling with no homozygous deletion of SMN1. 1D Quantitative PCR gel scans of Father (I: 1), Mother (I: 2) and sibling (II: 4). Band sizes in base pairs are shown in the upper boxes and peak areas in the lower boxes. IS, internal standards.
Table 1 Copy number of SMN1 gene in parents, sibs and SMA patients
Case number Subjects Copy number of SMN1 gene
Group I
Case 1 P 0
F 1
M 1
S 1
Case 2 P 0
F 1
M 1
Case 3 P 0
F 1
M 1
S 2
Case 4 P 0
F 1
M 1
Case 5 P 0
F 1
M 1
Case 6 P 0
S1 1
S2 1
Group II
Case 7 F 2
M 2
Case 8 F 2
M 2
Case 9 F 1
M 2
Case 10 F 1
M 1
S1 2
S2 2
P: Proband; F: Father; M: Mother; S: Sibling
==== Refs
Ogino S Wilson RB Genetic testing and risk assessment for spinal muscular atrophy Hum Genet 2002 111 477 500 12436240 10.1007/s00439-002-0828-x
Scriver CR Beaudet AL Sly WS Valle D The metabolic and molecular bases of inherited disorder 2001 231 8 McGraw-Hill: New York 5833 5843
Biros I Forrest S Spinal muscular atrophy: Untangling the knot J Med Genet 1999 36 1 8 9950358
Rochette CF Gilbert N Simard LR SMN gene duplication and the emergence of the SMN2 gene occurred in distinct hominids: SMN2 is unique to Homo sapiens Hum Genet 2001 108 255 66 11354640 10.1007/s004390100473
Panigrahi I Kesari A Phadke SR Mittal B Clinical and Molecular Diagnosis of Spinal Muscular Atrophy Neurol India 2002 50 117 22 12134171
Wirth B An update of the mutation spectrum of the survival motor neuron gene (SMN1) in autosomal recessive spinal muscular atrophy (SMA) Hum Mutat 2000 15 228 237 10679938 10.1002/(SICI)1098-1004(200003)15:3<228::AID-HUMU3>3.0.CO;2-9
Somerville MJ Hunter AGW Aubry HL Korneluk RG Mac Kenzie AE Surh LC Clinical application of the molecular diagnosis of spinal muscular atrophy Am J Med Genet 1997 69 159 165 9056553 10.1002/(SICI)1096-8628(19970317)69:2<159::AID-AJMG8>3.0.CO;2-K
Van der Steege G Grootscholten PM Van derrVlies P Draaijers TG Osinga J Cobben JM Scheffer H Buys CHC PCR-based DNA test to confirm clinical diagnosis of autosomal recessive spinal muscular atrophy Lancet 1995 345 985 986 7715313 10.1016/S0140-6736(95)90732-7
McAndrew PE Parsons DW Simard LR Rochette C Ray PN Mendell JR Prior TW Burghes AH Identification of proximal spinal muscular atrophy carriers and patients by analysis of SMNT and SMNC gene copy number Am J Hum Genet 1997 60 1411 22 9199562
Chen KL Wang YL Rennert H Joshi I Mills JK Leonard DG Wilson RB Duplications and de novo deletions of the SMNt gene demonstrated by fluorescence-based carrier testing for spinal muscular atrophy Am J Med Genet 1999 85 463 9 10405443 10.1002/(SICI)1096-8628(19990827)85:5<463::AID-AJMG6>3.0.CO;2-V
Ogino S Leonard DG Rennert H Ewens WJ Wilson RB Genetic risk assessment in carrier testing for spinal muscular atrophy Am J Med Genet 2002 110 301 307 12116201 10.1002/ajmg.10425
Kesari A Mittal B Genetic Testing in pediatric Practice Pediatric Today 2004
Kesari A Mukherjee M Mittal B Mutation analysis in spinal muscular atrophy using allele-specific polymerase chain reaction Indian J Biochem Biophys 2003 40 439 441
Feldkotter M Schwarzer V Wirth R Wienker TF Wirth B Quantitative analyses of SMN1 and SMN2 based on real-time light Cycler PCR: fast and highly reliable carrier testing and prediction of severity of spinal muscular atrophy Am J Hum Genet 2002 70 358 368 11791208 10.1086/338627
Zeesman S Whelan DT Carson N McGowan-Jordan J Stockley TL Ray PN Prior TW Parents of children with spinal muscular atrophy are not obligate carriers: carrier testing is important for reproductive decision-making Am J Med Genet 2002 107 247 249 11807908 10.1002/ajmg.10132
Markowitz JA Tinkle MB Fischbeck KH Spinal muscular atrophy in the neonate J Obstet Gynecol Neonatal Nurs 2004 33 12 20 14971549 10.1177/0884217503261125
Mailman MD Heinz JW Papp AC Snyder PJ Sedra MS Wirth B Burghes AH Prior TW Molecular analysis of spinal muscular atrophy and modification of the phenotype by SMN2 Genet Med 2002 4 20 6 11839954
Mishra VN Kalita J Kesari A Mittal B Shankar SK Misra UK A clinical and genetic study of spinal muscular atrophy Electromyogr Clin Neurophysiol 2004 44 307 12 15378871
| 15910686 | PMC1174872 | CC BY | 2021-01-04 16:03:34 | no | BMC Med Genet. 2005 May 23; 6:22 | utf-8 | BMC Med Genet | 2,005 | 10.1186/1471-2350-6-22 | oa_comm |
==== Front
BMC MicrobiolBMC Microbiology1471-2180BioMed Central London 1471-2180-5-341593263310.1186/1471-2180-5-34Methodology ArticleOptimization of culture conditions to obtain maximal growth of penicillin-resistant Streptococcus pneumoniae Restrepo Andrea V [email protected] Beatriz E [email protected] María [email protected] Carlos A [email protected] Andres F [email protected] Omar [email protected] GRIPE: Grupo Investigador de Problemas en Enfermedades Infecciosas, University of Antioquia Medical School, Medellín, Colombia2 Section of Infectious Diseases, Department of Internal Medicine, University of Antioquia Medical School, Medellín, Colombia3 Department of Pharmacology & Toxicology, University of Antioquia Medical School, Medellín, Colombia2005 2 6 2005 5 34 34 25 2 2005 2 6 2005 Copyright © 2005 Restrepo et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Streptococcus pneumoniae, particularly penicillin-resistant strains (PRSP), constitute one of the most important causes of serious infections worldwide. It is a fastidious microorganism with exquisite nutritional and environmental requirements to grow, a characteristic that prevents the development of useful animal models to study the biology of the microorganism. This study was designed to determine optimal conditions for culture and growth of PRSP.
Results
We developed a simple and reproducible method for culture of diverse strains of PRSP representing several invasive serotypes of clinical and epidemiological importance in Colombia. Application of this 3-step culture protocol consistently produced more than 9 log10 CFU/ml of viable cells in the middle part of the logarithmic phase of their growth curve.
Conclusion
A controlled inoculum size grown in 3 successive steps in supplemented agar and broth under 5% CO2 atmosphere, with pH adjustment and specific incubation times, allowed production of great numbers of PRSP without untimely activation of autolysis mechanisms.
==== Body
Background
Streptococcus pneumoniae, one of the most important human pathogens of all times, is today an even more serious menace because of its increasing resistance to penicillin, the first-choice antibiotic [1]. It is a fastidious organism to grow both in vitro and in animal infection models. Probably because of the wide genetic diversity of this organism and its exclusive nature as human pathogen, it is difficult to obtain comparable cell production when culturing different strains even within the same serotype [2].
In the field of experimental animal models there is also wide variation in the susceptibility of mice strains and specific organs to infection with S. pneumoniae [3,4]. While new drugs with activity against penicillin-resistant S. pneumoniae (PRSP) are being discovered, only penicillin-susceptible strains grow in the lungs of experimental animals with some consistency [5]. Although other organs such as the thighs of neutropenic mice are susceptible to progressive infection with PRSP, a reproducible pneumonia model remains elusive for most strains of PRSP.
Little attention has been given to the culture conditions of S. pneumoniae, critical to obtain the appropriate amount of viable log-phased cells to inoculate the animals. S. pneumoniae was reclassified within the group of anaerobic bacteria because its relatively small genome simply lacks many genes required for aerobic growth [6]. As a Gram-positive catalase-negative coccus that generates H2O2 via a flavoenzime system, it grows better in presence of a source of catalase such as red blood cells [7,8]. In contrast to most bacteria, S. pneumoniae also requires choline for growth in defined media, and reducing agents are almost essential. Growth of most strains requires vitamin B complex (biotin, nicotinamide, pantothenate, pyridoxal, riboflavin and thiamine), adenine, guanine, uracil, and 7–10 amino acids. Its energy metabolism is fermentative, yielding primarily low levels of lactic acid, but optimum pH for growth is 7.8 with a range of 6.5–8.3 [9]. Besides these limitations, S. pneumoniae displays a particularly effective quorum sensing system that activates several potent autolysins once certain conditions are met within the growing population, and this is perhaps the most important hurdle to overcome when trying to obtain viable log-phased cells [10]. Among several factors not well understood, culture acidification is one of the conditions met by the growing population that clearly contributes to start the autolysis process [11].
We hypothesized that optimization of culture conditions tending to prevent environmental acidification and its consequent activation of autolysis mechanisms could lead to better production of PRSP in vitro, and aimed to determine and provide such ideal conditions for the growth of this microorganism, so they could be applied to diverse strains of clinical importance to develop a reproducible animal model of pneumonia.
Results
Susceptibility testing
Minimal inhibitory (MIC) and bactericidal (MBC) concentrations of six antimicrobial agents for these strains are shown in Table 1. Reduced susceptibility to penicillin was demonstrated for 6 strains, the other 2 were susceptible. Among those non-susceptible, 4 strains displayed intermediate (MIC 0.12–1 μg/mL) and 2 full resistance to penicillin (MIC > 1 μg/mL). Only one strain (INS-E685) was susceptible to all six antibiotics tested, but all of them were susceptible to erythromycin and vancomycin. MIC of ceftriaxone against PRSP INS-E611, E674 and E676 was 1 μg/mL, a value conferring intermediate resistance for CSF isolates, as was the case for strain INS-E676 (Table 1).
Evaluation of baseline culture variables
The results are summarized in Table 2. For cryoprotection, skim milk invariably allowed recuperation of frozen organisms, while 17% glycerol failed once (no viable bacteria after thawing). During Phase 0, we found that two successive passes on solid media were necessary for complete recuperation after thawing, and that their optimal incubation time was 15 hours. After 15 hours colony umbilication became deeper, suggesting progression of the autolysis process [9]. The number of colonies from Phase 0 employed to inoculate Tube 1 in Phase 1 was critical to maximize the number of viable log-phased cells: 10 colonies into 10 ml of culture broth gave the best results. Since broth acidification is known to start autolysis, a low pH after 12 hours of incubation during Phase 1 was used as an indicator of poor viability. Compared with 10 colonies, 5 or less produced lower cell numbers with similar decrease in pH, and 15 or more colonies gave greater cell numbers but with a profound decrease in pH (Table 3). Besides 5% sheep blood, additional supplementation of trypticase soy agar (TSA) with 0.5% yeast extract (instead of 0%) during Phase 0 increased the number of viable cells produced by broth culture in Phase 1 from 6.60–7.60 to 8.32–8.63 CFU/ml.
The type of liquid media was tested before (standard broth) and after diverse adjustments (supplemented broth). Standard Todd Hewitt Broth (THB) was more reliable than Brain Heart Infusion (BHI) made by Oxoid and BBL, because it consistently produced 1.63–2.64 log10 CFU/ml per hour with all strains and lots tested. BBL-BHI production was low (1.34–1.63 log10 CFU/ml per hour) from the beginning and not used further. BHI Oxoid was similar to THB in cell production, but not reproducible with the different lots tested to grow strain INS-E611 (Figure 1). Analysis of these four media by repeated measures ANOVA demonstrated a non-significant difference (P = 0.7976), as expected from the similarity in production during the first few hours of cultivation. However, as illustrated by Figure 2, productivity at the end of each culture (5 hours) was significantly better with supplemented THB compared with standard THB and BHI (P = 0.0022, one way ANOVA). While stationary and death phase with ≤ 8.0 log10 CFU/ml was reached before 3 hours with standard broths, supplementation of THB consistently produced ≥ 9.0 log10 CFU/ml and prolonged logarithmic growth during Phase 2 up to 4 hours in Tube 6, 5 hours in Tube 7, and 7 hours in Tube 8 (Figure 3). Incubation time during Phase 1 was best at 12 hours, with longer periods resulting in poor growth during Phase 2, mainly determined by media acidification. Adjustment of pH was fundamental during both Phase 1 and Phase 2. For practical reasons, Phase 1 broth was adjusted at the beginning of the 12-hour culture, but Phase 2 broth was adjusted every hour. In Figures 1, 2, 3, the term "supplemented" also implies this protocol for pH adjustment.
Supplementation and optimization of culture conditions for PRSP
Optimized culture conditions produced ≥ 9 log10 CFU/ml of log-phased cells in Phase 2. These results were reproducible with all PRSP strains tested (Figure 4) under the following protocol:
Phase 0: from frozen stock (skim milk), cells are resuscitated by two consecutive passes on 5% sheep blood TSA supplemented with 0.5% yeast extract and incubated for 15 hours under 5% CO2 atmosphere.
Phase 1: 10 colonies from the second plate are inoculated into 10 ml of supplemented THB (Tube 1), followed by 4 successive 1:10 dilutions into identical liquid media (Tubes 2–5). Supplementation implies pH adjustment to 7.8 with 1N NaOH after addition of 2% yeast extract and 2.5% horse blood. All 5 tubes are incubated for 12 hours under 5% CO2 atmosphere, without additional pH adjustment.
Phase 2: 1 ml of bacterial suspension taken from the most diluted tube with visible growth from Phase 1 is inoculated into 9 ml of supplemented THB (Tube 6) in Phase 2. With most strains and under conditions just described for Phases 0 and 1, Tube 5 or 4 should have visible growth by the end of Phase 1. Tubes 3, 2, and 1 should not be used, because their pH is invariably acidic by the end of Phase 1, usually giving very low cell counts during Phase 2. After inoculation of Tube 6, two successive 1:10 dilutions are made into identical media (Tubes 7 and 8). Each one of these 3 tubes is incubated under 5% CO2 atmosphere, and their pH is adjusted every hour to 7.8 with 1N NaOH. We found that optimal incubation periods were 4, 5 and 7 hours for Tubes 6, 7, and 8, respectively.
Discussion
Our results demonstrate that optimization of culture conditions is an easy, non-expensive and reproducible way to attain penicillin-resistant S. pneumoniae growth over 9.5 log10 CFU/ml without the need of more complex methods (e.g. chemostat).
The S. pneumoniae serotypes used in this study are representative of the most frequent invasive isolates from blood and CSF in Colombia [20,21].
The three standard media used initially (BHI-Ox, BBL-BHI and THB) are recommended by American Society for Microbiology and Clinical Laboratory Standards Institute (CLSI, formerly NCCLS) for S. pneumoniae growth, but their productivity was quite different. In contrast with BHI, THB has soy bean peptone and higher pH (7.4 vs 7.8, respectively). Conversely, BBL-BHI has the highest dextrose concentration (3 g/L), which results in marked acidification of the media as anaerobic growth progresses [2]. In this aspect, we confirmed previous findings of acidic pH (under 6.5) inhibiting growth of penicillin resistant and susceptible S. pneumoniae strains [12]. To avoid media acidification, we limited the number of colonies produced during Phase 0 to inoculate supplemented THB in Tube 1 (Phase 1) and adjusted pH to 7.8 at the beginning of Phase 1 and every hour during Phase 2. Since pH control is not enough to attain growth over 8.7 log10 CFU/ml [12], we addressed two additional characteristics that influence S. pneumoniae growth: autolysis and competence. The process of autolysis takes place around the upper part of the logarithmic phase of growth, is started by pheromones liberated by a complex quorum sensor system, and is characterized by extensive cellular death [10,11]. Competence, or the capacity for genetic transformation, is well developed in S. pneumoniae, and particularly well in PRSP, but it does not happen without repercussion in bacterial growth, both in vitro and in vivo [3,5,13]. Although the relationship between autolysis and competence not always goes in the same direction, choline mutants (the target of most autolysins) have shown important changes respect to wild type pneumococci, including abnormal growth in long chains instead of diplococci, loss of competence for transformation, resistance to autolysis, and resistance to cell-wall active antibiotics [14]. Based on these facts, we hypothesized that prevention of autolysis could improve production of viable cells even in the logarithmic phase of bacterial growth. Besides offering a suitable culture media, shorter incubation times and separate culture process in three phases allowed generous bacterial growth before triggering autolysis mechanisms. Given that concentrations of 8 log10 CFU/ml were obtained 1–2 hours before reaching the maximum incubation periods shown in Figure 3, inoculation of healthy cells in large numbers is possible before triggering autolysis and, probably, while pneumococci are still virulent enough to induce acute pneumonia in animal models.
Goncalves et al specifically addressing the optimal growing conditions for S. pneumoniae, described a batch cultivation procedure for serotype 23F, suitable for large scale polysaccharide production to manufacture vaccines [2]. They used Hoeprich's medium with different glucose concentrations and supplemented with various amino acids and a vitamin solution, and provided an anaerobic environment with N2 or CO2. Their results, expressed in terms of biomass and polysaccharide production, are difficult to compare with ours. Moreover, they used autolysis as a way of releasing polysaccharide, whereas we tried to avoid this process to obtain viable bacteria.
This is the first report on culture optimization for maximum growth of PRSP with simple methods, and is directed to facilitate the use of these strains in experimental procedures such as the development of animal models of pneumonia. The cultivation conditions described here generate high concentrations of log-phased bacteria without early and inconvenient induction of autolysis.
Conclusion
In this study we demonstrate that a standardized inoculum grown in supplemented solid and liquid media with pH adjustment and control of incubation times in three phases produces viable PRSP far beyond the limiting 8 log10 CFU/mL. This method should allow improvement in experimental approaches to solve important questions regarding the biology, pathology, and therapeutics of PRSP.
Methods
Bacterial strains
Eight clinical strains of Streptococcus pneumoniae, obtained from very sick patients nationwide, were supplied by the Colombian Instituto Nacional de Salud (INS). These included six strains non susceptible to penicillin (called further penicillin-resistant): INS-E611, E674 (blood isolates), E676, E678, E683, E684 (cerebrospinal fluid isolates); and two penicillin-susceptible strains: INS-E682 (blood isolate) and E685 (cerebrospinal fluid isolate). A standard strain of PRSP (S. pneumoniae ATCC 49619) was used as a control for all experiments (Table 1).
Susceptibility testing
MIC and MBC to penicillin, ceftriaxone and vancomycin were determined by broth microdilution following CLSI procedures [15]. MIC to the same antibiotics and to chloramphenicol, trimethoprim-sulfamethoxazole and erythromycin had been determined at the Colombian INS using identical methodology.
Evaluation of baseline culture variables
Strain INS-E611, a penicillin-resistant strain (MIC = 2.0 μg/mL), was selected to test a wide group of variables involved in bacterial growth that could have importance for S. pneumoniae. Once the growth dynamics under each variable were determined for this strain, the most productive conditions were standardized and applied to the remaining strains. Bacterial stocks were stored at -70°C in two sets of 100 aliquots per strain. Cryoprotection media included 17% glycerol in trypticase soy broth for one set and skim milk for the other. The process from thawing to the end of variable evaluation was separated in three successive steps that we called phases: Phase 0 for resuscitating the frozen organism on to solid medium (two successive agar plates), Phase 1 for passing it to 10 ml liquid medium a first time (five 16 × 125 glass tubes labeled 1 to 5 with successive 1:10 dilutions), and Phase 2 for a second transfer of log-phased cells into 10 ml fresh liquid medium (three 16 × 125 glass tubes labeled 6 to 8 with successive 1:10 dilutions). The results were determined for all culture variables intervening at each phase and for the whole process by OD580 nm (standard broths) and OD600 nm (horse-blood supplemented broths) at 20–60 minutes intervals (SPECTRO 22, Labomed, Culver City, CA, USA); quantification of viable CFU per mL was made by dilution plating at least every hour during Phase 2.
These variables were tested by separate and combined at each step: cryoprotection media, addition of 0.5% yeast extract, and incubation time during Phase 0; inoculum size (number of CFU per tube), type of standard culture media, addition of 2% yeast extract and 2.5% horse blood, initial pH adjustment to 7.8 (744 pH Meter, Metrohm Ltd., Switzerland) with 1N NaOH (EK CHEM, Germany), incubation time, and number of the most diluted tube with visible turbidity (used to inoculate Tube 6) during Phase 1; and type of standard culture media, addition of 2% yeast extract and 2.5% horse blood, pH adjustment to 7.8 every hour, incubation time, and bacterial growth rate from Tubes 6, 7 and 8 during Phase 2. All three phases included incubation under 5% CO2 atmosphere. Three types of standard culture media were evaluated in Phases 1 and 2: Brain Heart Infusion (BHI-Ox, Oxoid LTD, Basingstoke, Hampshire, UK), BBL™ Brain Heart Infusion (BBL-BHI, Becton Dickinson & Co., Sparks, MD, USA) and Bacto™ Todd Hewitt Broth (THB, Becton Dickinson & Co., Sparks, MD, USA). Horse blood was obtained directly by the research group and defibrinated, lysed, centrifuged and filtered following standard procedures [16]. Sterile quality control agar or broth was set simultaneously with the inoculated media in all culture processes, and incubation temperature was set to 37 and 35°C for experiments with standard and supplemented media, respectively. Experiments were repeated 2–3 times.
Supplementation and optimization of culture conditions for PRSP
The evaluation of baseline cultures variables allowed the creation of a culture protocol that was tested with all strains using the same phases. The results for the whole process for each strain were standardized by hourly determination of OD600 nm and viable CFU/mL count during Phase 2.
Statistical analysis
Data are presented as means and standard deviations. The significance of differences in productivity of various culture media along time was determined by Repeated Measures ANOVA. For differences in maximal production between these media, one way ANOVA was applied followed by the Bonferroni t test. Data were stored, analyzed and graphed with Microsoft Excel v10.2 for Windows (Microsoft Corp., Seattle, WA, USA) and GraphPad Prism v4.0 for Windows (GraphPad Software, San Diego, CA, USA).
Authors' contributions
AVR carried out the microbiologic experiments, performed the analysis and interpretation of data and drafted the first versions of the manuscript. BES and MA carried out the microbiologic experiments and performed the analysis and interpretation of data. CAR carried out the microbiologic experiments, performed the analysis and interpretation of data and contributed in the study design. AFZ carried out the microbiologic experiments, performed the analysis and interpretation of data and drafted the first versions of the manuscript. OV conceived of the study, directed its design and coordination, obtained funding, directed data analysis, and re-wrote the final version of the manuscript. All authors contributed to the critical revision of the manuscript for important intellectual content and read and approved the final manuscript.
Role of the Sponsor
The funding organization had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript.
Acknowledgements
We are grateful to Dr. Elizabeth Castañeda from Instituto Nacional de Salud, Bogotá, Colombia, for her kind assistance during the execution phase of the study and for providing the bacterial strains. This study was funded by research grant Colciencias 1115-04-12981 and by University of Antioquia.
Figures and Tables
Figure 1 Comparison of three standard liquid media with a supplemented medium for culture of S. pneumoniae INS-E611. Five-hour growth curves for a PRSP strain in four different culture media: standard BHI (two brands), standard THB, and supplemented THB. All curves were obtained from Tube 7 (Phase 2). Bacterial growth was poor in standard BBL-BHI, and was not used for further experiments. As expected, there was no difference among these media when every data point along time was compared by repeated measures ANOVA (P = 0.7976). However, there was a significant difference in maximal growth between supplemented THB and standard media, as illustrated by Figure 2.
Figure 2 Maximal growth of S. pneumoniae INS-E611 cultured in three liquid media. The graph illustrates cell production of each medium after 5 hours of incubation; all cultures were obtained from Tube 7 (Phase 2). Supplemented THB was clearly superior to standard THB and BHI.
Figure 3 Growth of S. pneumoniae INS-E611 during Phase 2 in supplemented THB. One-mL bacterial suspension produced by Tube 5 (Phase 1) after 12 h of incubation was inoculated into 9-mL fresh broth (Tube 6, 1:10 dilution) and diluted further into Tube 7 (1:100) and Tube 8 (1:1000). Incubation under 5% CO2 atmosphere and pH adjustment every hour followed for 5–6 hours, with simultaneous plating for CFU counting. The graph illustrates the time to reach maximum growth for each inoculum size before triggering autolysis: 4, 5 and 7 hours for Tubes 6, 7 and 8, respectively. With this method, very similar maximum growths were attained under different inoculum sizes. Experiments with each tube were done 1–3 times.
Figure 4 Growth curves of six strains of PRSP under optimized culture conditions. The high productivity of viable log-phased bacteria demonstrated for S. pneumoniae INS-E611 was reproduced with other 5 strains of PRSP, including diverse serotypes (Table 1). More than 9.0 log10 CFU/mL were obtained in a consistent manner with all strains (curves shown correspond to Tube 7, Phase 2). The variability between strains, evidenced by the time to reach maximal growth, is characteristic of S. pneumoniae.
Table 1 Serotypes and minimal inhibitory and bactericidal concentrations (MIC / MBC) of six antimicrobial agents against nine strains of Streptococcus pneumoniae
Strains Serotype MIC / MBC [μg/mL] MIC* [μg/mL]
PEN CRO VAN ERY CHL SXT
INS-E611 6B 2.00 / 2.00 1.00 / 1.41 0.35 / 0.35 0.03 32.00 4.00
INS-E674 14 2.00 / 2.00 1.00 / 1.00 0.25 / 0.25 0.06 4.00 8.00
INS-E676 14 1.00 / 1.41 1.00 / 1.41 0.35 / 0.70 0.06 4.00 8.00
INS-E678 14 0.12 / 0.12 0.03 / 0.03 0.25 / 0.25 0.06 4.00 16.00
INS-E682† 6B 0.01 / 0.01 0.01 / 0.01 0.25 / 0.25 0.06 32.00 0.50
INS-E683 9V 1.00 / 1.00 0.35 / 0.35 0.35 / 0.35 0.06 4.00 32.00
INS-E684 14 1.00 / 1.00 0.50 / 0.50 0.25 / 0.25 0.13 4.00 1.00
INS-E685† 1 0.01 / 0.01 0.01 / 0.01 0.25 / 0.25 0.06 2.00 0.50
ATCC 49619 19F 0.46 / 0.46 0.06 / 0.12 0.33 / 0.44 0.06 4.00 0.25
PEN = penicillin; CRO = ceftriaxone; VAN = vancomycin; CHL = chloramphenicol; ERY = erythromycin; SXT = trimethoprim-sulfamethoxazole (only first component values shown)
* MIC determined by Instituto Nacional de Salud (INS), Bogotá, Colombia.
† Breakpoint for penicillin-susceptible strains: MIC ≤ 0.06 μg/mL.
Table 2 Comparison of diverse culture variables for productivity of PRSP cells
Variable Phase 0 Phase 1 Phase 2
Cryoprotection Media SM > 17%G:TSB NA NA
Inoculum (colonies) NA 10 > 5 or 15 or loopful NA
Culture Broth NA THB > BHI THB > BHI
Yeast Extract (% YE) 0.5 > 0 2.0 > 0 2.0 > 0
Horse Blood (% HB) NA 2.5 > 0 2.5 > 0
pH adjustment to 7.8 NA Initially > NoAdj q1h > NoAdj
Incubation time (h) 15 > 18 > 24 12 > 15, 18, 24 3*, 5†, 6‡
Ideal dilution (Tube #) NA 5 > 4 > 1–3 6 = 7 = 8
SM = Skim Milk; > implies "superior to"; NA = non applicable to that Phase; 17%G:TSB = 17% glycerol in trypticase soy broth; THB = Todd Hewitt Broth; BHI = Brain Heart Infusion; NoAdj= no adjustment; *Tube 6; †Tube 7; ‡Tube 8.
Table 3 Influence of inoculum size in cell production of Streptococcus pneumoniae INS-E611 after 12 hours of Phase 1 culture in supplemented THB
Number of CFU (Phase 0) OD600 nm (Phase 1) Broth pH (Phase 1) Phase 1 Production [log10 CFU/mL]
1 0.381 7.90 7.278
5 1.878 7.00 8.662
10 0.914 – 1.884 7.08 – 7.32 8.884 ± 0.047
15 1.766 6.28 9.067
Loopful 1.778 6.17 8.851
==== Refs
Austrian R The pneumococcus at the millennium: not down, not out J Infect Dis 1999 179 S338 S341 10081505
Gonçalves VM Zangirolami TC Giordano RLC Raw I Tanizak MM Giordano RC Optimization of medium and cultivation conditions for capsular polysaccharide production by Streptococcus pneumoniae serotype 23F Appl Microbiol Biotechnol 2002 59 713 717 12226729 10.1007/s00253-002-1075-8
Balachandran P Hollingshead SK Paton JC Briles DE The autolytic enzyme LytA of Streptococcus pneumoniae is not responsible for releasing pneumolysin J Bacteriol 2001 183 3108 3116 11325939 10.1128/JB.183.10.3108-3116.2001
Orihuela CJ Gao G McGee M Yu J Francis KP Tuomanen E Organ-specific models of Streptococcus pneumoniae disease Scand J Infect Dis 2003 35 647 652 14620149 10.1080/00365540310015854
Rieux V Carbon C Azoulay-Dupuis E Complex relationship between acquisition of beta-lactam resistance and loss of virulence in Streptococcus pneumoniae J Infect Dis 2001 184 66 72 11398111 10.1086/320992
Baltz RH Norris FH Matsushima P Dehoff BS Rockey P Porter G Burgett S Peery R Hoskins J Braverman L Jenkins I Solenberg P Young M McHenney MA Rosteck PR Skatrud PL Tomasz A DNA sequence sampling and gene disruption for identification of new antibacterial targets in Streptococcus pneumoniae Streptococcus pneumoniae: molecular biology and mechanisms of disease 2000 New York: Mary Ann Liebert 33 44
Rouff K Whiley RA Beighton D Murray P, Baron E, Pfaller M, Tenover F, Yolken R Streptococcus Manual of clinical microbiology 1999 7 Washington, DC: ASM Press 283 296
Musher D Mandell G, Bennett J, Dolin R Streptococcus pneumoniae Principles and practice of Infectious Diseases 2000 2 Fifth Philadelphia: Churchill Livingstone 2129 2145
Rotta J Sneath P, Mair N, Sharpe ME, Holt J Pyogenic haemolytic streptococci Bergey's Manual of Systematic Bacteriology 1986 2 Baltimore: Williams & Wilkins 1052
Lopez R Garcia E Garcia P Garcia JL Toumanen E, Mitchell T, Morrison D, Spratt B Cell wall hydrolases The pneumococcus 2004 Washington, DC: ASM Press 75 88
Morrison D Tomasz A Streptococcal competence for genetic transformation: regulation by peptide pheromones Streptococcus pneumoniae: molecular biology & mechanisms of disease 2000 New York: Mary Ann Liebert 45 53
Mazzola GJ Mortensen JE Millar LA Poupard JA The growth and survivability of Streptococcus pneumoniae clinical isolates subjected to various environmental conditions Diag Microb Infect Dis 2003 45 153 164 10.1016/S0732-8893(02)00526-6
Canvin JR Marvin AP Sivakumaran M Paton JC Boulnois GJ Andrew PW Mitchell PJ The role of pneumolysin and autolysin in the pathology of pneumonia and septicemia in mice infected with a type 2 pneumococcus J Infect Dis 1995 172 119 123 7797901
Lopez R Garcia E Garcia P Garcia JL Tomasz A The pneumococcal cell wall degrading enzymes: a modular design to create new lysins? Streptococcus pneumoniae: molecular biology & mechanisms of disease 2000 New York: Mary Ann Liebert 197 209
National Committee for Clinical Laboratory Standards Performance standards for antimicrobial susceptibility testing; fourteenth informational supplement M100-S14 2004 24 Wayne, PA
Hindler JF Hoechstein L Howell A Isenberg HD Preparation of routine media and reagents used in antimicrobial susceptibility testing Clinical microbiology and procedures handbook 2004 2 2 Washington, DC: ASM Press 5.14.3.9
| 15932633 | PMC1174873 | CC BY | 2021-01-04 16:03:39 | no | BMC Microbiol. 2005 Jun 2; 5:34 | utf-8 | BMC Microbiol | 2,005 | 10.1186/1471-2180-5-34 | oa_comm |
==== Front
BMC PediatrBMC Pediatrics1471-2431BioMed Central London 1471-2431-5-131591069410.1186/1471-2431-5-13Research ArticleOptimal fetal growth for the Caucasian singleton and assessment of appropriateness of fetal growth: an analysis of a total population perinatal database Blair Eve M [email protected] Yingxin [email protected] Klerk Nicholas H [email protected] David M [email protected] Centre for Child Health Research, The University of Western Australia, at the Telethon Institute for Child Health Research, P.O. Box 855, West Perth. WA. 6872, Australia2 Centre for Developmental Health, Curtin University of Technology and Telethon Institute for Child Health Research, P.O. Box 855, West Perth. WA. 6872, Australia2005 24 5 2005 5 13 13 20 7 2004 24 5 2005 Copyright © 2005 Blair et al; licensee BioMed Central Ltd.2005Blair et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The appropriateness of an individual's intra uterine growth is now considered an important determinant of both short and long term outcomes, yet currently used measures have several shortcomings. This study demonstrates a method of assessing appropriateness of intrauterine growth based on the estimation of each individual's optimal newborn dimensions from routinely available perinatal data. Appropriateness of growth can then be inferred from the ratio of the value of the observed dimension to that of the optimal dimension.
Methods
Fractional polynomial regression models including terms for non-pathological determinants of fetal size (gestational duration, fetal gender and maternal height, age and parity) were used to predict birth weight, birth length and head circumference from a population without any major risk factors for sub-optimal intra-uterine growth. This population was selected from a total population of all singleton, Caucasian births in Western Australia 1998–2002. Births were excluded if the pregnancy was exposed to factors known to influence fetal growth pathologically. The values predicted by these models were treated as the optimal values, given infant gender, gestational age, maternal height, parity, and age.
Results
The selected sample (N = 62,746) comprised 60.5% of the total Caucasian singleton birth cohort. Equations are presented that predict optimal birth weight, birth length and head circumference given gestational duration, fetal gender, maternal height, age and parity. The best fitting models explained 40.5% of variance for birth weight, 32.2% for birth length, and 25.2% for head circumference at birth.
Conclusion
Proportion of optimal birth weight (length or head circumference) provides a method of assessing appropriateness of intrauterine growth that is less dependent on the health of the reference population or the quality of their morphometric data than is percentile position on a birth weight distribution.
==== Body
Background
Being born small for one's gestational age is associated with adverse outcomes in both the short and long term [1-3]. However assessing whether a neonate is at risk of compromise on account of inappropriate intrauterine growth is complicated because not all fetuses should grow at the same rate [4]. Currently the appropriateness of fetal growth is usually inferred from the percentile position that the neonate's birth weight occupies on a gestation-specific birth weight distribution, that may also be specific for gender. This practice is unsuitable if the most appropriate birth weight for the neonate being assessed is not the same as that for all members of the population contributing to the distribution. Additional problems associated with percentile-based standards are (a) the implications of a given percentile position vary with the burden of growth restricting pathology in the source population, (b) the estimation of percentile position is imprecise at the extremes of a distribution where the information is of most clinical importance and (c) since percentile positions represent an ordinal rather than interval or ratio scale, the possibilities for valid statistical manipulation are limited [5].
This communication seeks to describe, demonstrate and justify the usefulness of an alternative method of assessing the appropriateness of fetal growth from information available in the neonatal record.
This method is based on three underlying concepts:
1. Appropriateness of growth can be expressed as the ratio of the observed birth dimension to the optimal birth dimension for that an individual neonate. Considering the dimension of weight we refer to this ratio as the proportion of optimal birth weight (POBW): a concept similar to the birth weight ratio [6]. Assessing appropriateness of growth then requires values for the optimal birth dimensions for the neonate in question.
2. Optimal intrauterine growth is most likely to be achieved during pregnancies unaffected by any maternal or fetal pathology or exposures that can pathologically affect fetal growth,
3. The many determinants of fetal growth can be classified as having either a pathological or a non-pathological effect on growth.
Factors with non-pathological effects on growth include gestational duration, gender [7-9], maternal size [10] and parity [7,11] and paternal size [12]. We define optimal birth weight as that achieved when no factors are present that can exert a pathological effect on growth. The central tendency of the distribution of birth weights in a population which experiences no factors that exert a pathological effect on intrauterine growth is taken as the optimal birth weight for neonates with the same combination of non-pathological determinants of fetal growth.
Pathological growth determining factors, such as maternal vascular disease or those associated with congenital malformations, usually restrict fetal growth. More rarely fetal weight is pathologically increased, the most well known example being fetal macrosomia induced by maternal diabetes [13].
It is less useful to categorise as either pathological or non-pathological those determinants of growth that cannot easily be altered. Multiple pregnancy and maternal race are two examples of such determinants.
This paper demonstrates our method of assessing the appropriateness of fetal growth by deriving equations for optimal birth weight, birth length and head circumference. Gestational duration, fetal gender and maternal height, age and parity are considered as potential independent variables representing the non pathological determinants of fetal growth. In order to select a population with optimal intrauterine growth, all births with evidence of having been exposed to pathological determinants of fetal growth are excluded. Appropriateness of fetal growth is then expressed as the ratio of the observed birth dimension to the estimated optimal birth dimension for a neonate with the same values for non-pathological determinants of fetal growth. The utility of this approach is discussed.
Methods
Sample selection
Records of the 126,393 births in Western Australia (WA) during the period of 1998 to 2002 were obtained from the Western Australian Maternal and Child Health Research Database (MCHRDB) [14]. This period was selected because data concerning whether the mother smoked during pregnancy, the most prevalent environmental exposure with a pathological effect on intrauterine growth, are available on this data base for births from 1998 onwards. The most recent available cohort at the time of writing was 2002.
Of 1998–2002 WA births, 85% were to Caucasian women and 96.8% were singletons. This example therefore derives standards for singletons born to Caucasian women, see Discussion for generalisability.
To achieve a cohort of Caucasian singletons anticipated to exhibit optimal fetal growth, any pregnancy with evidence to suggest that fetal growth may have been affected pathologically must be excluded. Stillbirths and deaths before 28 days were excluded as evidence of a suboptimal intrauterine course, which may be associated with abnormal growth, and, for stillbirths, because duration of intrauterine growth, as opposed to gestational duration to delivery, is not recorded.
The selection of further exclusion criteria was guided, in part, by the extensive literature concerning growth restriction as reviewed by Resnik [15]. Suggested exclusion criteria for which data are available in the MCHRDB are listed in Table 1 in order of their frequency observed in 1998–2002 WA births. Resnik [15] also suggests that maternal gestational use of anticonvulsants, cocaine, heroin or alcohol, maternal thrombophilic disorders and nutritional deprivation are risk factors for growth restriction. While these variables are not available on the MCHRDB, 0.5% of 1998–2002 WA mothers were recorded as having epilepsy and may have been on anticonvulsants. The use of cocaine and heroin are illegal in WA and very likely to be under-reported in medical records. Their use, along with that of excess alcohol, is associated with birth defects. Since both a birth defect and death before 28 days are exclusion criteria, it is anticipated that the majority of births significantly affected by these substances will be excluded. The incidence of thrombophilic disorders varies with ethnic background and no data are available concerning its frequency in WA pregnant women who are of mixed ethnic backgrounds. However neither thrombophilic disorders nor macro-nutrient deprivation are noted as problems in the WA pregnant population. Thus the factors listed in Table 1 are anticipated to represent the most frequently occurring pathological determinants of fetal growth in our population and were excluded from the sample for the purposes of deriving measures of optimal fetal growth. Socio-demographic variables of the selected sample were compared with those of excluded Caucasian singletons.
Table 1 Observed frequency of factors known to be associated with pathological deviations in fetal growth: All Western Australian births 1998–2002.
Factor N %
Maternal Smoking 27,326 21.62
Maternal vascular disease 8,334 6.59
Birth defects [29] 7,520 5.95
Maternal (pre-existing or gestational) diabetes 5,051 4.00
Multiple pregnancy 3,991 3.16
TORCH infectionsa 2,945 2.33
High altitudeb 0 0.0
a) Toxoplasmosis, Rubella, CMV, Herpes
b) No population centre in Western Australia is above 300 m. The highest peak (Mt. Meharry) is 1253 m, and like all Western Australian peaks, is situated in an unpopulated area.
Gestational age data
Since the primary determinant of birth dimensions is the duration of growth, reliable estimates of gestational duration (GA) are essential, yet exclusion criteria for poor quality gestational data are likely to exclude a biased sample. Details of the algorithm used to obtain the best estimate of GA from all available data are described and justified elsewhere [16]. Applying this method to the total 1998–2002 WA birth population resulted in no satisfactory gestational estimate being available for only 97 births (~0.1%) and being beyond the range of 23–42 weeks for a further 573 subjects. Birth weight and gestational duration data for remaining births were examined to exclude combinations so unlikely as to suggest error in the gestational datum. The cut off birth weights at each gestational age between 23 and 36 weeks, above which the observation was excluded, were selected with a view to excluding infants at least four gestational weeks older than reported, since break through bleeding at four-week intervals in early pregnancy is a source of gestational error in women claiming to be certain of the date of their last menstrual period. Due to the slower rate of weight accretion with respect to weight dispersion in infants born at term, this method of data cleaning is not applicable for births reported as being at greater than 36 weeks gestational duration [17].
Analysis
For each of the three response variables (birth weight, length and head circumference) the Box-Cox transformation [18] was used to identify the optimal transformation to reduce non-normality and heteroscedasticity of errors. Fractional polynomial regression was then used to identify the best fit transformation of gestational age to account for any non-linearity in the relationship between gestational duration and each response variable [19]. Fractional polynomials are a means of identifying the curve of best fit in cases where non-linearity is possible but there is no scientific reason to specify the shape of the non-linear relationship. Royston and Altman claim that their set of power transformations have the flexibility to cover almost all likely shapes of non-linear relationship. The number of possible inflection points is determined by the order of the fractional polynomials fitted. In this case, where a sideways "S" shape is expected, 2nd order fractional polynomials, which allow for up to two inflection points, are sufficient. To aid computation, gestational duration was included in the fractional polynomial regressions as GA/100.
Maternal height (cm) and maternal age (years) were included as linear predictor variables (see Discussion) centred on the population mean values of 162 cm and 25 years respectively. Infant sex and maternal parity were included as categorical variables. Parity was categorised with first birth as the reference, second and third birth as two separate categories, and fourth and subsequent births constituting the fourth category. Models were fitted using SAS (Version 8.2) (SAS Institute Inc., 2001). The fit of each model was tested by plotting residuals against GA and against the predicted dependent variable (weight, length or head circumference at birth). Additionally, POBWs for 3rd, 10th and 90th percentile birth weight were estimated within each completed gestational week for sub-samples of parity and maternal height, to ensure that the model adjusted appropriately for these non-pathological determinants. Finally, to aid clinical interpretation of POBW values, POBW was estimated for the 3rd, 10th and 90th percentile positions by taking the weighted mean of POBWs estimated for each parity/gender stratum of births between 38 and 41 gestational weeks inclusive, assuming a constant maternal height of 162 cm.
Results
Table 2 lists the numbers of births sequentially excluded by each exclusion criterion, and shows that 62,746 singleton Caucasian births remained for analysis. Equations for optimal growth were derived from this total population sample of singleton, Caucasian births without recognised risk factors for growth abnormality, which comprised 49.7% of all Western Australian births and 60.5% of the Caucasian singleton births. Table 3 compares the distributions of socio-demographic variables for included births with those of excluded Caucasian singleton births. As anticipated, given the sample size and the selection criteria, the difference in all distributions is statistically very significant, with the exception of gender, which is nonetheless significantly different at the p = .05 level. However the clinical differences tend to be small with the exception of the proportion of preterm and very preterm births.
Table 2 Sample selection: the number of births sequentially excluded by each exclusion criterion.
Exclusion Criterion Number (%) excluded Number remaining
All WA births 1998–2002 0 126,393
Not Caucasian 18,968 (15.0) 107,425
Multiple pregnancy 3,532 (3.3) 103,893
Stillbirth 606 (0.6) 103,287
Maternal gestational smoking 21,570 (20.9) 81,717
Growth restricting conditionsa 9,970 (12.2) 71,747
Gestation <23 or >42 weeks 146 (0.2) 71,601
Birth weight excessive for GA 75 (0.1) 71,526
Missing essential variableb 5,254 (7.3) 66,272
Death before 28 days 102 (0.2) 66,170
Birth defectc 3,424 (5.2) 62,746
a. as identified in Table 1.
b. Essential variables were maternal height and age, birth weight, length and head circumference: almost all exclusions at this stage were the result of missing values for maternal height.
c. [28]
Table 3 Comparison of distributions of selected characteristics among Caucasian singleton births which were or were not included in the study.
Included Excluded*
Denominator, N 62,747 maximum 41,146
Sex p
Male fetus, % (N) 50.62 (31,760) 51.63 (21,242) 0.0014
Gestational age
Mean GA (sd), wks 39.0 (1.6) 38.2 (2.6) <.0001
GA<37 weeks, % (N) 4.55 (2,856) 9.96 (4,099) <.0001
GA<33 weeks, % (N) 0.53 (334) 2.65 (1,092) <.0001
5th-95th percentile, wks 37–41 35–41
Birth weight
Mean weight (sd), g. 3,282 (641) <.0001
5th-95th percentile, g 2695–4270 2265–4170
Birth length
Mean length (sd), cm 50.4 (2.5) 49.5 (3.7) <.0001
5th-95th percentile, cm 46–54 45–54
Birth head circumference
Mean circumference, cm 34.7 (1.6) 34.2 (2.4) <.0001
5th-95th percentile, cm 32–37 31–37
Maternal characteristics
Mean height, cm 165.1 (6.7) 164.7 (6.7) <.0001
5th-95th percentile, cm 154–176 153–176
Mean age, y 29.4 (5.2) 28.5 (5.7) <.0001
5th-95th percentile, y 20–38 19–38
Socio-economic Disadvantage† 1009.3 (78.8) 984.7 (86.0) <.0001
* Variable denominators as a result of missing values.
† [29]
Birth weight
The Box-Cox procedure suggested that square root was the optimal transformation to use for normalising birth weight. The optimal fractional polynomial gestational age terms were GA3 and GA3ln(GA). In multivariate analysis, maternal age was not a significant predictor of birth weight. Parameter estimates for the selected best fitting regression equations for the square root of birth weight are given in Table 4. This model has an adjusted R2 of 40.5%. The best fitting regression equation for estimating optimal birth weight (grams) can therefore be expressed as:
Table 4 Parameter estimates modelling the square root of birth weight (grams)
Independent Variable Parameter Estimate Standard Error t value Pr >|t| 95% confidence limits
Intercept -14.08 0.73 -19.36 <.0001 -15.51 -12.66
GA3 * -1413.6 25.0 -56.43 <.0001 -1463 -1365
GA3 In (GA3) * -2782.5 39.4 -70.55 <.0001 -2860 -2705
Male infant 1.185 0.027 44.22 <.0001 1.13 1.24
Maternal height cm# 0.1077 0.0024 45.63 <.0001 0.103 0.112
2nd birth 1.0277 0.031 33.40 <.0001 0.967 1.088
3rd birth 1.318 0.0399 33.04 <.0001 1.24 1.40
4th and subsequent birth 1.571 0.054 29.33 <.0001 1.46 1.68
GA# × Maternal height 0.00667 0.00123 5.42 <.0001 0.0043 0.0091
Adjusted R2 for this model is 0.405
* indicates non-centred but scaled GA, divided by 100
# indicates a centred variable – GA centred around mean of 40 weeks, maternal height centres around a mean of 162 cm
and, of course,
This equation suggests that under our standard conditions of birth at 40 weeks gestation to a 162 cm primiparous woman, female infants should weigh 3436.0 g and males 3576.4 g. Second births should weigh 123 g more, third births 158 g more and fourth or subsequent births 189 g more than the first birth. An example of the curves obtained from this equation is shown in Figure 1. The weighted mean POBWs (across the 8 parity/gender combinations) observed at the 3rd, 10th and 90th percentile positions on the birth weight distribution are shown by gestational duration across the range 35–42 weeks in Figure 2. These ratios change little by gestational duration within this range. The weighted mean POBWs across the range 38–41 weeks for the 3rd, 10th and 90th percentile birth weights are 81%, 87% and 115% respectively, Table 7.
Figure 1 Mean of male and female optimal birth weight by gestational age at delivery and parity, estimated for births to women of height 162 cm.
Figure 2 Weighted mean POBW (across 8 parity/gender combinations) observed at the 3rd, 10th and 90th percentile positions on the birth weight distributions, by gestational age at delivery.
Birth length
The Box-Cox procedure suggested that birth length raised to the power of 0.75 was the optimal transformation to use for normalising birth length. The optimal fractional polynomial gestational age terms were GA2 and GA3. In multivariate analysis, maternal age was not a significant predictor of birth length. Parameter estimates for the selected best fitting regression equations for the square root of birth weight are given in Table 5. This model has an adjusted R2 of 32.2%. The best fitting regression equation for estimating optimal birth weight (grams) can therefore be expressed as:
Table 5 Parameter estimates modelling birth crown heel length to the power of 0.75 (cm)
Independent Variable Parameter Estimate Standard Error t value Pr >|t| 95% confidence limits
Intercept 5.684 0.176 32.30 <.0001 5.34 6.03
Scaled gestational age squared* 209.9 3.66 57.31 <.0001 203 217
Scaled gestational age cubed* -318.5 6.44 -49.43 <.0001 -331 -305
Male infant 0.2350 0.0047 50.08 <.0001 0.226 0.244
Maternal height c# 0.01665 0.00042 40.28 <.0001 0.0158 0.0174
2nd birth 0.07484 0.0054 13.89 <.0001 0.064 0.085
3rd birth 0.1161 0.0070 16.62 <.0001 0.102 0.130
4th and subsequent birth 0.1508 0.0094 16.09 <.0001 0.132 0.169
GA# × Maternal height 0.000763 0.000215 3.54 0.0004 0.00034 0.00120
Adjusted R2 for this model is 0.322
* indicates non-centred but scaled GA, divided by 100
# indicates a centred variable – GA centred around mean of 40 weeks, maternal height centres around a mean of 162 cm
and of course
This equation suggests that under our standard conditions females should be 50.3 cm long at birth and males should be 0.83 cm longer. The weighted mean proportions of optimal crown heel length at 3rd, 10th and 90th percentile positions of crown heel length were found to be 93%, 95% and 105% respectively, see Table 7.
Birth head circumference
The Box-Cox procedure suggested that it was neither necessary nor desirable to transform head circumference prior to modelling. The optimal fractional polynomial gestational age terms were GA and GAln(GA). All potential predictor variables significantly predicted head circumference, including maternal age. Parameter estimates for the selected best fitting regression equations for the square root of birth weight are given in Table 6. This model has an adjusted R2 of 25.2%. The best fitting regression equation for estimating optimal birth weight (grams) can therefore be expressed as:
Table 6 Parameter estimates modelling head circumference at birth (cm)
Independent Variable Parameter Estimate Standard Error t value Pr >|t| 95% confidence limits
Intercept -88.31 1.92 -46.31 <.0001 -92.04 -84.57
Scaled gestational age* 43.31 0.38 115.06 <.0001 42.57 44.05
GA ln (GA) * -287.6 5.1 -56.32 <.0001 -297.6 -277.6
Male infant 0.6072 0.0109 55.89 <.0001 0.5860 0.6285
Maternal height cm # 0.02745 0.000956 28.68 <.0001 0.0256 0.0293
2nd birth 0.2352 0.0127 18.52 <.0001 0.210 0.260
3rd birth 0.3151 0.0167 18.92 <.0001 0.282 0.348
4th and subsequent birth 0.3394 0.0225 15.09 <.0001 0.295 0.384
Maternal age # 0.01322 0.00110 12.01 <.0001 0.0111 0.0154
GA# × Maternal height 0.00107 0.000498 2.14 0.0321 0.000091 0.00205
Adjusted R2 for this model is 0.252
* indicates non-centred but scaled GA, divided by 100
# indicates a centred variable – GA centred around mean of 40 weeks, maternal height is centred around a mean of 162 cm, maternal age is centred around a mean of 25 years
Table 7 Percentage of optimal birth dimension equivalences of percentile cut points from which appropriateness of growth has traditionally been inferred: as observed in this sample of optimally grown neonates.
Percentage of optimal birth dimension (%)
Percentile Position Weight Length Head circumference
3rd 80 93 93
10th 87 95 96
90th 115 105 105
This equation suggest that under our standard conditions and for mothers of 25 years, the optimal head circumference for females was 34.4 cm and for males, 0.61 cm larger. The weighted mean proportions of optimal head circumference at 3rd, 10th and 90th percentile positions of head circumference were found to be 93%, 96% and 105% respectively, see Table 7.
Discussion
These regression equations, derived from a population based sample of more than 62,000 singleton Caucasian pregnancies without the major risk factors for intrauterine growth anomaly, demonstrate the method used at our Institute to assess appropriateness of fetal growth. The method is applicable to all populations with suitable data available. Our results may be directly applicable to other populations besides Western Australian Caucasian and Aboriginal singletons, and we consider it likely to be applicable to all Caucasian populations, but applicability should be verified as suggested below.
Advantages of the ratio method
The use of ratios, such as POBW, in the measurement of intrauterine growth is not a novel idea, [6,20] but has not been universally adopted despite the many advantages of ratios over the more commonly used percentile positions. The following discussion applies to all ratios of optimal dimensions, but, for simplicity POBW will be used as the example throughout. These advantages are:
a) Ratios, such as POBW, represent continuous interval measures.
b) Estimations of POBW require only a single standard value, the predicted optimal birth weight, rather than values at several points on the birth weight distribution. The precision of estimating a percentile position varies inversely with observation density and, since the majority of distributions have fewer observations at the extremes, extreme observations will be the least precise, whatever the size of the sample generating the distribution. Extreme observations are also most subject to error. When verification of individual observations is not possible (as with de-identified data), it is a common practice to exclude extreme values on the assumption that they are in error, significantly altering the estimated value of extreme percentile positions. The positions of percentile extremes are therefore both imprecise and sensitive to actual and perceived data quality. The most precise percentile estimates are those at the highest observation densities, which, since many distributions are akin to Gaussian (particularly those of birth dimensions), is often the 50th percentile or median[5].
c) Births affected by growth disturbing factors are over-represented in the extremes of the growth distribution. Hence the positions of extreme percentiles are sensitive to the incidence of growth affecting pathologies in the reference population and vary with the health of the reference population. For example, a newborn with a POBW of 85% might be at the 20th percentile position of the birth weight distribution for a population with a high burden of growth restricting pathologies, but the 8th percentile of a population with optimal fetal growth. For an extreme percentile position to be meaningful therefore, the health status of the population from which it is derived needs to be defined, whereas the predicted birth weight is less sensitive to disease burden. Though less sensitive, the proportion of the reference population with growth disturbing factors will also affect the predicted birth weight, except in the unlikely situation where pathological restriction is balanced by pathological acceleration. For this reason we sought to identify a population without growth disturbing factors. The ratio of observed birth weight to predicted birth weight is more generalisable than extreme percentile positions, and the ratio of observed to predicted optimal birth weight is even more generalisable[5].
d) POBW is a continuous scale that correlates with weight deficit, whereas percentile position is an ordinal scale that does not. For example, Table 8 considers a population sample with a normal (Gaussian) birth weight distribution, mean birth weight of 3,400 g and standard deviation of 345 g. Being Gaussian, the predicted weight equals the mean (and 50th percentile position) or 3,400 g. In Table 8, changes in percentile position of 4 or 5 percentile points are shown to represent changes in weight of between 43 g and 151 g depending on the particular percentile positions, whereas, within a population, there is a linear correlation between differences in weight and change in POBW. Equivalent changes in percentile position do not represent equivalent changes in weight. Furthermore, in a total population the presence of growth restricting factors creates a negatively skewed (non-Gaussian) birth weight distribution, so the observed range of birth weights covered by extreme percentiles is broader than indicated in Table 8 and is unpredictable.
Table 8 Comparison of changes in percentile position and in POBW for selected changes in birth weight, for a neonate with an estimated optimal birth weight of 3,400 g.
Percentile positions (percentile points) Change in birth weight (grams) Change in POBW (%)
Range Change Range Change Range Change
50th-45th 5 3400-3357 43 100-98.74 1.3
10th-5th 5 2958-2832 126 87.0-83.3 3.7
5th-1st 4 2832-2681 151 83.3-78.9 4.4
Failure to utilise the advantages of a ratio may in part be due to clinical unfamiliarity. In contrast to percentile position, there is little literature describing the clinical associations of appropriateness of growth expressed as proportions of a desirable birth dimension [21]. For this reason we have included Table 7 which gives the estimated mean, over gestational weeks 38 – 41 inclusive and each gender and parity group, proportion of optimal ratio values of the 3rd, 10th and 90th percentile positions, of each distribution of weight, length and head circumference at birth. This table of equivalences enables an approximate translation of the literature using percentile positions to percentages of optimal dimensions. The populations from which percentiles are derived will seldom be confined to those without factors affecting growth pathologically. The proportion of optimal equivalences in Table 7 will over-estimate the appropriateness of growth of percentile defined groups to an extent depending on the burden of growth restricting pathologies in their reference sample. When POBW becomes familiar, its numerical values will convey more precise and generalisable clinical meaning than the traditional percentile positions.
Sample selection
The sample was limited to singleton births to Caucasian mothers. It is not useful to classify all determinants of intrauterine growth according to whether or not they have a pathological effect on growth. For example, twin pregnancy slows fetal growth particularly in the third trimester; and gestation-specific perinatal outcomes for multiple births delivered at term are not as good as those for singletons [22-24]. However, it is seldom desirable to reduce twin pregnancies to singleton pregnancies and reasonable to ask whether a twin fetus is growing appropriately, given that it is a twin. Multiplicity-specific fetal growth standards would be required to answer this question.
Maternal race is also a problematic factor. The observed variation in intra-uterine growth rates between ethnic groups [7,6] may reflect genetically determined differences in optimal rates and/or systematic differences in incidence of growth restricting pathologies and/or environmental exposures. That is, the association between growth and maternal race may arise as a result of either or both non-pathological and pathological determinants of fetal growth. If racial variation in intrauterine growth arises purely as a result of pathological determinants, maternal race is merely associated with growth rate, rather than being a determinant, and should not be controlled. The balance between non-pathological and pathological influences is likely to vary between ethnic groups and between locations. For example, in Western Australian (WA) Indigenous communities the tendency to slower fetal growth relative to Caucasians is believed to be primarily a result of a higher incidence of growth restricting pathologies and environmental exposures [11]. In south east Asian communities living in WA women also tend to have small babies but their perinatal outcomes are similar to those of WA Caucasians. It is reasonable to differentiate birth weight distributions by race only if race is itself a (non-pathological) determinant of fetal growth. Whether this is the case may be determined by comparing the estimations for optimal fetal growth, adjusted for non-pathological determinants, between populations of different races, after excluding pregnancies with evidence of exposure to pathological growth determining factors. If the estimations are significantly and systematically different, race specific standards are required. If they are not, the same standards for optimal fetal growth may be used even if the observed distributions in birth weight differ.
The aim of many of our exclusion criteria was to select a sample of births that had not been exposed to factors that have a pathological effect on intrauterine growth. Although the selection criteria only consider causes of growth anomaly we anticipated that selected births would be more likely to be born at term, be larger and born to taller, older and less disadvantaged mothers for several reasons. For example, pathologically affected growth is more often restricted than accelerated and is also associated with preterm birth and maternal smoking, the most prevalent pathological determinant of intrauterine growth, is associated with maternal age and socio-economic circumstances. Table 2 shows these anticipations to be realised. Selected births were also somewhat more likely to be female, supporting the general observation of the female advantage during gestation.
We believe that the curves shown in Figure 2 demonstrate that the exclusion of pathologically affected intrauterine growth, and of erroneous gestational estimates, has been reasonably successful. Biological growth (of which fetal growth is an example) typically proceeds to produce a Gaussian distribution at any point in time, with the standard deviation being proportional to the mean. Charts of unselected birth weights against gestational duration usually demonstrate increasing dispersion with decreasing gestation of delivery for deliveries before 40 weeks. There are two reasons for this: (i) the proportion of erroneously reported gestational age values increases with decreasing gestation, simply because the number of births actually delivered at any gestation week decreases the further it is from the modal value. Typically, erroneous preterm gestational reporting underestimates true gestational duration, hence the birth weight associated with an erroneously reported preterm gestational age is typically higher than those of births actually delivered at the reported gestation.
(ii) Birth much before term often has a pathological cause that also influences fetal growth. Thus the distributions of birth weights delivered at preterm gestations are no longer Gaussian, but typically, because pathological restriction occurs much more frequently than acceleration, are negatively skewed. Thus weights of neonates born at preterm gestations tend to be lower than fetuses of the same gestational age who go on to deliver at term.
In this study we address (i) by using the best estimate of gestational age that can be derived from all available data[16] and to exclude the '4-week errors' arising from gestational break-through bleeding. If we have succeeded in excluding pregnancies exposed to factors known to pathologically affect fetal growth we have addressed (ii), though the cost is the exclusion of a disproportionate number of infants born preterm, particularly very preterm, as can be seen in Table 2. The observation that the POBWs of the percentile positions are independent of gestational age, Figure 2, indicates that the dispersion is proportional to the mean across gestational age, and is compatible with both (i) and (ii) having been successfully addressed.
Selection of independent variables
Some may consider our selection of predictor variables incomplete as it does not include measures of paternal size, maternal weight or maternal weight gain.
Paternal size
While paternal size is known to influence fetal growth it was not included because the biological father cannot routinely be identified and therefore measures of paternal size are not available on our database. The proportion of variability accounted for by the regression equations would, no doubt, be increased by the inclusion of paternal height as an independent variable.
Maternal weight
It has been suggested that maternal size affects fetal growth because it correlates with the area of uterine endometrium available for placentation. Since this area is not directly measurable, it is logical to seek and adjust for the maternal dimension(s) that correlates most closely with it. Maternal height measures skeletal size in the vertical dimension only, while maternal weight is associated with skeletal size and soft tissue mass, including adipose tissue. Skeletal height tends to correlate with skeletal size, but the proportion of weight consisting of soft tissue, particularly adipose tissue, is very variable, weakening the correlation between maternal weight and skeletal size. Data from the 5 month Dutch famine suggest that maternal pre-natal weight for height, a measure of soft tissue mass, is not a strong determinant of birth weight [25]. We therefore suggest that maternal height is likely to correlate better with the uterine area available for placentation than is prenatal maternal weight or weight for height.
Maternal weight gain
Maternal weight gain is occasionally considered to be a determinant of birth weight [26]. However fetal weight can be expected to correlate with maternal weight gain, because fetal weight, and its correlate placental weight, are significant components of maternal weight gain. Thus rather than being a non-pathological determinant of fetal weight, maternal weight gain partially measures fetal growth, whether or not it is optimal and should therefore not be adjusted for when estimating appropriateness of growth.
The non-pathological determinants of growth used in these models accounted for 40.5%, 32.2% and 25.2% of the variance in birth weight, length and head circumference respectively. The variation between these proportions may result from variation in the accuracy with which each birth dimension can be measured. Birth weight is routinely measured to within 5 g, representing about 0.15% of a median weight baby. Compared with birth weight birth, length is more difficult to measure reliably due to the tendency of the neonate to flex and the facility with which it may be stretched. Measured head circumference at delivery may be influenced by moulding of the head during passage through the birth canal. The effect of moulding on head circumference may be largely avoided by waiting until 2 days after birth before measurement. However with early discharge policies, such a wait risks failing to obtain any measurement of head circumference and in WA head circumference is routinely measured in the delivery room.
The highly significant, though small, dependence of head circumference on maternal age, despite adjustment for parity, was unexpected and requires confirmation in independent samples
Inclusion of maternal height and age as linear variables
All three dependent variables were found to have a linear dependence on maternal height in the range 147–183 cm. Outside this range there was a tendency for regression to the mean value of the dimension. This may occur because we could not include a term for paternal height and there will be a tendency for women at the extremes of height to have partners with heights that are less extreme. However since only 269 (~0.4%) of our selected sample had a height outside this range, maternal height was included as a linear variable.
Of the three dependent variables only head circumference had an association with maternal age, which was found to be linear up to age 45 years. Only 16 (0.03%) of our selected sample were older than 45 at the time of delivery, therefore maternal age was also included as a linear variable.
Comparisons with previous methods of assessing intrauterine growth
In 1963 Lubchenco and colleagues [27] presented the first percentile charts of gender- and gestation-specific birth weights for an unselected population of live births and thereby initiated the modern study of intrauterine growth. The next major innovation in the methods of assessing intrauterine growth was the development of customised computer generated charts for individual neonates [4] by adding to gender and gestational duration the following predictor variables: maternal height, weight, ethnic group, parity and the birth weight, gestational duration and gender of any previous siblings, with the option of further adding measures of growth taken during the index pregnancy. The charts were again presented as percentiles. These charts were designed to predict birth weight rather than assess appropriateness of growth, as not all the independent variables (eg. sibling growth and ethnic group) are necessarily non-pathological determinants of intrauterine growth.
In 1993, Wilcox and colleagues [6] introduced the concept of the birth weight ratio, the ratio of the observed birth weight to the birth weight predicted given gestational duration, fetal gender, maternal height, weight, parity and ethnic group. Their study sample excluded multiple births, stillbirths and congenitally abnormal babies, and limited their analysis to term births, but did not attempt to exclude pregnancies affected by other pathological determinants of growth. Poor growth was defined on the basis of a percentile position of the birth weight ratio, thereby retaining the problems inherent in the use of percentile positions as standards.
Summary
The method reported in this paper introduces two innovations, (i) using optimal, rather than expected, growth as the standard, and (ii) reporting the ratio of the observed to optimal birth dimension as the indicator of appropriateness of growth, rather than a percentile position.
We sought a sample with optimal opportunities for fetal growth for the creation of standard both because this is the logical standard and also to avoid the problem of the varying incidence of growth restricting pathology and environmental exposures between populations. Our previously published birth weight standards for Western Australia [17] excluded only perinatal deaths from the reference sample because other relevant data were not available at the population level. In subsequent models, we explored the possibility of using ratios rather than percentiles [11,20], the effects of maternal height and parity were estimated in broad strata and those of maternal age were not considered. Although births affected by some factors suggesting supoptimal growth could be excluded, data concerning the most commonly occurring pathological growth restricting exposure, maternal smoking, were not then available at the population level.
The creation of the standards for optimal fetal growth for Caucasian singletons presented here is possible in part due to additional methods of estimating gestational duration [16], the ability to exclude the large proportion of births to women who smoked or experienced factors known to affect fetal growth and more complete information concerning non-pathological determinants of growth. Computing and statistical methods have also been improved with the use of (a) the Box-Cox transformation to account for any non-normality in the distribution of the response variables (b) fractional polynomial regression which required no assumptions regarding the form of the relationship between gestational duration and each of the response variables and facilitated the use of continuous variables thereby allowing the effects of non-pathological determinants of growth to be estimated more precisely.
Conclusion
We have presented a comprehensive guide to an alternative method of creating standards for newborn dimensions and assessing appropriateness of intrauterine growth. It is based on the estimation of the optimal value for the dimension which we define as the value obtained by regression techniques from a large sample of women without risk factors for intrauterine growth anomaly. In this method, appropriateness of intrauterine growth is expressed as the ratio of the observed birth dimension to the optimal birth dimension rather than as being above or below a specified percentile position of the population distribution of that dimension, avoiding the problems inherent in the use of percentile position. Since POBW is a measure of appropriateness of intrauterine growth it may be used as a continuous variable and subjected to parametric statistical analysis. The use of POBW in clinical and research settings will prove whether it is a more precise predictor of compromise within individuals than previously available indicators of intrauterine growth status.
Abbreviations
GA: best available estimate of gestational age at delivery.
POBW: percentage of optimal birth weight.
WA: Western Australia.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
EB conceived of and directed the study and drafted the manuscript. YL carried out initial statistical analyses. NHdeK gave statistical advice. DML carried out subsequent statistical analyses.
All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
The authors are grateful to Vivien Gee and Western Australian midwives for collecting and providing the birth data, to Peter Cosgrove abstracting the required data and to R.S. Kirby for a very helpful review. The work reported in this paper has been supported financially by Program Grants #003209 and #353541 from the National Health and Medical Research Council of Australia.
==== Refs
Marlow N Kingdom J BP Paediatric implications - neonatal complications Intrauterine growth restriction: aetiology and management 2000 London, Springer 337 349
Blair E Kingdom J BP Paediatric implications of IUGR with special reference to cerebral palsy Intrauterine growth restriction: aetiology and management 2000 London, Springer 351 366
Reynolds RM Godfrey KM Kingdom J BP Long term implications for adults health Intrauterine growth restriction: aetiology and management 2000 London, Springer 367 384
Gardosi J Chang A Kalyan B Sahota D Symonds EM Customised antenatal growth charts Lancet 1992 339 283 287 1346292 10.1016/0140-6736(92)91342-6
Blair E Uses and Misuses of the Percentile (Centile or Quantile) Position. Australasian Epidemiologist 2003 10 26 28
Wilcox MA Maynard PV Chilvers CED The individualised birthweight ratio: a more logical outcome measure of pregnancy than birthweight alone. British journal of obstetrics and gynaecology 1993 100 342 347 8494835
Zhang J Bowes WA Birth-weight-for -gestational age patterns by race, sex, and parity in the United States population Obstetrics and Gynecology 1995 1995 200 208 7617350 10.1016/0029-7844(95)00142-E
Roberts CL Lancaster PAL Australian national birthweight percentiles by gestational age Medical Journal of Australia 1999 170 114 118 10065122
Skjerven R Gjessing HK Bakketeig LS Birthweight by gestational age in Norway Acta Obstetricia Gynecologica Scandinavica 2000 79 440 449 10.1034/j.1600-0412.2000.079006440.x
Voorhorst FJ Bouter LM Besemer PD Kurver PHJ Maternal characteristics and expected birth weight European Journal of Obstetrics & Gynecology and Reproductive Biology 1993 50 115 122 8405639 10.1016/0028-2243(93)90175-C
Blair E Why do Aboriginal newborns weigh less? Determinants of birthweight for gestation. J Paediatrics and Child Health 1996 32 498 503
Wilcox MA Newton CS Johnson IR Paternal influences on birthweight Acta Obstetricia et Gynecologica Scandinavica 1995 74 15 18 7856427
Richard J Identification of fetal growth abnormalities in diabetes mellitus Seminars in Perinatology 2002 26 190 195 12099308
Stanley FJ Croft ML Gibbins J Read AW A population database for maternal and child health research in Western Australia using record linkage Paediatric and Perinatal Epidemiology 1994 8 433 447 7870627
Resnik R Intrauterine growth restriction Obstet Gynecol 2002 99 490 496 11864679 10.1016/S0029-7844(01)01780-X
Blair E Liu Y Cosgrove P Choosing the best estimate of gestational age from routinely collected population based perinatal data. Paediatr Perinat Epidemiol 2004 18 270 276 15255880
Blair E Stanley FJ Intrauterine growth chart Commonwealth Department of Health 1985
Box GEP Cox DR An analysis of transformations Journal of the Royal Statistics Society 1964 B-26 211 252
Royston P Altman DG Regression using fractional polynomials for continuous covariates: parsimonious parametric modelling (with discussion) Applied Statistics 1994 43 429 467
Palmer L Petterson B Blair E Burton P Family patterns of gestational age at delivery and growth in utero in moderate and severe cerebral palsy. Developmental medicine and child neurology 1994 36 1108 1119 7958523
Palmer L Blair E Petterson B Burton P Antenatal antecedents of moderate and severe cerebral palsy Paediatric and perinatal epidemiology 1995 9 171 184 7596894
Taylor GM Owen P Mires GJ Foetal growth velocities in twin pregnancies Twin Research 1998 1 9 14 10051352 10.1375/136905298320566438
Liu Y Blair E Predicted birthweight for singleton and twins Twin Research 2002 5 529 537 12573184 10.1375/136905202762341991
Alexander GR Kogan M Martin J Papiernik E What are the fetal growth patterns of singletons, twins, and triplets in the United States? Clinical Obstetrics and Gynecology 1998 41 115 125 10.1097/00003081-199803000-00017
Morley R Owens J Blair E Dwyer T Is birthweight a good marker for gestational exposures that increase the risk of adult disease? Paediatric and Perinatal Epidemiology 2002 16 194 199 12123431 10.1046/j.1365-3016.2002.00428.x
Strauss RS Dietz WH Low maternal weight gain in the second or third trimester increases the risk for intrauterine growth retardation. Journal of Nutrition 1999 129 988 993 10222390
Lubchenco LO Hansman C Dressler M Boyd E Intrauterine growth as estimated from liveborn birth- weight data at 24 to 42 weeks of gestation Pediatrics 1963 32 793 800 14075621
Bower C Rudy E Ryan A Cosgrove P Report of the birth defects registry of Western Australia, 1980-2003 2004 Perth, Western Australia, King Edward Memorial Hospital, Women's and Children's Health Service.
McLennan W 1996 Census of population and housing: socio-economic indexes for areas. 1998 Canberra, Australia, Australian Bureau of Statistics
| 15910694 | PMC1174874 | CC BY | 2021-01-04 16:31:06 | no | BMC Pediatr. 2005 May 24; 5:13 | utf-8 | BMC Pediatr | 2,005 | 10.1186/1471-2431-5-13 | oa_comm |
==== Front
BMC Public HealthBMC Public Health1471-2458BioMed Central London 1471-2458-5-531592151210.1186/1471-2458-5-53Research ArticleAccumulation of health risk behaviours is associated with lower socioeconomic status and women's urban residence: a multilevel analysis in Japan Fukuda Yoshiharu [email protected] Keiko [email protected] Takehito [email protected] Health Promotion/International Health, Division of Public Health, Graduate School of Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8519, Japan2005 27 5 2005 5 53 53 19 10 2004 27 5 2005 Copyright © 2005 Fukuda et al; licensee BioMed Central Ltd.2005Fukuda et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Little is known about the socioeconomic differences in health-related behaviours in Japan. The present study was performed to elucidate the effects of individual and regional socioeconomic factors on selected health risk behaviours among Japanese adults, with a particular focus on regional variations.
Methods
In a nationally representative sample aged 25 to 59 years old (20,030 men and 21,076 women), the relationships between six risk behaviours (i.e., current smoking, excessive alcohol consumption, poor dietary habits, physical inactivity, stress and non-attendance of health check-ups), individual characteristics (i.e., age, marital status, occupation and household income) and regional (N = 60) indicators (per capita income and unemployment rate) were examined by multilevel analysis.
Results
Divorce, employment in women, lower occupational class and lower household income were generally associated with a higher likelihood of risk behaviour. The degrees of regional variation in risk behaviour and the influence of regional indicators were greater in women than in men: higher per capita income was significantly associated with current smoking, excessive alcohol consumption, stress and non-attendance of health check-ups in women.
Conclusion
Individual lower socioeconomic status was a substantial predictor of risk behaviour in both sexes, while a marked regional influence was observed only in women. The accumulation of risk behaviours in individuals with lower socioeconomic status and in women in areas with higher income, reflecting an urban context, may contribute to their higher mortality rates.
==== Body
Background
Health inequality within the population is a major public health concern [1,2]. Individual socioeconomic status, as measured by educational level, occupational class and income, has been shown to be closely related to both mortality and morbidity in combination with race/ethnicity, access to health care services and a number of other factors [3,4]. Previous studies confirmed by that the relationship between socioeconomic factors and health status can be explained in part by differences in health-related behaviour according to socioeconomic status [5,6]. Generally, people with lower socioeconomic status have a higher likelihood of exposure to risk behaviours, such as smoking, excessive alcohol consumption, physical inactivity, poor diet and non-attendance of health check-ups, as well as psychological stress [7-13].
A number of recent studies, especially a series of multilevel analyses, have focused on the regional and contextual effects on health [14-16], and have demonstrated effects on mortality, morbidity and self-rated health [17-20]. Regional socioeconomic conditions have been shown to influence health-related behaviour independently and interactively with individual socioeconomic status, and most previous studies indicated a higher likelihood of health risk behaviour in people living in socioeconomically disadvantaged areas [21-23].
Japan has the longest life expectancy and healthy life expectancy in the world [24], which has been shown to be due not only to improvements in the standard of living due to economic growth, but also to the relatively small degree of socioeconomic disparity within the population [25,26]. We are especially interested in the degree of socioeconomic inequalities in health and its relation to the status of the Japanese population as the healthiest worldwide, as well as the future implications.
There is only limited evidence of socioeconomic inequality in health in Japan. National data indicate a substantial gap in mortality among occupations: lower occupational classes, such as service workers and primary industry workers, show higher mortality rates than higher occupational classes, such as professionals and technical workers [27]. Previous studies demonstrated regional variations in mortality among municipalities, and areas with lower socioeconomic status were associated with higher rates of mortality [28]. However, recent figures have indicated clear gender differences in the relationships between area socioeconomic status and mortality: areas with higher income and educational level show lower mortality rates among men, but not among women [28].
Although a few studies in Japan have demonstrated a significant correlation between lower socioeconomic status and increased likelihood of risk behaviour at the individual level, the relationship was moderate as compared to that in other industrial nations [29,30]. At the regional level, there are substantial variations in the prefectural prevalence of smoking, alcohol drinking and other health risk behaviours [31]. Nevertheless, little is known about the independent impacts of individual and regional socioeconomic conditions on health risk behaviour, paying attention to integration of the influences at both levels in Japan.
The present study was performed to examine the relationships of individual and regional socioeconomic factors to health risk behaviour with regard to smoking, alcohol consumption, diet, exercise, psychological stress and attendance at health check-ups, by multilevel analysis of a nationally representative sample in Japan.
Methods
Data source
The 2001 Comprehensive Survey of the Living Conditions of People on Health and Welfare conducted by the Ministry of Health, Labour and Welfare [32] was used for analysis in the present study. All members of households within 5,240 area units, sampled randomly with stratification by prefecture in Japan, were interviewed. The survey included household and individual basic information regarding demographics, health, illness profiles, lifestyle and other items. The total number of households sampled for basic information was 247,195, of which 30,386 were interviewed with regard to income and savings. The response rates were 87.4% for the basic information survey and 79.5% for the income survey. Microdata files from this survey were used with permission from the Ministry of Public Management, Home Affairs, Posts and Telecommunications; we analysed the data for 20,030 men and 21,076 women aged 25 to 59 years whose basic and income data were surveyed.
Health risk behaviour
The following six health risk behaviours were used in the analyses: current smoking, excessive alcohol consumption, poor dietary habits, physical inactivity, stress and non-attendance of health check-ups.
Current smoking
Smoking habits were surveyed based on the following four categories: (a) "I don't smoke"; (b) "I smoke every day"; (c) "I smoke on occasion but not every day"; and (d) "I have stopped smoking for more than one month". We categorised (b) and (c) as current smokers.
Alcohol consumption
Alcohol consumption per day was surveyed, and excessive alcohol consumption was defined as more than 2.0 "gou" per day (one "gou" is a measure of 180 ml of Japanese sake, and contains almost 20 g of ethanol).
Poor dietary habits
Dietary habits were surveyed using the question, "What are your daily dietary practices?": have regular meals; have balanced meals; have bland (less salty) meals; and do not overeat. People who answered "no" to all these questions were defined as having poor dietary habits.
Physical inactivity
The question was, "Do you exercise or play sports regularly?" People who answered "no" were defined as being physically inactive.
Stress
The question was, "Do you have any stress or worries in your daily life?" People who answered "yes" were defined as being stressed.
Non-attendance of health check-ups
The subjects' attendance of health check-ups in the past year was surveyed. Health check-ups included all types of general health check-up, such as periodical health examinations in the workplace, health examinations in communities and multiphasic examinations ("Ningen Dock"), but not health examinations only for cancer screening, maternal health check-ups, dental health examinations or clinical examinations at medical facilities.
Individual socioeconomic variable
We used age, marital status, occupation and income as individual socioeconomic variables.
Subjects' marital status was categorised as married, single, widowed or divorced. Occupation classification was based on the Vital Statistics in Japan [27]: professional and technical workers (professional); managers and officials (manager); clerical and related workers (clerk); sales workers (sales work); service workers (service work); protective service workers (protective service); agriculture, forestry and fishery workers (agriculture); workers in transport and communications (transport); craftsmen, mining, production process and construction workers and labourers (labour); housework; and others including workers not classifiable by occupation, the unemployed and students.
Annual household income before tax, including benefits and inheritance, was used as income information. To adjust for family size and composition, we applied the modified OECD equivalence scale of 1.0 for the first adult, 0.5 for the second and each subsequent person aged 14 and over, and 0.3 for each child under 14 [33]. The subjects were divided into quintiles according to equivalent income, and the income quintile was used as an independent variable.
Regional socioeconomic variables
Japan consists of 47 prefectures, and the prefecture was used the basic regional unit in the present study. For prefectures including a metropolitan city/cities, i.e., the Tokyo metropolitan area (23 special wards, "ku") and 12 cities designated by ordinance (Sapporo, Sendai, Chiba, Yokohama, Kawasaki, Nagoya, Kyoto, Osaka, Kobe, Hiroshima, Kitakyushu and Fukuoka), the regional unit was divided into prefectures excluding metropolitan cities. Consequently, the individual data were linked with 47 prefectures and 13 metropolitan cities.
Per capita income and unemployment rate of prefectures and metropolitan cities were obtained from the database of census data and governmental surveys [34], and were used as regional socioeconomic variables.
Statistical analysis
Multilevel analysis was performed with data for 20,003 men and 21,076 women (level-1) nested within 60 prefectures/metropolitan cities (level-2). To estimate the average relationships between health risk behaviours and individual variables across all regions (individual fixed parameters), the variations between prefectures/metropolitan cities that could not be accounted for by individual factors (regional random variance), and the effects of regional variables on health risk behaviours (regional fixed parameter), a multilevel binomial non-linear logit link model using the iterative generalised least-squares (IGLS) was fitted [35].
Individual and regional fixed parameters were expressed by the adjusted odds ratio (OR) with 95% confidence interval (CI). The proportion of variance related to the region (intra-regional correlation:%) was approximated as: regional variance/(regional variance + π2/3) ×100 [36,37]. Statistical analyses were performed using MLwiN version 1.10 [38].
Results
Table 1 shows the basic characteristics of the subjects included in the present study. The mean age was 42.9 years for both men and women. A large proportion of the subjects were classified as "married": 73.4% for men and 76.9% for women. The majority of women reported their occupation as "housework" (27.4%). In the following analysis, "protective service" and "housework" for men, and "manager", "protective service," and "transport" for women were included in "others", because they each comprised only a very small proportion of the total. In multilevel analysis, the references were "profession" for men and "housework" for women. Per capita income and unemployment rate of 60 prefectures/metropolitan cities ranged from 865 to 2,042 (thousand yen) and 3.0 to 9.4 (%), respectively.
Table 1 Basic characteristics of study subjects and regions.
Men Women
Individual variable (20030 men and 21076 women)
Age (mean, range: years) 42.9 (25–59) 42.9 (25–59)
Marital status (N, %) Married 14701 (73.4) 16208 (76.9)
Single 4635 (23.1) 3217 (15.3)
Separated 137 (0.7) 541 (2.6)
Divorced 557 (2.8) 1110 (5.3)
Occupation a (N, %) Profession 3747 (18.7) 2243 (10.6)
Manager 1736 (8.7) 310 (1.5)
Clerk 1776 (8.9) 3051 (14.5)
Sales work 1972 (9.8) 1785 (8.5)
Service work 1697 (8.5) 2242 (10.6)
Agriculture 680 (3.4) 535 (2.5)
Transport 1063 (5.3) 80 (0.4)
Labour 4752 (23.7) 2167 (10.3)
Housework 28 (0.1) 5765 (27.4)
Others 2579 (12.9) 2898 (13.8)
Income quintile (median: thousand yen) 5th (highest) 6035
4th 4000
3rd 2989
2nd 2200
1st (lowest) 1250
Regional variable (60 prefectures and metropolitan cities)
Per capita income (mean, range: thousand yen) 1423.4 (865–2042)
Unemployment rate (mean, range: %) 4.68 (3.0–9.4)
a Profession = professional and technical workers; clerk = clerical and related workers; agriculture = agriculture, foresty and fishery workers; labour = craftmen, mining, production process and construction workers, and labourers.
The prevalence of difference health risk behaviours are shown in Table 2. Current smoking, excessive alcohol consumption and poor dietary habits were much more prevalent in men (57.3, 27.4 and 40.6%, respectively) than in women (15.6, 5.8 and 27.4%, respectively). Physical inactivity and stress were not markedly different between men and women, but women showed a higher prevalence of non-attendance of health check-ups than men (42.3% vs. 27.5%, respectively).
Table 2 Prevalence of health risk behaviour in Japanese adults
Health risk behavior Men Women
N (%) N (%)
Current smoking 10784 (57.3) 3099 (15.6)
Excess alcohol consumption 5177 (27.4) 1155 (5.8)
Physical inactivity 13989 (71.2) 14926 (72.0)
Poor dietary habits 7974 (40.6) 5671 (27.4)
Stress 9639 (51.9) 12151 (61.3)
Non-attendance of health check-ups 5243 (27.5) 8532 (42.3)
Table 3 shows the results of multilevel analysis of the relationships between health risk behaviours and individual and regional socioeconomic variables for men. Compared to married subjects, those who were divorced showed significantly higher OR for current smoking, poor dietary habits and non-attendance of health check-ups. Subjects in lower occupational classes, such as "service work", "transport," and "labour", showed significantly higher OR for risk behaviours as compared to those classified as "professional". There was a significant gradient of increased OR according to income, except with regard to excessive alcohol consumption and stress.
Table 3 Results of multilevel analysis for health risk behaviour in Japanese men: odds ratio (OR) and 95% confidence interval (95%CI) of individual and regional variables.
Variable Current smoking (N = 19816) Excess alcohol consumption (N = 18922) Physical inactivity (N = 20728) Poor dietary habits (N = 20728) Stress (N = 19817) Non-attendance of health check-up (N = 20148)
Individual OR (95%CI) OR (95%CI) OR (95%CI) OR (95%CI) OR (95%CI) OR (95%CI)
Individual
Age (10 years) 0.81 (0.78, 0.84) *** 1.17 (1.12, 1.21) *** 0.93 (0.90, 0.97) * 0.71 (0.69, 0.74) *** 1.01 (0.98, 1.04) 0.88 (0.84, 0.91) ***
Marital status
Married 1.00 1.00 1.00 1.00 1.00 1.00
Single 0.72 (0.66, 0.78) *** 0.62 (0.56, 0.68) *** 0.98 (0.90, 1.07) 1.04 (0.96, 1.13) 0.82 (0.75, 0.88) *** 1.73 (1.58, 1.89) ***
Widow 1.35 (0.93, 1.95) 0.83 (0.55, 1.25) 1.08 (0.72, 1.62) 1.28 (0.87, 1.87) 0.97 (0.67, 1.41) 1.11 (0.71, 1.74)
Divorced 1.89 (1.55, 2.31) *** 1.10 (0.91, 1.33) 1.14 (0.93, 1.40) 1.28 (1.06, 1.53) * 1.06 (0.88, 1.27) 1.75 (1.43, 2.12) ***
Occupation a
Profession 1.00 1.00 1.00 1.00 1.00 1.00
Manager 1.14 (1.01, 1.28) * 1.07 (0.93, 1.22) 0.94 (0.83, 1.07) 0.92 (0.81, 1.05) 1.03 (0.92, 1.16) 0.78 (0.67, 0.92) ***
Clerk 0.94 (0.84, 1.06) 0.97 (0.85, 1.11) 1.06 (0.93, 1.20) 0.96 (0.85, 1.09) 0.99 (0.88, 1.11) 0.59 (0.50, 0.69) ***
Sales work 1.31 (1.17, 1.47) *** 1.18 (1.04, 1.34) * 1.11 (0.98, 1.25) 1.06 (0.94, 1.19) 1.01 (0.90, 1.13) 1.68 (1.47, 1.91) ***
Service work 1.26 (1.12, 1.42) *** 1.13 (0.98, 1.29) 1.20 (1.05, 1.37) * 1.19 (1.06, 1.35) ** 1.07 (0.95, 1.21) 1.34 (1.17, 1.54) ***
Agriculture 1.16 (0.97, 1.39) 1.20 (0.99, 1.46) 1.03 (0.85, 1.25) 0.93 (0.77, 1.11) 0.73 (0.61, 0.87) *** 1.84 (1.52, 2.24) ***
Trasnport 1.61 (1.39, 1.87) *** 1.29 (1.10, 1.51) ** 1.19 (1.01, 1.39) * 1.35 (1.16, 1.56) *** 0.92 (0.80, 1.06) 1.08 (0.91, 1.28)
Labour 1.49 (1.36, 1.63) *** 1.26 (1.14, 1.39) *** 1.13 (1.03, 1.25) * 1.12 (1.02, 1.23) * 0.89 (0.81, 0.98) * 1.10 (0.99, 1.22)
Others 1.08 (0.97, 1.20) 0.91 (0.80, 1.03) 0.90 (0.80, 1.01) 1.00 (0.89, 1.11) 1.03 (0.92, 1.15) 1.33 (1.17, 1.50) ***
Income quintile
5th (highest) 1.00 1.00 1.00 1.00 1.00 1.00
4th 1.11 (1.01, 1.21) * 0.96 (0.87, 1.06) 1.17 (1.06, 1.29) *** 1.08 (0.98, 1.19) 1.10 (1.00, 1.20) * 1.10 (0.98, 1.24)
3rd 1.12 (1.02, 1.23) * 0.99 (0.89, 1.10) 1.31 (1.19, 1.45) *** 1.16 (1.06, 1.28) ** 1.07 (0.97, 1.17) 1.34 (1.19, 1.50) ***
2nd 1.30 (1.18, 1.43) *** 1.03 (0.92, 1.14) 1.43 (1.29, 1.58) *** 1.26 (1.14, 1.39) *** 1.09 (0.99, 1.20) 1.99 (1.77, 2.23) ***
1st (lowest) 1.29 (1.17, 1.43) *** 0.99 (0.89, 1.10) 1.42 (1.28, 1.58) *** 1.28 (1.16, 1.42) *** 1.15 (1.05, 1.27) ** 3.14 (2.80, 3.52) ***
Region
Per capita income (million yen) 0.89 (0.71, 1.11) 1.03 (0.82, 1.29) 1.05 (0.92, 1.20) 1.09 (0.94, 1.26) 1.44 (1.24, 1.67) *** 1.42 (1.15, 1.76) **
Unemployment (%) 1.06 (1.01, 1.12) * 1.07 (1.01, 1.13) * 0.99 (0.96, 1.02) 1.02 (0.99, 1.06) 1.01 (0.98, 1.05) 1.11 (1.05, 1.16) **
Regional random variance (SE) b 0.029 (0.008) *** 0.028 (0.008) *** 0.000 (0.000) 0.004 (0.003) 0.005 (0.003) 0.029 (0.009) **
-2 log likelihood 26317.0 22089.1 23313.1 25487.2 26207.1 19463.1
Intra-regional correlation (%) 8.0 7.7 0.0 1.1 1.6 8.0
* p < 0.05, ** p < 0.01, *** p < 0.001
a Profession = professional and technical workers; manager = managers and officials; clerk = clerical and related workers; agriculture = agriculture, forestiry and fishery workers; transport = workers in transport and communications; labour = craftmen, mining, production process, and construction workers and labourers.
b Variance at the regional level in a logit model
With regard to regional indicators, significant correlations were found between per capita income and stress and non-attendance of health check-ups, and between unemployment rate and current smoking, excessive alcohol consumption and non-attendance of health check-ups. Significant regional variance was observed for current smoking, excessive alcohol consumption and non-attendance of health check-ups.
Table 4 shows the results of multilevel analysis for women. Being divorced showed a significantly higher OR for current smoking, excessive alcohol consumption, poor dietary habits and stress. Non-attendance of health check-ups showed significantly lower OR for all types of employment as compared to "housework", while no significant differences were observed between OR for other health risk behaviours and any type of employment except for the relationship between current smoking and "agriculture". Higher income was associated with reduced likelihood of all risk behaviours. The strongest relationships were found for current smoking and non-attendance of health check-ups, which showed approximately two-fold greater odds in the lowest quintile as compared to the highest quintile.
Table 4 Results of multilevel analysis for health risk behaviour in Japanese women: odds ratio (OR) and 95% confidence interval (95%CI) of individual and regional variables.
Variable Current smoking (N = 19816) Excess alcohol consumption (N = 18922) Physical inactivity (N = 20728) Poor dietary habits (N = 20728) Stress (N = 19817) Non-attendance of check-up (N = 20148)
OR (95%CI) OR (95%CI) OR (95%CI) OR (95%CI) OR (95%CI) OR (95%CI)
Individual
Age (10 years) 0.76 (0.73, 0.79) *** 0.93 (0.87, 1.00) 0.70 (0.65, 0.75)*** 0.70 (0.68, 0.73) *** 0.93 (0.90, 0.96) ** 0.65 (0.63, 0.68) ***
Marital status
Married 1.00 1.00 1.00 1.00 1.00 1.00
Single 1.00 (0.89, 1.14) 1.00 (0.82, 1.21) 0.73 (0.65, 0.81) ** 1.12 (1.01, 1.24) * 0.73 (0.66, 0.80) *** 0.63 (0.57, 0.70) ***
Widow 1.07 (0.82, 1.40) 0.76 (0.49, 1.19) 1.11 (0.91, 1.35) 1.09 (0.87, 1.36) 0.92 (0.76, 1.11) 0.82 (0.67, 1.02)
Divorced 2.67 (2.30, 3.09) *** 1.68 (1.33, 2.11) *** 1.09 (0.93, 1.27) 1.23 (1.06, 1.42) ** 1.16 (1.01, 1.34) * 0.88 (0.76, 1.02)
Occupation a
Housework 1.00 1.00 1.00 1.00 1.00 1.00
Profession 1.08 (0.92, 1.25) 1.34 (1.06, 1.70) * 1.01 (0.90, 1.14) 1.03 (0.91, 1.17) 1.32 (1.18, 1.47) *** 0.26 (0.24, 0.30) ***
Clerk 0.97 (0.84, 1.12) 1.46 (1.19, 1.80) *** 1.15 (1.03, 1.28) ** 1.06 (0.95, 1.19) 1.20 (1.09, 1.33) *** 0.28 (0.25, 0.31) ***
Sales work 1.57 (1.35, 1.82) *** 1.79 (1.42, 2.25) *** 1.11 (0.98, 1.25) 1.43 (1.26, 1.62) *** 1.19 (1.06, 1.34) ** 0.48 (0.43, 0.54) ***
Service work 1.58 (1.37, 1.81) *** 1.91 (1.55, 2.36) *** 1.10 (0.98, 1.23) 1.37 (1.22, 1.54) *** 1.14 (1.03, 1.28) * 0.42 (0.37, 0.47) ***
Agriculture 0.70 (0.52, 0.96) * 1.21 (0.79, 1.86) 0.99 (0.81, 1.22) 1.06 (0.85, 1.32) 0.84 (0.70, 1.02) 0.56 (0.46, 0.68) ***
Labour 1.10 (0.94, 1.28) 1.07 (0.83, 1.38) 1.35 (1.20, 1.52) *** 1.38 (1.23, 1.56) *** 1.06 (0.95, 1.18) 0.36 (0.32, 0.40) ***
Others 1.30 (1.13, 1.48) *** 1.34 (1.08, 1.66) *** 1.03 (0.93, 1.15) 1.22 (1.09, 1.36) ** 1.06 (0.96, 1.17) 0.53 (0.48, 0.59) ***
Income quintile
5th (highest) 1.00 1.00 1.00 1.00 1.00 1.00
4th 1.12 (0.97, 1.29) 0.96 (0.78, 1.17) 1.06 (0.96, 1.17) 1.18 (1.06, 1.32) ** 1.06 (0.97, 1.17) 1.21 (1.09, 1.34) ***
3rd 1.34 (1.16, 1.54) ** 1.04 (0.85, 1.27) 1.11 (1.00, 1.22) * 1.24 (1.11, 1.39) *** 1.11 (1.01, 1.22) * 1.42 (1.29, 1.58) ***
2nd 1.66 (1.44, 1.90) *** 1.06 (0.86, 1.29) 1.21 (1.09, 1.34) *** 1.32 (1.19, 1.48) *** 1.14 (1.04, 1.26) ** 1.75 (1.58, 1.93) ***
1st (lowest) 2.03 (1.76, 2.33) *** 1.28 (1.04, 1.56) * 1.24 (1.12, 1.38) *** 1.64 (1.47, 1.83) *** 1.26 (1.14, 1.39) *** 2.23 (2.01, 2.47) ***
Region
Per capita income (million yen) 1.93 (1.64, 2.28) ** 1.49 (1.05, 2.10) * 0.76 (0.64, 0.90) ** 1.12 (0.90, 1.40) 1.43 (1.23, 1.67) *** 1.75 (1.13, 2.27) ***
Unemployment (%) 1.08 (0.99, 1.18) 1.10 (1.03, 1.19) * 0.95 (0.92, 0.99) ** 1.03 (0.98, 1.08) 1.00 (0.97, 1.04) 1.09 (1.03, 1.15) **
Regional random variance (SE) b 0.118 (0.026) *** 0.043 (0.019) * 0.009 (0.004) * 0.032 (0.009) *** 0.007 (0.004) 0.045 (0.011) ***
-2 log likelihood 13564.2 -2382.9 23140.5 23489.1 26988.4 23946.7
Intra-regional correlation (%) 26.1 11.4 2.6 8.8 2.1 11.8
* p < 0.05, ** p < 0.01, *** p < 0.001
a Profession = professional and technical workers; clerk = clerical and related workers; agriculture = agriculture, foresty and fishery workers; labour = craftmen, mining, production process and construction workers, and labourers.
b Variance at the regional level in a logit model
Per capita income and unemployment rate showed significant positive associations with current smoking, excessive alcohol consumption, stress and non-attendance of health check-ups. With the exception of stress, all risk behaviours showed significant regional variance. All risk behaviours showed higher intra-regional correlations as compared to men: in particular, 26.1% for current smoking, 11.8% for non-attendance of health check-ups and 11.4% for excessive alcohol consumption.
Discussion
The results of the present study indicated a substantial relationship between health risk behaviour and lower socioeconomic status at the level of the individual in Japanese men and women. The regional variance and the influence of regional socioeconomic indicators on risk behaviour were marked in women, but small in men.
Among men, those in the lower occupational classes showed a higher likelihood of risk behaviours, except for stress, as compared to "professionals". Especially, "service work," "transport" and "labour" showed significantly higher likelihood of current smoking, excessive alcohol consumption, physical inactivity and poor dietary habits. These observations suggest a plausible explanation for the higher mortality rates among these occupational classes noted in the national data [27].
Individual income was significantly related to risk behaviour of smoking, exercise, diet and health check-ups in both men and women, and lower income increased the likelihood of these behaviours. In men, a clear gradient of OR was found only for non-attendance of health check-ups and the gradient of OR for risk behaviours was not clearer than that of women. The model without occupation showed a clear gradient of OR in men [39] indicating a substantial degree of income-related inequality in health behaviour and its interaction with occupational class in Japanese men.
For women, no occupation showed a significantly lower OR as compared to "housework", with the exception of current smoking for "agriculture" and non-attendance of health check-ups for all occupational categories. This suggests that employment is associated with risk behaviour in women. Previous studies demonstrated that women's participation in society was related to a higher prevalence of smoking in accordance with the reduced intolerance toward this habit in women [40,41]. In addition, the tendency for higher OR of current smoking, excessive alcohol consumption and poor dietary habits in "sales workers" and "service workers" among the occupational categories implies the accumulation of risk behaviours in Japanese women in lower occupational classes.
It is interesting that excessive alcohol consumption did not show an income-related gradient in either men or women: the second highest quintile in men showed a significantly lower OR and the lowest quintile in women showed a significantly higher OR as compared to the highest quintile. For women, the difference in excessive alcohol consumption by occupation was greater than those in the other health risk behaviours. A previous study confirmed that participation in the workforce increases women's drinking habit in Japan [42].
The relationship between individual socioeconomic status and non-attendance of health check-ups showed a different pattern from other behaviours. Women in the "housework" category and men in the "agriculture" category were less likely to attend health check-ups. Health check-ups in people of working age are strongly dependent on the workplace [43]. Employees, particularly in large companies, have greater benefits of not only occupational health services but also preventive health services [44,45]. The steepest gradient of OR for non-attendance of health check-ups among all health risk behaviours suggests substantial inequality in receiving preventive health services according to socioeconomic status in the Japanese population.
There was a clear gender difference in the influence of regional socioeconomic indicators on health risk behaviour. All risk behaviours showed higher intra-regional correlations in women than in men, and marked differences were found for current smoking and excessive alcohol consumption. Area socioeconomic conditions have been shown to influence health-related behaviour, and in general those living in socioeconomically disadvantaged areas have a higher likelihood of health risk behaviour [7,13,23,46]. However, similarly to previous studies in France showing correlations between higher gross domestic product per capita in residential areas and both smoking and alcohol consumption [47,48], the results of the present study indicated that women living in areas with higher per capita income had higher likelihood of current smoking, excessive alcohol consumption, stress and non-attendance of health check-ups. This may be explained by the following two points.
First, the regional differences in socioeconomic conditions in Japan are relatively small, and thus, regional disadvantage and deprivation would have little influence on individual health-related behaviour in the Japanese population. The data indicated that the degree of income inequality in Japan is smaller than in other industrial countries [49,50], and the regional inequality in per capita income has decreased over the past several decades [51]. As a previous study demonstrated that national financial adjustment policy contributed to a reduction of regional disparity in health levels [52], a national minimum across the country was achieved by egalitarian social policies in Japan.
A second explanation is related to the linkage between per capita income and urban-rural differences. In Japan, indicators reflecting urbanisation, such as population size and population density, are strongly correlated with higher income, and therefore income-related indicators represent not only socioeconomic conditions but also aspects of urban-rural differences – higher per capita income indicates an urban context [28,53,54]. Therefore, the results of the present study imply a relationship between a higher likelihood of health risk behaviour and urbanisation. Urbanisation accompanied by social participation of women is likely to increase the likelihood of smoking and alcohol drinking in Japan [40-42].
One notable feature of the geographical variation in Japan is the deteriorating relative health levels of urban populations, especially for women. Osaka Prefecture, which is the second largest metropolis after the Tokyo Metropolitan Area, had the second shortest life expectancy for women among the 47 prefectures in Japan [55]. In addition, life expectancy in the Tokyo Metropolis, which had the longest life expectancy before 1965, ranked 15th for men and 37th for women in 2000 [55]. Urban areas showed higher mortality rates from cancer and ischemic heart disease than rural areas [56], and individual health-related behaviour contributes strongly to these diseases [23,57,58].
For men, higher mortality rates are found in areas with lower income- and education-related indicators [28]. As shown in the present study, regional socioeconomic indicators had little influence on health risk behaviour in men taking individual socioeconomic indicators into consideration. As mentioned above, a previous study indicated a gender-related difference in the relationship between mortality and area socioeconomic status: higher mortality rates in areas with lower per capita income were seen only in men [28]. The higher likelihood of health risk behaviour in men on the lower socioeconomic scale suggests one plausible explanation for the higher mortality in areas with lower per capita income, where lower-income individuals are more likely to live. In contrast, the relationship between health risk behaviour and higher per capita income can explain the marked deterioration of health level in women in urban areas with higher per capita income.
Finally, it is necessary to discuss the limitations of this study, as well as its strengths. The present study was performed in a large nationally representative sample with stratified random sampling, although potential differences in response rates based on socioeconomic status and region may have caused selection bias [59,60]. The questionnaire regarding health risk behaviour was comprehensive, although it was self-reported and a few behaviours (e.g., poor dietary habits and physical inactivity) are quite subjective. The items of dietary habits were based on the national guidelines for a healthy diet [61], and they (or the index formulated using these items) have been shown to be related to some aspects of physical health status [62,63]. Nevertheless, the validity and reliability of these questions have not been examined in detail, and those of single-item questions about physical activity were verified but only moderately so [64,65]. These issues probably induced misclassification bias [59,60].
In this study, we effectively applied multilevel analysis to elucidate the influence of socioeconomic factors of two levels and to demonstrate regional variances. However, our models did not consider random effects of the variables or interactions between the variables. Supplemental analysis using random slope models for household income did not show statistically significant regional variance in the effects of household income on any health risk behaviour in either men or women (data not shown).
Other limitations are related to the measurements of socioeconomic status. In the present study, household income was used as the main measure of individual socioeconomic status, which was estimated from the details of income-related information and adjusted for family size and composition. Educational attainment is also commonly used as another major measure of socioeconomic status [66,67]. As the survey used in the present study did not include educational information, no education-related variable could be introduced into the analyses. Previous studies have shown that differences in health-related behaviour among groups stratified according to educational attainment in Japan are substantial [29,30,68]. Further studies are required to clarify the independent and interactive influences of different socioeconomic measures on health risk behaviour.
Conclusion
The results of the present study demonstrated a close link between selected health risk behaviours and individual socioeconomic status in the Japanese population. A lower socioeconomic status measured according to income and occupation was generally associated with higher likelihood of health risk behaviours, such as current smoking and excessive alcohol consumption. Regional variance and the independent influence of regional indicators were marked in women, but small in men, and higher per capita income, reflecting an urban context, was related to accumulation of risk behaviours in women.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
YF designed the study, analyzed the data, and drafted the article. KN helped to interpret the results and edited the draft. TT supervised the data analysis and writing article.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study was supported by Grant-in-Aid for Scientific Research (C) by the Japan Society for the Promotion of Science (Grant No. 14570326 and 16590497).
==== Refs
Acheson D Independent inquiry into inequalities in health 2000 London: Stationary Office
Marmot M Inequalities in health N Engl J Med 2001 345 134 136 11450663 10.1056/NEJM200107123450210
Graham H Understanding health inequalities 2000 Buckingham: Open University Press
Townsend P Davidson N Whitehead M Inequalities in health: the Black Report and the health divide 1992 London: Penguin Books
Evans RG Barer ML Marmor TR Why are some people healthy and others not? 1994 New York: Aldine de Gruyter
Marmot M Wilkinson RG Social determinants of health 1999 New York: Oxford University Press
Shohaimi S Luben R Wareham N Day N Bingham S Welch A Oakes S Khaw K Residential area deprivation predicts smoking habit independently of individual education level and occupational social class. A cross sectional study in the Norfolk cohort of the European Investigation into Cancer (EPIC-Norfolk) J Epidemiol Community Health 2003 57 270 276 12646543 10.1136/jech.57.4.270
Turrell G Blakely T Patterson C Oldenburg B A multilevel analysis of socioeconomic (small area) differences in household food purchasing behaviour J Epidemiol Community Health 2004 58 208 215 14966233 10.1136/jech.2003.011031
Parks SE Housemann RA Brownson RC Differential correlates of physical activities in urban and rural adults of various socioeconomic backgrounds in the United States J Epidemiol Community Health 2003 57 29 35 12490645 10.1136/jech.57.1.29
Galobardes B Morabia A Bernstein MS Diet and socioeconomic position: does the use of different indicators matter? Int J Epidemiol 2001 30 334 340 11369739 10.1093/ije/30.2.334
Stahl T Rutten D Nutbeam D Bauman A Kannas L Abel T Luschen G Rodriquez DJA Vinck J van der Zee J The importance of the social environment for physically active lifestyle – results from an international study Soc Sci Med 2001 52 1 10 11144909 10.1016/S0277-9536(00)00116-7
Duncan C Jones K Moon G Do places matter? A multi-level analysis of regional variations in health-related behaviour in Britain Soc Sci Med 1993 37 725 733 8211288 10.1016/0277-9536(93)90366-C
Reijneveld SA The impact of individual and area characteristics on urban socioeconomic differences in health and smoking Int J Epidemiol 1998 27 33 40 9563691 10.1093/ije/27.1.33
O'Campo P Advancing theory and methods for multilevel models of residential neighborhoods and health Am J Epidemiol 2003 157 9 13 12505885 10.1093/aje/kwf171
Krieger N Theories for social epidemiology in the 21st century: an ecosocial perspective Int J Epidemiol 2001 30 668 677 11511581 10.1093/ije/30.4.668
Diez Roux AV Place, people, and health Am J Epidemiol 2002 155 516 519 11882525 10.1093/aje/155.6.516
Diez Roux AV Merkin SS Arnett D Chambless C Massing M Nieto J Sorlie P Szklo M Tyroler HA Watson R Neighborhood of residence and incidence of coronary heart disease N Engl J Med 2001 345 99 106 11450679 10.1056/NEJM200107123450205
Malmstrom M Johansson SE Sundquist J A hierarchical analysis of long-term illness and mortality in socially deprived areas Soc Sci Med 2001 53 265 275 11439812 10.1016/S0277-9536(00)00291-4
Subramanian SV Kawachi I Kennedy BP Does the state you live in make a difference? Multilevel analysis of self-rated health in the US Soc Sci Med 2001 53 9 19 11380164 10.1016/S0277-9536(00)00309-9
Sundquist K Winkleby M Ahlen H Johansson S Neighborhood socioeconomic environment and incidence of coronary heart disease: a follow-up study of 25,319 women and men in Sweden Am J Epidemiol 2004 159 655 662 15033643 10.1093/aje/kwh096
Pickett KE Pearl M Multilevel analysis of neighbourhood socioeconomic context and health outcomes: a critical review J Epidemiol Community Health 2001 55 111 122 11154250 10.1136/jech.55.2.111
Diez Roux AV Merkin SS Hannan P Jacobs DR Kiefe CI Area characteristics, individual-level socioeconomic indicators, and smoking in young adults. The coronary artery disease risk development in young adults study Am J Epidemiol 2003 157 315 326 12578802 10.1093/aje/kwf207
Sundquist J Malmstrom M Johansson SE Cardiovascular risk factors and the neighbourhood environment: a multilevel analysis Int J Epidemiol 1999 28 841 845 10597980 10.1093/ije/28.5.841
World Health Organization (WHO) World Health Report 2003 2003 Geneva: WHO
Marmot MG Davey Smith G Why are the Japanese living longer? BMJ 1989 299 1547 51 2514896
Wilkinson R National mortality rates: the impact of inequity Am J Public Health 1992 82 1082 1083 1636827
Saeki N Hirako T Sakata F Ishikawa M Nakata T Vital statistics by occupation and industry Koseino Shihyo 2000 47 10 16
Fukuda Y Nakamura K Takano T Municipal socioeconomic status and mortality in Japan: sex and age differences, and trends in 1973–1998 Soc Sci Med 2004 59 2435 2445 15474199 10.1016/j.socscimed.2004.04.012
Anzai Y Ohkubo T Nishino Y Tsuji I Hisamichi S Relationship between health practices and education level in the rural Japanese population J Epidemiol 2000 10 149 156 10860298
Nakamura Y Sakata K Kubo N Akizawa Y Nagai M Yanagawa H Smokng habits and socioeconomic factors in Japan J Epidemiol 1994 4 157 161
Kanda A Ojima T Miura Y Oguri S Okayama A Matsumura R Yanagawa H Prefectural difference in drinking, smoking, physical exercise, and obesity, and its trend Koseino Shihyo 2002 49 7 15
Ministry of Health Labour and Welfare 2001 Comprehensive Survey ofthe Living Conditions of People on Health and Welfare 2003 Tokyo: Health and Welfare Statistics Association
Mejer L Siermann C Income poverty in the European Union: children, gender and poverty gaps. Statistics in focus 2000 Luxembourg: Eurostat
Statistical Information Institute for Consulting and Analysis Municipal data of social indicators 2003 Tokyo: Statistical Information Institute for Consulting and Analysis
Layland AH Goldstein H Multilevel modelling of health statistics 2001 Chichester: John Wiliy & Sons
Snijders TAB Bosker RJ Multilevel analysis: An introduction to basic and advanced multilevel modelling 1999 London: Sage Publication
Lindstrom M Moghaddassi M Merlo J Social capital and leisure time physical activity: a population based multilevel analysis in Malmo, Sweden J Epidemiol Community Health 2003 57 23 28 12490644 10.1136/jech.57.1.23
Rashash J Browne W Goldstein H Yang M Plewis I Healy M Woodhouse M Draper D Langford I Lewis T A user's guide to MLwiN 2002 London: Centre for Multilevel Modelling, Institute of Education, University of London
Fukuda Y Nakamura K Takano T Socioeconomic pattern of smoking in Japan: income inequality and gender and age differences Ann Epidemiol 2005 15 365 72 15840550 10.1016/j.annepidem.2004.09.003
Ohida T Kamal AMM Takemura S Sone T Mochizuki Y Kawaminami K Relation between smoking prevalence and various social factors in Japan Keio J Med 2001 50 263 268 11806504
Ohida T Sone T Motizuki Y Kawaguti T Kido M Harita A Household size related to prevalence of smoking in women in Japan J Epidemiol 2000 10 305 309 11059512
Takano T Nakamura K Watanabe M Increased female drinking in accordance with post-industrial urbanization in Japan Alcohol Alcohol 1996 31 41 49 8672173
Ministry of Health and Welfare Survey of demands of health and welfare services 2000 Tokyo: Health and Welfare Statistics Association
Hamashima C Yoshida K What is important for the introduction of cancer screening in the workplace? Asian Pac J Cancer Prev 2003 4 39 43 12718699
Hirata M Kumagai S Tabuchi T Tainaka H Andoh K Oda H Actual conditions of occupational health activities in small-scale enterprises in Japan: system for occupational health, health management and demands by small-scale enterprises Sangyo Eiseigaku Zasshi 1999 41 190 201 10637944
Duncan C Jones K Moon G Smoking and deprivation: are there neighbourhood effects? Soc Sci Med 1999 48 497 505 10075175 10.1016/S0277-9536(98)00360-8
Chaix B Chauvin P Tobacco and alcohol consumption, sedentary lifestyle and overweightness in France: a multilevel analysis of individual and area-level determinants Eur J Epidemiol 2003 18 531 538 12908718 10.1023/A:1024642301343
Chaix B Guilbert P Chauvin P A multilevel analysis of tobacco use and tobacco consumption levels in France: are there any combination risk factors? Eur J Public Health 2004 14 186 190 15230507 10.1093/eurpub/14.2.186
Institute for Posts and Telecommunications Policy Indicators of regional affluence 1997 Tokyo: Institute for Posts and Telecommunications Policy
Tachibanaki T Japanese economy from the point of view of household economy 2004 Tokyo: Iwanami
Cabinet Office Report of regional economics 2001 2001 Tokyo: National Printing Bureau
Takano T Nakamura K The national financial adjustment policy and the equalization of health levels among prefectures J Epidemiol Community Health 2001 55 748 754 11553660 10.1136/jech.55.10.748
Hotta Y Araki S Sato H Yokoyama K Social life factors for the mortality from motor vehicle accident Jpn J Occup Med Traumatol 2000 48 483 489
Fukuda Y Nakamura K Takano T Wide range of socioeconomic factors associated with mortality among cities in Japan Health Promot Int 2004 19 177 188 15128709 10.1093/heapro/dah205
Health and Welfare Statistics Association Prefectural Life Table 2000 2003 Tokyo: Health and Welfare Statistics Association
Fukuda Y Nakamura K Takano T Cause-specific mortality differences across socioeconomic position of municipalities in Japan, 1973–77 and 1993–98: increased importance of injury and suicide in inequality for ages under 75 Int J Epidemiol 2005 34 100 109 15561754 10.1093/ije/dyh283
Gersten O Wilmoth JR The cancer transition in Japan since 1951 Demographic Research 2002 7 271 306
Kodama K Sasaki H Shimizu Y Trend of coronary heart disease and its relationship to risk factors in a Japanese population: a 26-year follow-up, Hiroshima/Nagasaki Study Jpn Circ J 1990 54 414 421 2398621
Sackett DL Bias in analytic research J Chron Dis 1979 32 51 63 447779 10.1016/0021-9681(79)90012-2
Delgado-Rodriguez M Llorca J Bias J Epidemiol Community Health 2004 58 635 641 15252064 10.1136/jech.2003.008466
Health and Welfare Statistics Association Kokumin Eiseino Doko 2004 Tokyo: Health and Welfare Statistics Association
Nakanishi N Tatara K Nakamura K Suzuki K Association of lifestyle with serum lipid levels: a study of middle-aged Japanese men J Epidemiol 2000 10 216 225 10959603
Morimoto K Lifestyle and health – Health theory and demonstrable research 1991 Tokyo: Igaku Shoin
Schechtman KB Barzilai B Rost K Fisher EB Measuring physical activity with a single question Am J Public Health 1991 81 771 773 2029054
Iwai N Hisamichi S Hayakawa N Inaba Y Nagaoka T Sugimori H Seki N Skata K Suzuki K Tamakoshi A Nakamura Y Yamamoto A Nishino Y Ogihara A Okamoto N Suziki H Morioka S Ito Y Wakai K Ojima T Tanaka H Nose T Ohono Y Validity and reliability of single-item questions about physical activity J Epidemiol 2001 11 211 218 11579928
Kunst AE Mackenbach JP Measuring socioeconomic inequalities in health 1997 Copenhagen: WHO Regional Office for Europe
Lynch J Kaplan G Berkman L, Kawachi I Socioeconomic position Social epidemiology 2000 New York: Oxford University Press 13 35
Nishi N Makino K Fukuda H Tatara K Effects of socioeconomic indicators on coronary risk factors, self-related health and psychological well-being among urban Japanese civil servants Soc Sci Med 2004 58 1159 1170 14723910 10.1016/S0277-9536(03)00287-9
| 15921512 | PMC1174875 | CC BY | 2021-01-04 16:28:56 | no | BMC Public Health. 2005 May 27; 5:53 | utf-8 | BMC Public Health | 2,005 | 10.1186/1471-2458-5-53 | oa_comm |
==== Front
BMC UrolBMC Urology1471-2490BioMed Central London 1471-2490-5-81589007310.1186/1471-2490-5-8Research ArticleBCG directly induces cell cycle arrest in human transitional carcinoma cell lines as a consequence of integrin cross-linking Chen Fanghong [email protected] Guangjian [email protected] Yoshiki [email protected] William A [email protected] Department of Urology, Medical College of Wisconsin, Milwaukee, WI, USA2 Human and Molecular Genetics Center, Medical College of Wisconsin, Milwaukee, WI, USA3 Co-first authors2005 12 5 2005 5 8 8 8 2 2005 12 5 2005 Copyright © 2005 Chen et al; licensee BioMed Central Ltd.2005Chen et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Current models of the mechanism by which intravesical BCG induces an anti-tumor effect in urothelial carcinoma propose a secondary cellular immune response as principally responsible. Our group has demonstrated that BCG mediated cross-linking of α51 integrin receptors present on the tumor surface elicits a complex biologic response involving AP1 and NF-κB signaling as well as the transactivation of immediate early genes. This study evaluated the direct biologic effect of cross-linking α5β1 integrin on cell cycle progression and apoptosis in two human urothelial carcinoma cell lines.
Methods
Two independent assays (MTT and Colony forming ability) were employed to measure the effect of α5β1 cross-linking (antibody mediated or BCG) on cellular proliferation. Flow cytometry was employed to measure effect of BCG and α5β1 antibody mediated cross-linking on cell cycle progression. Apoptosis was measured using assays for both DNA laddering and Caspase 3 activation.
Results
Results demonstrate that integrin cross-linking by BCG, or antibody mediated crosslinking of α5β1 resulted in a decrease in proliferating cell number. BCG treatment or α5β1 cross-linking increased the percentage of cells in G0/G1, in both 253J and T24 cell lines. Peptide mediated blockade of integrin binding site using RGDS reversed the effect BCG on both proliferation and cell cycle arrest. Apoptosis in response to BCG was not identified by either DNA laddering or Caspase 3 activation.
Conclusion
These findings show that BCG exerts a direct cytostatic effect on human urothelial carcinoma cell lines. Cell cycle arrest at the G1/S interface is a mechanism by which BCG inhibits cellular proliferation. This effect is duplicated by antibody mediated cross-linking of α5β1 and likely occurs as a consequence of crosslink-initiated signal transduction to cell cycle regulatory genes.
==== Body
Background
Bacille Calmette Guerin (BCG) remains the most effective available treatment option for non-muscle invasive urothelial carcinoma. Its superiority, both in terms of preventing recurrence and progression, has been demonstrated in multiple studies.[1-3] The mechanism responsible for BCG's superior anti-tumor activity is felt to be principally a consequence of an immune mediated response.[4] Investigators have shown the importance of a cellular immune response in orthotopic animal models of urothelial malignancy. The "effector" cell population responsible for the anti-tumor activity is currently felt to be natural killer (NK) cells.[4]
While an extensive literature supports an immune mechanism as being responsible for a portion of BCG's anti-tumor activity, a direct effect of BCG has been demonstrated by other authors. Multiple studies have demonstrated an in vitro anti-proliferative effect of BCG against human urothelial carcinoma cell lines.[5,6] Other authors have demonstrated a direct effect of BCG on other important biologic end points such as invasion.[7] The precise mechanism responsible for this direct effect is ill defined.
Work by others together with recent studies by our group has demonstrated that BCG has a direct gene regulatory effect in urothelial carcinoma cell lines. We have shown that this response is mediated via signal transduction initiated as a consequence of BCG induced cross-linking of α5β1 integrins present on the surface membrane of urothelial carcinoma cells.[8] Activation of signaling through NF-κB and AP1 initiate the transactivation of immediate early genes including interleukin 6 (IL-6).[9] Given the prevalence of NF-κB and AP1 response elements in the promoters of genes, it is likely that multiple genes are activated as a consequence of BCG/α5β1 cross-linking.
Historically, studies assessing a direct effect of BCG on tumor cells were hampered by the need to add a bacterial preparation to the culture media. In this setting it is difficult to separate a true direct anti-tumor effect of BCG from culture artifact.[6] Changes in pH, byproducts of the BCG preparation, and/or bacterial toxins have the potential to influence experimental outcome and yet fail to represent a relevant in vivo mechanism. The current study employed a non-biologic model that reproduces BCG induced signaling to determine whether BCG exerts a direct anti-tumor effect. Our results show that BCG decreases cell proliferation as measured by two separate assays of cell viability. Cell cycle arrest at the G1/S interface, rather than apoptosis, appears to be the mechanism by which this response occurs. These results are reproduced by the non-biologic signaling model in which α5β1 integrin receptors are cross-linked via antibodies and blocked by peptide fragments that inhibit the ability of fibronectin (FN) to function as a bridge for BCG mediated cross linking of these receptors.
Methods
Cell lines
The human transitional carcinoma cell lines 253J and T24 were obtained from American Type Cell Culture (Rockville, MD). Cells were maintained at 37°C, 5% CO2 in RPMI 1640 (Gibco BRL, Grand Island, NY) supplemented with 10% fetal bovine serum (FBS), penicillin, and streptomycin (complete media).
Bacillus Calmette-Guerin (BCG)
TICE BCG, living organisms of an attenuated, Bacillus of Calmette and Guerin strain of Mycobacterium bovis were used in the experiments (Organon Inc, West Orange, NJ). Freeze dried BCG was reconstituted in complete media at an estimated concentration of 2.5 × 107 viable organisms/ml. (dilution assumed average viability of 4 × 108 organisms per vial based upon manufacturer's specified range of 1 to 8 × 108 per vial)
Antibody mediated cross-linking
Antibody mediated cross-linking of α5β1 integrin was carried out as previously described.[8] Briefly, the following mouse mAbs were used: anti-α5 and anti-β1 Abs were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Affinity-purified F(ab')2 goat anti-mouse (GAM) antibody was purchased form BioSource International. Cells were resuspended in 10% FBS RPMI-1640 and cultured at 37°C for 45 min after trypsinization, then resuspended in RPMI-1640 serum-free medium (4 × 107 cells/300 μl/tube) and incubated with anti-5 or anti-1 integrin mAb (5 μg/ml) for 30 min at 4°C and washed with cold phosphate buffered saline (PBS). To initiate receptor cross-linking, F(ab')2 goat anti-mouse F(ab')2 Ab (100 μg/ml) was added after warming the cells to 37 C in RPMI-1640 medium Cells were incubated at 37C. Cells were harvested by pelleting the cells for 5 min at 500 × g.
Cell viability assay (MTT)
For the determination of the cell viability, we estimated cellular bioreduction of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) subsequent to cell exposure to either BCG or antibody mediated α5β1 cross-linking.[10] A 200 μl volume of an exponentially growing cell suspension (1 × 104 cells/ml) was seeded into a 96-well microtiter plate. Twenty-four hours later 20 μl of BCG at a ratio of 50:1 BCG to cells was added. After incubation for 1, 2, 3, or 6 days at 37°C, 20 μl of MTT solution (5 mg/ml in phosphate buffered saline, PBS) was added to each well and the plates were incubated for a further 2 hr at 37°C. 200 μl of lysis buffer (20% SDS in 50% DMF) was added to each well and cells were continuously incubated for a further 2 hr at 37°C. Optical density was measured at 570 nm. Each experiment was performed in 3 replicate wells for each treatment and carried out independently a minimum of 3 times.
Colony forming assay
Colony outgrowth of human urothelial carcinoma cells following exposure to BCG or antibody mediated cross-linking of α5β1 was measured using a colony forming assay. 1 × 105 of 253J or T24 cells were seed in 48-well plate with 1640 medium plus 10% FBS. After 24 hours cells were treated with 1 × 107 BCG. After 1, 2, 3, or 6 days incubation, the cells were trypsinized with 0.1 ml of Trypsin-EDTA per well. Trypsinized cells were resuspended in a 10 ml total volume of minimal essential medium (MEM) plus 20%FBS. 0.2 ml of this cell suspension was transferred into MEM plus 5%FBS at total volume of 25 ml. Five ml of this final cell suspension was seeded onto grid marked 60 mm culture plate with grid. Three plates were included for each group. The cells were incubated at 37°C. After 7 days of incubation, 0.5 ml of 2.5% glutaraldehyde (Sigma, St. Louis, MO) was directly added to each plate. The plates were incubated at room temperature for 45 min. The media and glutaraldehyde was aspirated and 1 ml of Wright-Giesma solution (EM Science, Gibbstown, NJ) added for 30 minutes to stain the colonies. One ml of Buffer Solution (LabChem Inc, Pittsburgh, PA) was added and the plates incubated at room temperature for an additional 30 min. The plates were gently rinsed under flowing tap water. After drying, the colonies were manually counted under magnification. For assay purposes, colonies were defined as consisting of 20 or more cells. For experiments employing antibody mediated cross-linking of α5 integrin, 1 × 105 of 253J or T24 cells were cultured in 48-well plate with 1640 plus 10% FBS at 37°C. After 24 hours, the media was changed to medium containing 5 μg/ml mouse monoclonal α5 antibody. The cells were incubated at 4°C for 45 min and then washed twice with serum free 1640 medium. Cross linking was achieved by incubating the cells with media containing GAM (10 μg/ml) at 37°C for 48 hrs. The CFA was performed as described as the above. Untreated cells, cells exposed to only primary antibody, and cells exposed to secondary antibody served as controls.
Flow cytometry
Flow cytometry was employed to quantify the cell cycle compartmentalization of TCC lines following exposure to BCG or antibody mediated cross-linking of α5β1. 5 × 106 cells/per 6 cm dish were synchronized using serum starvation for 60 hours prior to the start of the experiment. At time zero, cells were exposed to BCG and 10% fetal bovine serum (FBS) at a ratio of 50:1 BCG to cells or to 10% FBS alone. After 24 or 48 hours cells were trypsinized, centrifuged at 1500 rpm for 5 min, washed with PBS, and then treated with 50 μg/ml RNase A (Sigma). DNA was stained with 50 μg/ml propidium iodide for 10 min at room temperature and cell cycle analysis performed.
Apoptosis assay
Induction of apoptosis in response to BCG exposure or antibody mediated cross-linking of α5β1 was determined by measuring DNA fragmentation and activation of the Caspase 3 pathway.[11] For DNA fragmentation, 1 × 106 253J cells in 1640 plus 10%FBS were seeded in 6 cm plates. The following day cells were treated with 1 × 108of BCG for 48 hrs. Non-adherent BCG was washed from the plate and cells harvested in a digestion solution, consisting of 100 mM NaCL, 10 mM Tris,10 mM EDTA, 0.5% SDS, and 0.1 mg/ml Proteinase K, pH8.0. After overnight digestion at 50°C, the cell lysates were extracted with phenol/chloroform and the DNA in the aqueous phase precipitated with ethanol and resuspended in Tris-EDTA buffer. After digestion with 5 μg/ml RNase A for 1 hr, DNA was re-precipitated and DNA samples were electrophoretically separated on 1.5% agarose gel containing ethidium bromide (0.5 μg/ml). DNA was visualized by a UV transilluminator and the gels photographed.
In the caspase 3 assay, 253J cells were seeded at 1 × 106 in 6 cm plates. The next day cells were treated with BCG at a ratio of 1 to 50 for 6 hours. At intervals following BCG exposure, cells were harvested with trypsinization, washed once in PBS, fixed and permeabilized using the Cytofix/Cytoperm Kit (BD Biosciences) for 20 min at room temperature (RT), pelleted and washed with Perm/Wash buffer. Cells were then stained with anti-active caspase-3 mAb using 20 ug/1 × 106 cells for 60 min at room temperature in the dark. Following incubation with the Ab, cells were washed in Perm/Wash Buffer and analyzed by flow cytometry.
All experiments were performed in triplicate. Results are shown as the average value +/- one standard error.
Results
BCG inhibits urothelial carcinoma cell proliferation
Figures 1 and 2 demonstrate the effect of BCG exposure on the viability of the human urothelial carcinoma cell lines 253J and T24 as measured by the MTT and CF assays respectively. Both cell lines, in both assays, demonstrated a decrease in cell viability. Three days following BCG exposure, the viability of 253J and T24 cells was 41% and 58% of control values in the MTT assay. In the CF assay, colony counts were decreased in the 253J and T24 cells to 27% and 24 % respectively of control values after 3 days of exposure.
Figure 1 BCG decreases the proliferation of human urothelial carcinoma cell lines as measured by the MTT assay. Exposure of either 253J or T24 cells to BCG resulted in a decrease in viable cell number as measured by the optical density of the metabolic product. The day represents the duration of exposure to BCG
Figure 2 BCG decreases the proliferation of human urothelial carcinoma cell lines as measured by Colony Forming assay. Exposure of either 253J or T24 cells to BCG results in a time dependent decrease in viable cell number as measured by colony outgrowth. The day represents the duration of exposure to BCG.
Antibody mediated cross-linking of α5β1 integrin inhibits urothelial carcinoma cell proliferation
Figures 3 and 4 demonstrate the effect of antibody mediated α5β1 cross-linking on the viability of the human urothelial carcinoma cell lines 253J and T24 as measured by the MTT and CF assays respectively. Cross-linking decreased viability as measured by the MTT assay to 78% and 77% of untreated controls in the 253J and T24 lines at 3 days. Colony forming ability was similarly decreased in both cell lines with the average colony count reduced to 23% and 42% of untreated controls in the respective cell lines.
Figure 3 Antibody mediated cross-linking of a5b1 integrin decreases the proliferation of human urothelial carcinoma cell lines as measured by the MTT assay. Results are shown for a5 cross-linking in both 253J cells (figure 3a) and T24 cells (Figure 3b). Identical results were obtained when cross-linking was performed using antibodies to b1 (data not shown).
Figure 4 Antibody mediated cross-linking of α5β1 integrin decreases the proliferation of human urothelial carcinoma cell lines as measured by Colony Forming assay.
BCG induces cell cycle arrest at the G1/S interface
Figure 5 demonstrates the effect of BCG exposure on cell cycle compartmentalization in 253J and T24 cells. Following BCG exposure, the percentage of cells in G0/G1 increased 40% and 75% in the respective cell lines. The increase in G phase population was mirrored by a decrease in the percent of cells in S phase. The S phase cell fraction in BCG treated cells represented 60% and 65% of the control groups in 253J and T24 cells respectively.
Figure 5 Human urothelial carcinoma cells exposed to BCG undergo cell cycle arrest at the G1/S interface. Both 253J and T24 cell lines increased the percentage of cells in G1, with a concomitant decrease in the S phase fraction, following BCG exposure.
Antibody mediated α5β1 integrin cross-linking induces cell cycle arrest at the G1/S interface
Figure 6 demonstrates the effect of antibody mediated α5β1 cross-linking on cell cycle compartmentalization in 253J and T24 cells. Following cross-linking, the percentage of cells in G0/G1 increased 29% and 48% in the respective cell lines. The increase in G phase population was mirrored by a decrease in the percent of cells in S phase. The S phase cell fraction in cross-linked cells represented 71% and 60% of the control groups in 253J and T24 cells respectively.
Figure 6 6a and 6b. Cell cycle compartmentalization in 253J (Figure 6a) and T24 (Figure 6b) cells following antibody mediated cross-linking of α5β1 integrin. Similar to the effect observed in response to BCG, antibody mediated cross-linking resulted in an accumulation of cells in G0-G1 with a concomitant decrease in the S-phase fraction.
Peptide blockade of integrin fibronectin binding sites using RGDS reverses the effect of BCG on proliferation and cell cycle arrest
BCG attaches to the cell surface via a fibronectin bridge linking mycobacterial receptors to cell surface integrins.[12-14] Mycobacterial binding of multiple integrin bound fibronectin molecules can affect crosslinkage of integrin receptors.[8] Blocking the ability of FN to bind to either cellular or mycobacterial receptors precludes BCG attachment to the cell surface. [8,14] This series of experiments employed a competitive inhibitor of FN binding to integrins (RGDS) to assess the effect of inhibiting BCG binding to integrin receptors on the biologic response to BCG. While RGDS competitively inhibits the ability of FN to function as an opsonin for BCG binding (effectively inhibiting BCG/integrin interaction), prior studies have demonstrated that RGDS does not effect integrin mediated signaling initiated in response to antibody mediated cross linking of α5β1.[8] 253J cells were pre-incubated with RGDS or the control (non-blocking) peptide RGES for 1 hour prior to BCG exposure. The MTT and flow cytometric assays were then carried out as described above. Controls included each peptide alone, BCG alone, and untreated cells. In the MTT assay 2 concentrations of both the blocking and control peptides were employed (100 or 250 ug/ml). The 250 ug/ml concentration was used in the flow cytometry assay.
Figure 7 demonstrates the effect of integrin blockade on the anti-proliferative effect of BCG as measured by MTT assay. RGDS reduced the BCG effect in a "dose-response" manner with a maximal 4 fold inhibition at the 250 ug/ml concentration. RGES had no effect. In the flow cytometric assay, RGDS inhibited G1 arrest. The BCG induced increase in the G1 fraction was reduced by 60% in RGDS treated cells. RGES pretreatment had no effect on G1 or S phase fraction compared to controls (Figure 8).
Figure 7 The effect of BCG on 253J Cellular Proliferation as measured by MTT assay. Peptide mediate integrin blockade (RGDS) reversed the anti-proliferative effect of BCG in a dose-response manner. The control peptide RGES had no effect.
Figure 8 The effect of BCG on 253J cell cycle compartmentalization as measured by flow cytometry. Peptide mediate integrin blockade (RGDS) reversed the G1 cell cycle arrest effect of BCG. The control peptide RGES had no effect.
BCG does not induce an apoptotic response in urothelial carcinoma cell lines
A series of experiments were conducted to determine if BCG exposure induced apoptosis in either the 253J or T24 cell lines. Gel electrophoresis of cellular DNA obtained at intervals following BCG exposure failed to demonstrate the laddering associated with apoptosis. Flow cytometry for the apoptotic pathway enzyme caspase 3 failed to demonstrate activation of this enzyme in response to BCG. (Data not shown)
Discussion
The anti-tumor effect of BCG is well established. Clinical experience, the results of animal models, as well as in vitro experiments provide data in clear support of this effect. While an anti-tumor effect of BCG is certain, the mechanism through which this occurs is less clear. In vivo animal models of BCG's anti-tumor effect suggest a critical role for a cell mediated immune response. At the same time, in vitro studies using human urothelial cell lines identify a direct effect of BCG. The relative contribution of these two possible mechanisms to BCG's clinical anti-tumor activity is unknown.
A number of studies suggest that a direct anti-tumor effect is at least in part responsible for BCG's activity. Work by Pryor, Ciao, and Sasaki all have reported a direct anti-proliferative effect of BCG.[5,6,15] Liu et al demonstrated that BCG, as a consequence of its interaction with cell surface bound fibronectin, abrogates invasion and motility of urothelial carcinoma cell lines.[7] Cellular internalization of BCG, with resultant alterations in reactive oxygen species and nitric oxide, has been proposed as a mechanism contributing to direct BCG mediated cytotoxicity. [16,17]
Fibronectin functions as an opsonin, linking BCG to cell surface fibronectin receptors of which α5β1 integrin is the predominate FN receptor on urothelial cells. This linkage is a requisite step for BCG's antitumor activity.[14] Rather than constituting a passive interaction, our prior reports have shown that FN mediated BCG adherence to the urothelial carcinoma surface has a pharmacogenetic effect as exemplified by transactivation of IL-6.[9] This effect is mediated through signal transduction pathways involving NF-κB and AP-1. BCG induced signaling and gene transactivation pathways are identical to those observed in response to antibody mediated cross linking of α5β1.[8] Strategies that prevent the ability of FN to function as an opsonin, including competitive inhibition of FN receptors using RGDS and simultaneous saturation of mycobacterial and cell surface FN receptors using excess exogenous FN, inhibit the signaling and transactivation responses to BCG.[8,18] The current report moves beyond the details of the signaling pathway to examine the biologic effect of BCG induced signaling on a tumor biologic end point with clear treatment relevance.
Consistent with what has been reported by others, our results demonstrate that BCG exerts a direct anti-proliferative effect on human urothelial carcinoma cell lines. The observed anti-proliferative effect does not appear to occur as a consequence of apoptosis. To our knowledge this report is the first to identify cell cycle arrest at the G1/S interface as a mechanism by which BCG exerts an anti-proliferative effect. As was the case for BCG induced signaling and gene transactivation, this effect was reproduced using antibody mediated cross-linking of α5β1 integrin. Elimination of the opsonin function of fibronectin using the peptide competitive inhibitor RGDS, effectively blocking the ability of BCG to bind integrin receptors, reversed the biologic effects of BCG. These findings, together with the results of prior studies, demonstrate that a portion of the biologic response to BCG occurs as a consequence of FN mediated binding of BCG to integrin receptors present on the urothelial cell surface. The failure of simple integrin ligation to simulate the BCG response, the ability of antibody mediated crosslinking of α5β1 integrin to duplicate the BCG response, the central role of FN in the biologic response to BCG, together with the predominance of α5β1 as the principal FN binding integrin on urothelial cells, strongly supports a model in which the biologic response of the tumor cell to BCG occurs as a consequence of BCG crosslinking of α5β1 integrin receptors.
The results of this study have direct clinical relevance. The demonstration that BCG has a direct anti-tumor effect opens this pathway for potential manipulation to improve treatment outcome. This is important given the fact that 30% of patients are BCG refractory, and that the long-term durability of response to BCG is limited. Finally, and equally important, the dissection of the pathway through which BCG exerts an anti-tumor effect will allow us to move further towards our ultimate goal of eliminating the need to expose patients to the risk of a viable biologic organism and perhaps allow the development of less toxic equally effective treatment approaches.
Conclusion
BCG, as a consequence of integrin cross-linking, exerts direct anti-tumor effect against human urothelial carcinoma cell lines. Cell cycle arrest at the G1/S interface, rather than apoptosis, is a mechanism contributing to BCG's anti-proliferative effect. Additional studies will be required to define the molecular pathways that contribute to this response.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
FC carried out the flow cytometric experiments for cell cycle arrest and apoptosis and participated in drafting the manuscript. GZ performed the cell viability and colony forming assays and aided in drafting the manuscript. YI developed the protocols for flow cytometric experiments, aided in the design of the studies and drafted the manuscript. WAS conceived of the study, participated in its design and coordination, and drafted the manuscript. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This work was supported by a grant from the Department of Veterans Affairs and the Milwaukee Veterans Affairs Medical Center.
==== Refs
Malkowicz SB Intravesical therapy for superficial bladder cancer Semin Urol Oncol 2000 18 280 8 11101091
Bohle A Jocham D Bock PR Intravesical bacillus Calmette-Guerin versus mitomycin C for superficial bladder cancer: a formal meta-analysis of comparative studies on recurrence and toxicity J Urol 2003 169 90 5 12478111 10.1097/00005392-200301000-00023
Sylvester RJ van der MEIJDEN AP Lamm DL Intravesical bacillus Calmette-Guerin reduces the risk of progression in patients with superficial bladder cancer: a meta-analysis of the published results of randomized clinical trials J Urol 2002 168 1964 70 12394686 10.1097/00005392-200211000-00016
Bohle A Brandau S Immune mechanisms in bacillus Calmette-Guerin immunotherapy for superficial bladder cancer J Urol 2003 170 964 9 12913751 10.1097/01.ju.0000073852.24341.4a
Pryor K Stricker P Russell P Antiproliferative effects of bacillus Calmette-Guerin and interferon alpha 2b on human bladder cancer cells in vitro Cancer Immunol Immunother 1995 41 309 16 8536277
Seow SW Rahmat JN Mohamed AA Lactobacillus species is more cytotoxic to human bladder cancer cells than Mycobacterium Bovis (bacillus Calmette-Guerin) J Urol 2002 168 2236 9 12394766 10.1097/00005392-200211000-00105
Liu BC Redwood SM Weiss RE Bacillus Calmette-Guerin abrogates in vitro invasion and motility of human bladder tumor cells via fibronectin interaction J Urol 1992 148 900 5 1512857
Chen F Zhang G Iwamoto Y See WA Bacillus Calmette-Guerin initiates intracellular signaling in a transitional carcinoma cell line by cross-linking α5β1 integrin J Urol 2003 170 605 10 12853840 10.1097/01.ju.0000067623.43110.4c
Chen F-H Crist SA Zhang G-J Interleukin-6 production by human bladder tumor cell lines is upregulated by BCG through NF-κb and AP-1 via an immediate early pathway J Urol 2002 168 786 797 12131369 10.1097/00005392-200208000-00099
Mosmann T Rapid colorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays J Immunol Methods 1983 65 55 63 6606682 10.1016/0022-1759(83)90303-4
Schultz DR Harrington WJ Jr Apoptosis: programmed cell death at a molecular level Semin Arthritis Rheum 2003 32 345 69 12833244 10.1053/sarh.2003.50005
Cheng DL Shu WP Choi JC Margolis EJ Droller MJ Liu BC Bacillus Calmette-Guerin interacts with the carboxyl-terminal heparin-binding domain of fibronectin: implications for BCG-mediated anti-tumor activity J Urol 1994 152 1275 80 8072119
Zhao W Schorey JS Bong-Mastek M Ritchey J Brown EJ Ratliff TL Role of a bacillus Calmette-Guerin fibronectin attachment protein in BCG-induced anti-tumor activity Int J Cancer 2000 86 83 8 10728599 10.1002/(SICI)1097-0215(20000401)86:1<83::AID-IJC13>3.0.CO;2-R
Kavoussi LR Brown EJ Ritchey JK Ratliff TL Fibronectin-mediated Calmette-Guerin bacillus attachment to murine bladder mucosa. Requirement for the expression of an anti-tumor response J Clin Invest 1990 85 62 7 2404029
Sasaki A Kudoh S Mori K Are BCG effects against urinary bladder carcinoma cell line T24 correlated with apoptosis in vitro Urologia Internationalis 1997 59 142 8 9428429
Pook SH Esuvaranathan K Mahendran R N-acetylcysteine augments the cellular redox changes and cytotoxic activity of internalized mycobacterium bovis in human bladder cancer cells J Urol 2002 168 780 5 12131368 10.1097/00005392-200208000-00098
Jansson OT Morcos E Brundin L Lundberg JO Adolfsson J Soderhall M Wiklund NP The role of nitric oxide in bacillus Calmette-Guerin mediated anti-tumour effects in human bladder cancer Br J Cancer 1998 78 588 92 9744496
Zhang GJ Chen FH Xu Y Cao Y Crist S McKerrow A Iwamoto Y See WA Autocrine overexpression of fibronectin by human transitional carcinoma cells impairs BCG adherence and signaling J Urol 2004 172 1496 1500 15371878 10.1097/01.ju.0000140193.95528.91
| 15890073 | PMC1174876 | CC BY | 2021-01-04 16:30:02 | no | BMC Urol. 2005 May 12; 5:8 | utf-8 | BMC Urol | 2,005 | 10.1186/1471-2490-5-8 | oa_comm |
==== Front
Cardiovasc DiabetolCardiovascular Diabetology1475-2840BioMed Central London 1475-2840-4-61589289410.1186/1475-2840-4-6Original InvestigationMetabolic effect of telmisartan and losartan in hypertensive patients with metabolic syndrome Vitale Cristiana [email protected] Giuseppe [email protected] Carlotta [email protected] Alessandra [email protected] Arianna [email protected] Massimo [email protected] Maurizio [email protected] Giuseppe MC [email protected] Department of Medical Sciences and Rehabilitation, IRCCS San Raffaele, Rome, Italy2 Department of Cardiology, University of Cagliari, Calgliari, Italy2005 15 5 2005 4 6 6 10 3 2005 15 5 2005 Copyright © 2005 Vitale et al; licensee BioMed Central Ltd.2005Vitale et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Metabolic syndrome is a cluster of common cardiovascular risk factors that includes hypertension and insulin resistance. Hypertension and diabetes mellitus are frequent comorbidities and, like metabolic syndrome, increase the risk of cardiovascular events. Telmisartan, an antihypertensive agent with evidence of partial peroxisome proliferator-activated receptor activity-gamma (PPARγ) activity, may improve insulin sensitivity and lipid profile in patients with metabolic syndrome.
Methods
In a double-blind, parallel-group, randomized study, patients with World Health Organization criteria for metabolic syndrome received once-daily doses of telmisartan (80 mg, n = 20) or losartan (50 mg, n = 20) for 3 months. At baseline and end of treatment, fasting and postprandial plasma glucose, insulin sensitivity, glycosylated haemoglobin (HBA1c) and 24-hour mean systolic and diastolic blood pressures were determined.
Results
Telmisartan, but not losartan, significantly (p < 0.05) reduced free plasma glucose, free plasma insulin, homeostasis model assessment of insulin resistance and HbAic. Following treatment, plasma glucose and insulin were reduced during the oral glucose tolerance test by telmisartan, but not by losartan. Telmisartan also significantly reduced 24-hour mean systolic blood pressure (p < 0.05) and diastolic blood pressure (p < 0.05) compared with losartan.
Conclusion
As well as providing superior 24-hour blood pressure control, telmisartan, unlike losartan, displayed insulin-sensitizing activity, which may be explained by its partial PPARγ activity.
angiotensin II receptor blockerstelmisartanlosartanhypertensionmetabolic syndrome
==== Body
Background
Metabolic syndrome describes the presence of a cluster of common cardiovascular risk factors, including hypertension, insulin resistance or glucose intolerance, visceral obesity, atherogenic dyslipidemia, prothrombotic state and proinflammatory state in a single individual [1,2]. The lack of a universally agreed definition has impeded epidemiologic work on the prevalence and antecedents of this syndrome. Nevertheless, it has been proposed that the metabolic syndrome is present in about 10–25% of individuals in industrialized countries [3,4]. The increasing availability and abundance of high-calorie, low-fiber foods and the adoption of more sedentary lifestyles are also leading to increased prevalence of the metabolic syndrome in developing countries [5]. Its presence predicts a two- to four-fold increase in the risk of cardiovascular disease and death [6,7] and the risk of developing type 2 diabetes is increased five- to nine-fold [3,8].
In general, components of the metabolic syndrome are treated individually, there being no current treatment that targets all features. Some classes of antihypertensives, notably calcium channel blockers, angiotensin-converting enzyme (ACE) inhibitors and angiotensin II receptor blockers (ARBs), have been shown to reduce the incidence of new-onset diabetes, particularly when compared with diuretics and β-blockers [9]. This suggests that antihypertensive agents have differential effects on hyperglycemia in patients with metabolic syndrome. However, there are few data on intra-class differences. Recent in vitro and animal studies suggest that telmisartan, unlike other ARBs, acts as a partial peroxisome proliferator-activated receptor-gamma (PPARγ) agonist at concentrations that are achievable with oral doses recommended for the treatment of hypertension, thus suggesting its insulin-sensitizing effect [10-12]
The aim of the present study was to compare the glucometabolic effect of telmisartan and losartan, two ARBs with potentially different effects on glycemia, in patients with metabolic syndrome.
Materials and methods
The study population included men and women aged between 18 and 75 years with arterial hypertension and the diagnosis of metabolic syndrome. All subjects were newly diagnosed as being hypertensive (office systolic blood pressure [SBP] ≥ 135 mmHg, diastolic blood pressure [DBP] ≥ 85 mmHg). Patients were required to have insulin resistance, impaired glucose tolerance (IGT) or type 2 diabetes, according to the diagnostic criteria for the metabolic syndrome of the World Health Organization [1]. Insulin resistance was defined as HOMA-IR > 3.5, impaired glucose tolerance (IGT) was defined as 2-hour values in the oral glucose tolerance test (OGTT) of ≥ 140 mg/dl (≥ 7.8 mmol/l), but <200 mg/dl (<11.1 mmol/l). Diabetes was diagnosed as free plasma glucose (FPG) ≥ 126 mg/dl (≥ 7.0 mmol/l) or 2-hour post-glucose load of ≥ 200 mg/dl (≥ 11.1 mmol/l). Patients with hyperkalemia or serum creatinine >2 mg/dl were excluded.
After evaluation of all inclusion and exclusion criteria, eligible patients entered a randomized, parallel-group, double-blind study. After a baseline 24-hour ambulatory blood pressure monitoring and an OGTT, they were assigned to the two treatment groups using equal weighting and electronic randomization, and received either once-daily telmisartan 80 mg or losartan 50 mg for 3 months. These dosages were employed because they are the highest approved for the treatment of hypertension on the basis of Italian licensing. Patients were asked to adhere to their standard eating habits and physical activity throughout the study.
Patients were assessed at baseline (first visit) and after 3 months' treatment. Fasting (minimum 12 hours) blood samples (10 ml) were obtained for laboratory evaluation of hematology and clinical chemistry parameters, including total cholesterol, LDL cholesterol, high-density lipoprotein (HDL) cholesterol, triglycerides, glucose and insulin. An OGTT was conducted using 75 g glucose. Blood samples (10 ml) were withdrawn at 30-minute intervals over 120 minutes for determination of glucose and insulin response. An autoanalyzer (Olympus) was used to assay plasma glucose using the hexokinase method, plasma triglycerides using the glycerol-3-phosphate oxidase-p-aminophenazone method; cholesterol using cholesterol oxidase phenol ampyrone method; HDL cholesterol using immunoinhibition; glycosylated hemoglobin (HbA1c) with the Abbott AxSYM analyzer (Abbott SpA Divisione Diagnostici, Rome, Italy) using microparticle enzyme immunoassay; and free plasma insulin (FPI).
Insulin resistance was measured using the homeostasis model assessment (HOMA-IR) [13], defined by the following formula:
Trough clinical blood pressures were recorded at baseline and after treatment using cuff sphygmomanometry. Ambulatory blood pressure monitoring (ABPM) was performed with an oscillometric device (Tonoport V; GE Medical Systems IT Inc., Milwakee, WI, USA) recommended for clinical use and that had previously been validated. The monitor was programmed to measure SBP and DBP every 20 minutes between 06.00 and 22.59 (daytime period) and every 30 minutes between 23.00 and 05.59 (night-time period). A standard, adult-sized arm cuff (length 12 cm) was positioned in the middle of the non-dominant arm covering the brachial artery above the antecubital fossa. The correct position for the cuff was confirmed when three SBP and DBP values were concordant (within 5 mmHg) with those obtained from the opposite arm with a standard sphygmomanometer. The arm cuff was inflated automatically by a pump, and the blood pressure was digitally recorded on the hard disk of a personal computer to which the device was connected. Patients were instructed to continue their usual daily activities, but to keep their arm still and parallel to the trunk during ambulatory blood pressure measurements, and to return to the hospital 24 hours after initiation of the ABPM.
ABPM data were excluded from analysis if >30% of the measurements were lacking, if data were missing for >3-hour spans, or if collected from patients who were experiencing an irregular rest-activity schedule or a night-time sleep span was <6 hours or >12 hours during ABPM. Mean SBP and DBP values for the daytime (06.00–22.59) and night-time (23.00-05.59) periods were calculated as mean values of the hourly averages. Smoothness index was calculated as the ratio of the standard deviation of the hourly blood pressure value to the 24-hour mean [14].
Body mass index (BMI) was measured as the ratio of weight (kg) to height (m2). Waist circumference was measured with a tape measure placed horizontally around the abdomen at the level of iliac ridge at the end of a normal expiration, keeping the tape well tense, adhered to the skin and parallel to the floor. Any adverse event was recorded.
Data are presented as mean ± 1 SD or percentages when appropriate. After testing data for normality, Wilcoxon Signed Rank test was used to compare values before and after each therapy and the relative changes in values in response to each therapy. The effects of the losartan and telmisartan on blood pressure and glucose metabolism were analyzed by one way repeated measures analysis of variance (ANOVA) or Friedman Repeated ANOVA on Ranks. A p value <0.05 was considered statistically significant.
Results
A total of 40 patients were enrolled, with 20 randomized to each treatment group, baseline clinical characteristics of study patients are shown in table 1, no significant differences were noted between groups. All but four patients had IGT, whereas one patient in the losartan group and three in the telmisartan group had a diagnosis of type 2 diabetes.
Table 1 Patient baseline characteristics
Losartan (n = 20) Telmisartan (n = 20) P value
Mean ± SD age (years) 56.2 ± 11.0 55.3 ± 12.4 NS
Males/females 11/9 12/8
Office blood pressure (mmHg)
Mean ± SD SBP 149.7 ± 9.0 151.3 ± 7.1
Mean ± SD DBP 91.2 ± 7.4 89.8 ± 8.7
24-hour mean blood pressure (mmHg)
Mean ± SD SBP 142.8 ± 12.0 143.6 ± 14.0 NS
Mean ± SD DBP 88.8 ± 10.2 88.3 ± 9.5 NS
Mean ± SD body mass index (kg/m2) 32.1 ± 7.2 34.5 ± 6.3 NS
Impaired glucose tolerance (n) 19 17 NS
Type 2 diabetics (n) 1 3 NS
Metabolic syndrome components (n)a
3 11 10 NS
4 7 8 NS
5 2 2 NS
Total cholesterol (mg/dl) 212.6 ± 45.8 209.6 ± 50.8 NS
Low-density lipoprotein cholesterol (mg/dl) 134 ± 44 138 ± 34 NS
High-density lipoprotein cholesterol (mg/dl) 51.2 ± 11 56.3 ± 17 NS
Triglycerides (mg/dl) 221 ± 32 210 ± 23 NS
a Based on World Health Organization criteria [1].
Changes in metabolic parameters were observed at the end of treatment compared with baseline (table 2). Compared with losartan, telmisartan reduced FPG by 8% (p < 0.05), FPI by 10% (p < 0.06), HOMA-IR by 26% (p < 0.05) and HbA1c by 9% (p < 0.05) as shown in figure 1. Losartan did not have a meaningful effect on these parameters. Levels of FPG and FPI following OGTT were also significantly reduced by telmisartan compared with losartan (figures 2 and 3).
Table 2 Mean ± SD metabolic parameters at baseline and end of treatment
Baseline p value End of treatment p value
Losartan Telmisartan Losartan Telmisartan
HOMA-IR 5.78 ± 3.53 5.74 ± 3.35 NS 5.82 ± 2.66 4.24 ± 2.36 < 0.05
FPG 110.05 ± 14.56 109.08 ± 16.67 NS 113.20 ± 12.68 100.00 ± 11.99 < 0.05
FPI 20.47 ± 9.64 18.86 ± 10.89 NS 20.14 ± 9.49 16.93 ± 8.91 < 0.06
2-hour OGTT 137.42 ± 32.5 131.31 ± 42.05 NS 134.6 ± 26.71 113.85 ± 42.14 < 0.01
HbA1c 6.27 ± 0.29 6.45 ± 0.35 NS 6.28 ± 0.21 5.85 ± 0.18 < 0.05
Figure 1 Effect of telmisartan and losartan on measures of glycaemia and insulin resistance in 40 patients with metabolic syndrome. FPG = fasting plasma glucose, FPI = fasting plasma insulin, HOMA-IR = homeostatic model assessment – insulin resistance, HbA1c = glycosylated haemoglobin.
Figure 2 Effect of telmisartan and losartan during the OGTT in patients with metabolic syndrome. A) Plasma glucose at baseline. B) Plasma glucose at endpoint. C) Plasma insulin at baseline. D) Plasma insulin at endpoint.
Figure 3 Effect of telmisartan and losartan on the smoothness index at endpoint.
After 3 months' treatment, telmisartan reduced 24-hour mean SBP and DBP significantly more than losartan. The superior blood pressure control with telmisartan was also apparent when changes in mean daytime SBP (13.5 ± 0.8 vs 10.0 ± 1.1 mmHg; p < 0.05) and DBP (8.9 ± 0.6 vs 5.6 ± 0.8 mmHg; p = 0.04) and mean night-time SBP (8.7 ± 0.9 vs 5.6 ± 1.3 mmHg; p < 0.05) and DBP (7.8 ± 1.1 vs 4.7 ± 0.8 mmHg; p < 0.05) were compared. There was no significant correlation between the decrease in blood pressure and the change in FPG (r = 0.28; p = 0.020) or FPI (r = 0.036; p = 0.012). Telmisartan also improved the SBP and DBP smoothness indices (figure 3).
Both telmisartan and losartan were well tolerated, with no adverse events reported.
Discussion
This study found that, compared with once-daily losartan 50 mg, once-daily telmisartan 80 mg reduced 24-hour mean blood pressure and blood pressure variability, and improved glucose tolerance and insulin sensitivity. Improvements were found in all three indices of glucose and insulin metabolism- FPG, OGTT and HbA1c suggesting that Telmisartan may be effective in reducing the progression of metabolic syndrome.
Losartan is an ARB that has been shown in the Losartan Intervention For Endpoint reduction in hypertension (LIFE) to reduce new-onset diabetes compared with atenolol [15]. However, β-blocker therapy is a risk factor for the development of diabetes [16,17], so the hyperglycemic effect of atenolol may explain this result. Telmisartan is an ARB with a longer duration of action than losartan [18]. Given once daily, telmisartan 80 mg significantly reduced 24-hour blood pressure compared with losartan 50 mg, with especially large benefit in the last 6 hours of the dosing interval [18].
There is also clinical evidence that telmisartan has favourable metabolic effects. Previous studies showed that telmisartan 80 mg, but not valsartan 160 mg has an insulin-sensitizing effect [19]. A 1-year study in patients with type 2 diabetes treated with telmisartan or eprosartan found that only telmisartan improved plasma lipid profiles [20], but did not significantly affect glycemia or insulin sensitivity. However, a relatively low dose of telmisartan (40 mg once daily) was used in that study and, since telmisartan acts as a partial PPARγ agonist, this may have been insufficient for a full manifestation of any hypoglycemic effects. Telmisartan has been shown to improve plasma total cholesterol and low-density lipoprotein (LDL) cholesterol compared with nifedipine gastrointestinal therapeutic system in patients with type 2 diabetes and mild hypertension [21]. Furthermore, in a German surveillance study of hypertensive patients receiving telmisartan, the patients with type 2 diabetes had substantially reduced plasma glucose and serum triglyceride concentrations after 6 months' treatment [22].
FPG is the standard test used to diagnose Type 2 diabetes, but it is also a marker for cardiovascular disease in its own right [23]. Non-diabetic individuals with an FPG ≥ 100 mg/dl (≥ 5.6 mmol/l) but <126 mg/dl (<7.0 mmol/l) are considered to have impaired fasting glucose and are at increased risk of cardiovascular complications [24]. The physiologic basis of the response to OGTT differs from that of impaired FPG, with postprandial hyperglycemia closely linked to a blunting of early-phase insulin release [25]. It is often one of the earliest abnormalities that can be detected in clinical practice (although OGTT is not recommended for routine clinical use). HbA1c provides an index of plasma glucose concentrations during the previous 2–3 months [26]. It reflects both fasting and postprandial plasma glucose and, therefore, represents an independent parameter [27]. In patients such as ours, with high-normal levels of HbA1c, it is more closely related to postprandial plasma glucose than to fasting values [27,28].
In addition to these measures of glycemia, this study also used the HOMA-IR index, a measure of insulin resistance derived from fasting levels of glucose and insulin and a physiologically-based model [29], which is an effective, easily-derived surrogate for the more complex euglycemic clamp [30]. HOMA-IR was predictive of future diabetes in the Mexico City Diabetes Study [31]. The reduction of HOMA-IR seen in our study may, therefore, reduce the progression from metabolic syndrome, although this has not been studied experimentally.
In this study, losartan had no effect on measures of glycemia or on HOMA-IR; a result that may seem surprising given that losartan reduced the incidence of new-onset diabetes in LIFE [15]. However, a previous study in hyperinsulinemic, hypertensive patients found no effect on insulin sensitivity or glucose tolerance following 12 weeks' treatment with either losartan or metoprolol [32]. This supports our finding that losartan is metabolically neutral, but leaves open the question of whether the results of LIFE were due to a pro-diabetogenic effect of atenolol. They also support previous studies which suggest that the PPARγ agonism exhibited by telmisartan in preclinical studies [10-12] has meaningful effects at the clinical level.
There have been relatively few studies of PPARγ agonists in patients without diabetes. In one study, 24 hypertensive, non-diabetic patients (mean BMI of 30 kg/m2) were given rosiglitazone in addition to non-ACE inhibitor antihypertensive therapy for 12 weeks. The result was a reduction in FPI (but not FPG) and an increase in glucose disposal (measured using euglycemic clamp) [33]. In an 8-week, placebo-controlled study, 50 non-diabetic patients who met a modified National Cholesterol Education Program definition for the metabolic syndrome were randomized to receive either rosiglitazone 4 mg/day or placebo for 8 weeks. In these patients, rosiglitazone reduced FPI by 40% and HOMA-IR by 45% compared with placebo [34]. The magnitude of the sensitizing effect with rosiglitazone was somewhat greater than that observed with telmisartan in our study. Although these difference may be a function of the differing population and study protocol, they may also relate to the in vitro observations that telmisartan is a selective PPARγ modulator (SPPARM) [10]. SPPARMs activate only a subset of genes targeted by full PPARγ agonists [35] and they may, in particular, have a better adverse event profile. For this reason, it is notable that telmisartan was well tolerated in our study as in previous ones, with none of the peripheral oedema and fluid retention that are characteristic of full PPARγ agonists [36].
As expected, both telmisartan and losartan reduced blood pressure in our patients; however, reductions in 24-hour mean SBP and DBP were significantly greater with telmisartan. A superior reduction in 24-hour mean SBP and DBP with telmisartan 80 mg compared with losartan 50 mg has been found in a meta-analysis of previous studies, partly due to telmisartan's longer duration of action [18]. The greater improvement in the smoothness index with telmisartan compared with losartan is also significant, given that this is an independent prognostic marker for cardiovascular events [37].
Compared to other AT(1) receptor blockers telmisartan may have further additional beneficial effects in patients with the metabolic syndrome as suggested by this study and by a recent report of Zhang et al that have shown that that AT(1) receptor-mediated coronary constriction that is augmented in the prediabetic metabolic syndrome and contributes to impaired control of coronary blood flow is beneficially affected by telmisartan[38].
Conclusion
This study found that telmisartan, but not losartan, improves metabolic parameters in patients with metabolic syndrome. Although treatment conventionally focuses on each risk factor individually, multifactorial intervention reduces significantly the incidence of cardiovascular disease in type 2 diabetics with microalbuminuria. The multifactorial effects of telmisartan shown in this study could, therefore, provide synergistic benefits in patients with hypertension and other cardiovascular risk factors, such as glucose intolerance. Such a provocative hypothesis will require confirmation in large clinical trials, such as the ONgoing Telmisartan Alone and in combination with Ramipril Global Endpoint Trial (ONTARGET) [39].
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
All authors contributed to the conduct of the study. GR conceived the study and prepared the manuscript.
==== Refs
Organization WH WHO Definition, diagnosis and classification of diabetes mellitus and its complications. Report of a WHO consultation. Part 1: Diagnosis and classification of diabetes mellitus 1999 World Health Organization 1 28
Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) final report Circulation 2002 106 3143 3421 12485966
Balkau B Vernay M Mhamdi L Novak M Arondel D Vol S Tichet J Eschwege E The incidence and persistence of the NCEP (National Cholesterol Education Program) metabolic syndrome. The French D.E.S.I.R. study Diabetes Metab 2003 29 526 532 14631330
Ford ES Giles WH A comparison of the prevalence of the metabolic syndrome using two proposed definitions Diabetes Care 2003 26 575 581 12610004
Lieberman LS Dietary, evolutionary, and modernizing influences on the prevalence of type 2 diabetes Annu Rev Nutr 2003 23 345 377 12651966 10.1146/annurev.nutr.23.011702.073212
Isomaa B Almgren P Tuomi T Forsen B Lahti K Nissen M Taskinen MR Groop L Cardiovascular morbidity and mortality associated with the metabolic syndrome Diabetes Care 2001 24 683 689 11315831
Lakka HM Laaksonen DE Lakka TA Niskanen LK Kumpusalo E Tuomilehto J Salonen JT The metabolic syndrome and total and cardiovascular disease mortality in middle-aged men JAMA 2002 288 2709 2716 12460094 10.1001/jama.288.21.2709
Sattar N Gaw A Scherbakova O Ford I O'Reilly DS Haffner SM Isles C Macfarlane PW Packard CJ Cobbe SM Shepherd J Metabolic syndrome with and without C-reactive protein as a predictor of coronary heart disease and diabetes in the West of Scotland Coronary Prevention Study Circulation 2003 108 414 419 12860911 10.1161/01.CIR.0000080897.52664.94
Sierra C Ruilope LM New-onset diabetes and antihypertensive therapy: comments on ALLHAT trial J Renin Angiotensin Aldosterone Syst 2003 4 169 170 14608521
Benson SC Pershadsingh HA Ho CI Chittiboyina A Desai P Pravenec M Qi N Wang J Avery MA Kurtz TW Identification of telmisartan as a unique angiotensin II receptor antagonist with selective PPARg-modulating activity Hypertension 2004 43 993 1002 15007034 10.1161/01.HYP.0000123072.34629.57
Fujimoto M Masuzaki H Tanaka T Yasue S Tomita T Okazawa K Fujikura J Chusho H Ebihara K Hayashi T Hosoda K Nakao K An angiotensin II AT1 receptor antagonist, telmisartan augments glucose uptake and GLUT4 protein expression in 3T3-L1 adipocytes FEBS Lett 2004 576 492 497 15498586 10.1016/j.febslet.2004.09.027
Schupp M Janke J Clasen R Unger T Kintscher U Angiotensin type 1 receptor blockers induce peroxisome proliferator-activated receptor-g activity Circulation 2004 109 2054 2057 15117841 10.1161/01.CIR.0000127955.36250.65
Matthews DR Hosker JP Rudenski AS Naylor BA Treacher DF Turner RC Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man Diabetologia 1985 28 412 419 3899825 10.1007/BF00280883
Parati G Omboni S Rizzoni D Agabiti-Rosei E Mancia G The smoothness index: a new, reproducible and clinically relevant measure of the homogeneity of the blood pressure reduction with treatment for hypertension J Hypertens 1998 16 1685 1691 9856370 10.1097/00004872-199816110-00016
Dahlöf B Devereux RB Kjeldsen SE Julius S Beevers G de Faire U Fyhrquist F Ibsen H Kristiansson K Lederballe-Pedersen O Lindholm LH Nieminen MS Omvik P Oparil S Wedel H group LIFE Cardiovascular morbidity and mortality in the Losartan Intervention For Endpoint reduction in hypertension study (LIFE): a randomised trial against atenolol Lancet 2002 359 995 1003 11937178 10.1016/S0140-6736(02)08089-3
Gress TW Nieto FJ Shahar E Wofford MR Brancati FL Hypertension and antihypertensive therapy as risk factors for type 2 diabetes mellitus. Atherosclerosis Risk in Communities Study N Engl J Med 2000 342 905 912 10738048 10.1056/NEJM200003303421301
Sowers JR Bakris GL Antihypertensive therapy and the risk of type 2 diabetes mellitus N Engl J Med 2000 342 969 970 10738057 10.1056/NEJM200003303421310
Neutel J Smith DH Evaluation of angiotensin II receptor blockers for 24-hour blood pressure control: meta-analysis of a clinical database J Clin Hypertens (Greenwich) 2003 5 58 63 12556655
Pershadsingh HA Kurtz TW Insulin-sensitizing effects of telmisartan: implications for treating insulin-resistant hypertension and cardiovascular disease Diabetes Care 2004 27 1015 15047668
Derosa G Ragonesi PD Mugellini A Ciccarelli L Fogari R Effects of telmisartan compared with eprosartan on blood pressure control, glucose metabolism and lipid profile in hypertensive, type 2 diabetic patients: a randomized, double-blind, placebo-controlled 12-month study Hypertens Res 2004 27 457 464 15302981 10.1291/hypres.27.457
Derosa G Cicero AFG Bertone G Piccinni MN Fogari E Ciccarelli L Fogari R Comparison of the effects of telmisartan and nifedipine gastrointestinal therapeutic system on blood pressure control, glucose metabolism, and the lipid profile in patients with type 2 diabetes mellitus and mild hypertension: A 12-month, randomized, double-blind study Clin Ther 2004 26 1228 1236 15476904 10.1016/S0149-2918(04)80049-3
Michel MC Bohner H Koster J Schafers R Heemann U Safety of telmisartan in patients with arterial hypertension. An open-label observational study Drug Saf 2004 27 335 344 15061687
The Expert Committee on the Diagnosis and Classification of Diabetes Mellitus Follow-up report on the diagnosis of diabetes mellitus Diabetes Care 2003 26 3160 3167 14578255
Association AD American Diabetes Association: clinical practice recommendations 2004 Diabetes Care 2004 27(Suppl 1) S1 S143
Tuomilehto J Point: a glucose tolerance test is important for clinical practice Diabetes Care 2002 25 1880 1882 12351496
Sacks DB Bruns DE Goldstein DE Maclaren NK McDonald JM Parrott M Guidelines and recommendations for laboratory analysis in the diagnosis and management of diabetes mellitus Clin Chem 2002 48 436 472 11861436
Woerle HJ Pimenta WP Meyer C Gosmanov NR Szoke E Szombathy T Mitrakou A Gerich JE Diagnostic and therapeutic implications of relationships between fasting, 2-hour postchallenge plasma glucose and hemoglobin a1c values Arch Intern Med 2004 164 1627 1632 15302632 10.1001/archinte.164.15.1627
Monnier L Lapinski H Colette C Contributions of fasting and postprandial plasma glucose increments to the overall diurnal hyperglycemia of type 2 diabetic patients: variations with increasing levels of HbA(1c) Diabetes Care 2003 26 881 885 12610053
Wallace TM Levy JC Matthews DR Use and abuse of HOMA modeling Diabetes Care 2004 27 1487 1495 15161807
Bonora E Targher G Alberiche M Bonadonna RC Saggiani F Zenere MB Monauni T Muggeo M Homeostasis model assessment closely mirrors the glucose clamp technique in the assessment of insulin sensitivity: studies in subjects with various degrees of glucose tolerance and insulin sensitivity Diabetes Care 2000 23 57 63 10857969
Haffner SM Kennedy E Gonzalez C Stern MP Miettinen H A prospective analysis of the HOMA model. The Mexico City Diabetes Study Diabetes Care 1996 19 1138 1141 8886564
Laakso M Karjalainen L Lempiainen-Kuosa P Effects of losartan on insulin sensitivity in hypertensive subjects Hypertension 1996 28 392 396 8794822
Raji A Seely EW Bekins SA Williams GH Simonson DC Rosiglitazone improves insulin sensitivity and lowers blood pressure in hypertensive patients Diabetes Care 2003 26 172 178 12502676
Wang TD Chen WJ Lin JW Chen MF Lee YT Effects of rosiglitazone on endothelial function, C-reactive protein, and components of the metabolic syndrome in nondiabetic patients with the metabolic syndrome Am J Cardiol 2004 93 362 365 14759393 10.1016/j.amjcard.2003.10.022
Kurtz TW Pravenec M Antidiabetic mechanisms of ACE inhibitors and AII receptor antagonists: beyond the renin-angiotensin system J Hypertens 2004 22 1 9 10.1097/00004872-200412000-00003
Nesto RW Bell D Bonow RO Fonseca V Grundy SM Horton ES Le Winter M Porte D Semenkovich CF Smith S Young LH Kahn R Thiazolidinedione use, fluid retention, and congestive heart failure: a consensus statement from the American Heart Association and American Diabetes Association Diabetes Care 2004 27 256 263 14693998
Mancia G Parati G Importance of smooth and sustained blood pressure control in preventing cardiovascular morbidity and mortality Blood Press Suppl 2001 3 26 32 11683474
Teo K Yusuf S Anderson C Mookadam F Ramos B Hilbrich L Pogue J Schumacher H Rationale, design, and baseline characteristics of 2 large, simple, randomized trials evaluating telmisartan, ramipril, and their combination in high-risk patients: the Ongoing Telmisartan Alone and in Combination with Ramipril Global Endpoint Trial/Telmisartan Randomized Assessment Study in ACE Intolerant Subjects with Cardiovascular Disease (ONTARGET/TRANSCEND) trials Am Heart J 2004 148 52 61 15215792 10.1016/j.ahj.2004.03.020
| 15892894 | PMC1174877 | CC BY | 2021-01-04 16:25:04 | no | Cardiovasc Diabetol. 2005 May 15; 4:6 | utf-8 | Cardiovasc Diabetol | 2,005 | 10.1186/1475-2840-4-6 | oa_comm |
==== Front
Harm Reduct JHarm Reduction Journal1477-7517BioMed Central London 1477-7517-2-81598751110.1186/1477-7517-2-8ReviewThe evolutionary origins and significance of drug addiction Saah Tammy [email protected] Stanford University School of Medicine, Transplant Immunobiology Laboratory, USA2005 29 6 2005 2 8 8 9 8 2004 29 6 2005 Copyright © 2005 Saah; licensee BioMed Central Ltd.2005Saah; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
By looking at drug addiction from an evolutionary perspective, we may understand its underlying significance and evaluate its three-fold nature: biology, psychology, and social influences. In this investigation it is important to delve into the co-evolution of mammalian brains and ancient psychotropic plants. Gaining an understanding of the implications of ancient psychotropic substance use in altering mammalian brains will assist in assessing the causes and effects of addiction in a modern-day context. By exploring addiction in this manner, we may move towards more effective treatment early prevention, treating the root of the issue rather than the symptoms.
==== Body
1. Introduction
As we find ourselves in the beginning of a new millennium, we are faced with challenges to our survival as a human population. Some of the greatest threats to our survival are sweeping epidemics that affect millions of individuals worldwide. Drug addiction, although often regarded as a personality disorder, may also be seen as a worldwide epidemic with evolutionary genetic, physiological, and environmental influences controlling this behavior. Globally, the use of drugs has reached all-time highs. On average, drug popularity differs from nation to nation. The United Nations Office on Drugs and Crime identified major problem drugs on each continent by analyzing treatment demand [1]. From 1998 to 2002, Asia, Europe, and Australia showed major problems with opiate addiction, South America predominantly was affected by cocaine addiction, and Africans were treated most often for the addiction to cannabis. Only in North America was drug addiction distributed relatively evenly between the use of opiates, cannabis, cocaine, amphetamines, and other narcotics. However, all types of drugs are consumed throughout each continent. Interpol reported over 4000 tons of cannabis were seized in 1999, up 20% from 1998, with the largest seizures made in Southern Africa, the US, Mexico, and Western Europe [2]. Almost 150 tons of cocaine is purchased each year throughout Europe and in 1999 opium production reached an estimated 6600 tons, the dramatic increase most likely due to a burst of poppy crops throughout Southwest Asia. This rapid increase in drug use has had tremendous global effects, and the World Health Organization cited almost 200,000 drug-induced deaths alone in the year 2000 [3]. The Lewin group for the National Institute on Drug Abuse and the National Institute on Alcohol Abuse and Alcoholism estimated the total economic cost of problematic use of alcohol and drugs in the United States to be $245.7 billion for the year 1992, of which $97.7 billion was due to drug abuse [4]. The White House Office of National Drug Control Policy (ONDCP) found that between 1988 and 1995, Americans spent $57.3 billion on drugs, of which $38 billion was on cocaine, $9.6 billion was on heroin and $7 billion was on marijuana.
Among the different approaches for diagnosis, prevention, and treatment of drug addiction, exploring the evolutionary basis of addiction would provide us with better understanding since evolution, personality, behavior and drug abuse are tightly interlinked. It is our duty as scientists to explore the evolutionary basis and origins of drug addiction so as to uncover the underlying causes rather than continuing to solely focus on the physiological signs and global activity of this epidemic. Too often the treatment of addiction simply works to alleviate the symptoms of addiction, dealing with overcoming the physiological dependence and working through withdrawal symptoms as the body readjusts to a non-dependent state of homeostasis. However, we must not only concentrate on this aspect of addiction when considering global treatments and preventative programs. We must take into consideration that it is not purely the physiology of addiction we are battling.
Drug addiction is thought of as an adjunctive behavior, or a subordinate behavior catalyzed by deeper, more significant psychological and biological stimuli. It is not just a pharmacological reaction to a chemical but a mode of compensation for a decrease in Darwinian fitness [5]. There are three main components involved in substance addiction: developmental attachment, pharmacological mechanism, and social phylogeny including social inequality, dominance, and social dependence [6]. Developmental attachment created by environmental influences, such as parental care or lack thereof, may influence children's vulnerability to drug addiction. Evolutionarily speaking, children that receive care that is more erratic may focus more so on short-term risks that may have proved to be an adaptive quality for survival in ancient environments. Compounding that attachment, the pharmacological mechanism describes the concept of biological adaptation of the mesolimbic dopamine system to endogenous substance intake. These factors combined with the influence of social phylogeny create a position for predisposition to drug addiction. They attribute to the common belief that many substances of abuse have great powers to heal, and that is often the driving motivation for overuse and addiction. Evolutionary perspective shows an intermediate and fleeting expected gain associated with drug addiction correlated with the conservation in most mammals of archaic neural circuitry [7], most often being a falsified sense of increased fitness and viability related to the three components of drug abuse [5,8]. The chemical changes associated with fitness and viability are perceived by mammals as emotions, driving human behavior.
Human behavior is mediated primarily by dopaminergic and serotonergic systems, both of ancient origins probably evolving before the phylogenetic splits of vertebrates and invertebrates [9]. 5-HT (serotonin), stimulated by a small range of drugs, mediates arousal. It is believed to be inhibited by hallucinogens and also helps control wanting for ethanol and cocaine consumption. The cortico-mesolimbic dopaminergic system, on the other hand, is believed to be the target of a wide range of drugs, including marijuana and cocaine, increasing the transmission of dopamine to the nucleus accumbens [10]. This system mediates emotion and controls reinforcement, and is the primary pathway acted on by antipsychotic drugs such as chlorprothixene and thioridazine. Problematic use of drugs develops into addiction as the brain becomes dependent on the chemical neural homeostatic circuitry altered by the drug [7]. No matter the theory of drug addiction, there remains one constant: withdrawal is inevitable. As a drug is administered continuously and an individual becomes addicted, the brain becomes dependent on the presence of the drug. With an absence of the drug, withdrawal symptoms are experienced as the brain attempts to deal with the chemical changes. There are believed to be evolutionary origins of drug addiction, which will be discussed further, as well as a link between physiological addiction and the evolution of emotion.
2. Drugs distribution and use in ancient environments
When examining the distribution of natural drugs in ancestral environment we see that there was often a limited amount of resources, meaning there was little overactivity of salient (wanting) behavior, causing no need for the adaptive development within the cortico-mesolimbic dopaminergic system of a built-in regulatory system of salience [6,11]. Genetic and environmental factors increasing substance abuse liability may have been of no consequence in ancestral environments due to their limitations. We originally relied on the limitations of ancient environments in that same manner, so when we are introduced to excessive amounts of salience in modern environment, we have no internal control. Basically, our ancient-wired bodies have not yet evolved to adapt to modern environment, leaving us vulnerable to addiction.
A common belief is that psychotropic plant chemicals evolved recurrently throughout evolutionary history [12]. Archaeological records indicate the presence of psychotropic plants and drug use in ancient civilizations as far back as early hominid species about 200 million years ago. Roughly 13,000 years ago, the inhabitants of Timor commonly used betel nut (Areca catechu), as did those in Thailand around 10,700 years ago. At the beginning of European colonialism, and perhaps for 40,000 years before that, Australian aborigines used nicotine from two different indigenous sources: pituri plant (Duboisia hopwoodii) and Nicotiana gossel. North and South Americans also used nicotine from their indigenous plants N. tabacum and N. rustica. Ethiopians and northern Africans were documented as having used an ephedrine-analog, khat (Catha edulis), before European colonization. Cocaine (Erythroxylum coca) was taken by Ecuadorians about 5,000 years ago and by the indigenous people of the western Andes almost 7,000 years ago. The substances were popularly administered through the buccal cavity within the cheek. Nicotine, cocaine, and ephedrine sources were first mixed with an alkali substance, most often wood or lime ash, creating a free base to facilitate diffusion of the drug into the blood stream. Alkali paraphernalia have been found throughout these regions and documented within the archaeological record. Although the buccal method is believed to be most standard method of drug administration, inhabitants of the Americas may have also administered substances nasally, rectally, and by smoking.
Many indigenous civilizations displayed a view of psychotropic plants as food sources, not as external chemicals altering internal homeostasis [12]. The perceived effects by these groups were tolerance to thermal fluctuations, increased energy, and decreased fatigue, all advantageous to fitness by allowing longer foraging session as well as greater ability to sustain in times of limited resources. The plants were used as nutritional sources providing vitamins, minerals, and proteins rather than recreational psychotropic substances inducing inebriation. Due to limited resources within ancient environments, mammalian species most probably sought out CNS neurotransmitter (NT) substitutes in the form of psychotropic allelochemicals, because nutrient NT-precursors were not largely available in the forms of food. Therefore, drugs became food sources to prevent decreased fitness from starvation and death. It is believed that early hominid species evolved in conjunction with the psychotropic flora due to constant exposure with one another. This may be what eventually allowed the above civilizations to use the flora as nutritional substances, therefore increasing both their fitness and viability.
Over time, psychotropic plants evolved to emit allelochemical reactivity to deter threats from herbivores and pathogenic invasions. These allelochemical responses evolved to imitate mammalian NT so as to act as competitive binders and obstruct normal CNS functioning. The allelochemical NT analogs were not anciently as potent as forms of abused substances used in modern environments, but instead were milder precursors that had an impact on the development of the mammalian CNS. The fit of allelochemicals within the CNS indicates some co-evolutionary activity between mammalian brains and psychotropic plants, meaning they interacted ecologically and therefore responded to one another evolutionarily. Basically, series of changes occurred between the mammalian brain and psychotropic plants allowing them affect one another during their processes of evolving. This would have only been possible with mammalian CNS exposure to these allelochemicals, therefore to ancient mammalian psychotropic substance use. The evidence for this theory is compelling. For example, the mammalian brain has evolved receptor systems for plant substances, such as the opioid receptor system, not available by the mammalian body itself. The mammalian body has also evolved to develop defenses against overtoxicity, such as exogenous substance metabolism and vomiting reflexes.
3. Evolutionary advantage of emotion
The evolution of brain systems brought about indicators of levels of fitness in the form of chemical signals perceived as emotion [7,8,11,13]. These emotions help direct physiology and behavior of an individual towards increasing Darwinian fitness. They essentially were tools chosen for by the mechanisms of natural selection. Positive emotions, such as euphoria and excitation, motivate towards increased gain and fitness state, whereas negative emotions, for instance anxiety and pain, evolved as defenses by motivating towards managing potential threats or decreases in fitness level.
Mammalian drive to escape danger is fueled by a capacity to feel negative emotions [14]. Negative emotions can be defenses, and in their suppression we may find ourselves unarmed and unprepared to deal with problems much more detrimental than the original warning emotions. Those individuals that lack the capacity to suffer, including the inability to experience pain, are unable to put up basic physiological and behavioral defenses and often find themselves dying at relatively young ages. Negative emotions (pain, fear, stress, anxiety, etc.) have evolved in mammals to allude to even the slightest, most harmless potential indicator of a more serious problem, leading to what may be known as a modern-day personality disorder. Personality disorders can be characterized as anything from over-anxiety to schizophrenia [13]. Many emotional disorders that drugs mask, such as anxiety disorders, develop from the ancient adaptive mechanisms expressed by the evolved mode of personality, and may in fact not be disorders but hypersensitive neural adaptations. Since personality evolved as an information gating mechanism to transmit culture among people, as well as within an individual from external environmental stimuli to internal neural circuitry for personal regulation, negative emotion may be simply transmitted and can be enhanced through personality [15].
There are two defined types of positive emotion [7]. The first includes feelings of anticipation and excitation induced by a promise of an increase in fitness (+ Positive Affect, or PA), while the second includes emotions of relief and security due to a removal of a threat to fitness (- Negative Affect or NA). + PA emotions fall into the behavioral activation system, or the BAS [16]. The BAS attempts to propagate positive emotions and appetitive conditioning, resulting in a motivation to reach goals and, essentially, the positive affect. – NA emotions fall into the behavioral inhibition system (BIS), which attempts to regulate and compensate for negative emotions and aversive conditioning. As mammals expose themselves to fitness-increasing situations and avoid fitness-decreasing situations, they tend to motivate towards pleasure-inducing, or + PA, stimuli that indicate these increases in fitness. Even if unrelated to fitness in modern environment, emotions continue to process events in the same archaic way. Many pleasant feelings may now not indicate an increase in fitness at all, but the evolutionary brain may still correlate the two.
Modern environments include medical and social technologies that bring comfort and longer living than was experienced in ancient environment, so much of modern human emotion does not serve the same function as was evolutionarily performed. As our emotions become less indicative of fitness and more superfluous, there comes to be confusion within the intended signals of emotion. The pursuit of "happiness" involves gain, and while evolutionarily these gains were increased fitness, the emotion of happiness is no longer directly related to fitness. While one may become happy due to a casual and pleasing relationship, the euphoric emotion may have evolutionarily corresponded with an indication of successful reproduction and therefore a gain in fitness and viability. This can also be applied to the euphoria associated with wealth, which in ancient environments may have been an indicator of increases in fitness due to plentiful food and water resources, but now may indicate status.
4. Effects of drugs on emotion
Psychoactive drugs induce emotions that at one point in mammalian evolutionary history signaled increased fitness, not happiness [11]. In ancient environments positive emotion correlated with a sign of increased fitness, such as successful foraging sessions or successful breeding. Mammals would feel euphoric only during times where fitness levels were high, the euphoria being indicative of survival and not a superfluous feeling of "happiness." Mammals would otherwise feel negative emotions when fitness levels were low. The effect of many psychoactive substances provided the same euphoric feeling, and may have had some increasing effects on fitness levels in ancient mammalian species. However, drug use today does not carry the same predicted increases in fitness, and in fact may act as a pathogen on neural circuitry. Yet, these same drugs continue to target archaic mechanisms of the brain with the intent of inducing positive emotion, essentially blocking many neurological defenses.
Drugs that stimulate positive emotion virtually mediate incentive motivation in the nucleus accumbens and the neural reward system [11]. Modern drug addiction fundamentally indicates a false increase of fitness, leading to increasing drug abuse to continue gain, even if the gain is realized as being false. This is the quintessential paradox among drug addicts. The motivation towards gain begins to take precedence over adaptive behaviors among addicted individuals. Some stimuli that simulate increased fitness may become greater priorities than true adaptive stimuli necessary for increased fitness, such as food and sleep [7]. Individuals can, in turn, decrease their fitness by ignoring necessary behaviors for survival and fitness and focusing on a false positive emotion. The appetite for a drug may also override the drive to consummate, causing a drastic decrease in viability. Their emotional systems are now concentrated on drug-seeking rather than survival.
In modern humans, drugs that may block negative emotions may be more useful than the endurance of ancient warnings of harm, like pain and fever [11]. Certain drugs can aid in pathology treatment, and while negative emotions may have been entirely necessary for the survival of ancient mammals, they may no longer be exclusively indicative of nociceptive or otherwise harmful stimuli [11,13]. Hypersensitivity of our bodies' defense mechanisms has evolved, leading to unnecessary negative emotions for non-nociceptive stimuli as preventative defense. When there is a threat towards an individual's fitness, the modern body often responds with several different warning signs, perhaps several different types of negative emotions (pain, fever, and hallucination, for example). Therefore, blocking a few of the negative emotions will ideally not disrupt the message. I emphasize the word "ideally" for this is not always the case. Frequently there are situations in which drugs that block these defenses, such as anxiolytics, may contribute to the decreases in fitness by temporarily removing a small negative emotions but leaving the individual vulnerable to a much larger harm [17].
Emotional disposition has shown to specifically correlate with problematic use of alcohol [16]. If the perceived emotion before alcohol consumption is negative, the individual most likely is drinking to cope, with less control over his/her own use. In the case of a positive disposition before consumption, the user is said to drink to enhance, with more greatly controlled use of the substance. Since alcohol consumption alters normally functioning cognitive processes, it does not prove to be equal to evolutionarily superior internal coping mechanisms. Instead, alcohol mediates not only negative feelings by their suppression, but also encourages the habituated continuance of positive emotion. Recovering alcoholics often document reasons of relapse surrounding the drive to compensate for negative feelings, resulting in a motivation to cope and therefore to drink.
5. Physiology of addiction and reward
Mammalian brains work heavily on a motivational system with two types of motivation: like and want [11]. Like is controlled by opioid and brain stem systems, and refers to pleasure upon receiving a reward, whereas want (salience), mediated by the cortico-mesolimbic dopaminergic system, is an anticipatory motivation to pursue reward. We receive "pleasure" through intracellular signaling of adaptive chemical pathways of a reward system that bring our attention to what we need. The nucleus accumbens (NAcb) and globus pallidus are involved in reward pathways for alcohol, opiates, and cocaine [18]. NTs involved in these pathways are dopamine (primarily within the NAcb and hippocampus), serotonin (hypothalamus), enkephalins (ventral tegmental area and NAcb), GABA (inhibitory – ventral tegmental area and NAcb), and norepinephrine (hippocampus). When there is a disturbance within the reward intracellular cascade, a chemical imbalance occurs that triggers negative emotions to be indicative of the disturbance. This is referred to as "reward deficiency syndrome," where the chemical imbalances within the intracellular cascade manifest themselves as behavioral disorders, indicating a deficiency within the adaptive reward pathway. Drug addiction may initially cause and then further proceed to exacerbate "reward deficiency syndrome."
Another theory of drug addiction, the "drugs for reward" theory, states that addiction is the malfunctioning collision of both motivational systems (like vs. want), stimulating pursuit of a substance that most probably no longer provides pleasure and in fact may be pathogenic [11]. Different drugs stimulate different types of positive emotion [7]. Opioids contribute to – NA states, while dopamine-releasing drugs contributes to + PA states. In this theory, dopamine is believed to mediate a state of addiction through the activation of the cortico-mesolimbic system passing through the ventral tegmental area to the nucleus accumbens, all regulating reward-seeking motivation. It is also involved in withdrawal from psychostimulants, as the sudden removal of a chemical drug stimulant from the body causes a massive alteration within the dopaminergic system, leading to negative emotions. Opioids are believed to mediate the consumption of reward, with opioid addiction following a well-defined route: 1) first ensues as a pleasure-seeking behavior, 2) tolerance to the opioid builds and pleasure resulting from drug use reduces, yet use is increased in an attempt to regain the hedonic pleasure, and 3) withdrawal may occur with a cessation of the opioid substance differing from withdrawal from psychostimulants, but also leading to negative emotions. With the "drugs for reward" theory, adaptive hard-wired (physiologically determined to serve a specific role) dopamine function is believed to induce a feeling of reward for a particular action that indicates an increase in the level of fitness of an individual [6]. It encourages the continuation of habit that increases dopamine release, therefore leading to a perception of increased levels of fitness (although often falsely when referring to drug use). Problems with this theory are encountered when we take into consideration that dopamine also signals negative reinforcement, not just positive reinforcement through reward. Dopamine is therefore referred to as simply altering an emotional state from one to another, even if it means going from positive emotion to negative emotion.
Dopamine is otherwise argued to be a mediator of salience [6]. Although dopamine is believed to control the cortico-mesolimbic system, it does not rule the consummatory/satiatory/seeking behavior in this particular theory. It instead mediates appetitive/approach behavior, placing an importance on things by demanding attention on either their strength (positive emotion) or their potential harm (negative emotion), then increasing the motivation to move towards an action to change, not to satiate (stop). If upregulated, a feeling of "wanting" is induced for a specific substance, leading to addiction with overuse [10]. This explains dopamine action as integrated activity rather than hard-wired function, and best explains how drug addiction is obsessively saliatory without ever reaching satiation. This concept is referred to as IS, or incentive salience. Earlier theories discussed unconditioned stimuli, such as a specific drug, as stimulants of an unconditioned response of neural regulation [19]. In this model, the drug is not the unconditioned stimulus causing guaranteed changes of the CNS, as was previously thought, but the chemical activity caused by the drug within the CNS is the unconditioned stimuli. The brain then becomes adapted to the chemical response of the drug, producing a salient conditioning response within the brain's association context. The prefrontal cortex directs associative context, in turn regulating the cortico-mesolimbic dopamine system to induce an amalgamation of abnormal behavior and salience; the individual is now driven by uncontrolled craving and wanting. We originally relied on the limitations of the ancestral environment to be the regulatory influences as we used drugs for food, and our bodies still remain adapted to ancestral environments in that aspect. Therefore, when we are introduced to excessive amounts of salience, we have no internal control.
Candidate gene polymorphisms within the above pathway receptors may contribute to substance abuse [20]. Substance abuse tendencies and liabilities (the vulnerability to a disease and the possibility of becoming affected due to genetic and environmental susceptibility) may be inherited through phenotypic liabilities. The expression of substance abuse is therefore dependent on this phenotypic liability and environmental influences. The phenotypic liability may be a result of a genetic polymorphism within the DRD2 dopamine receptor gene (A.sub.1 allele) [18]. The DRD2 dopamine receptors are targeted by antipsychotics [9]. This particular receptor gene polymorphism correlates with alcohol and substance addiction as well as obsessive compulsive disorders. The DRD4 dopamine receptor has documented polymorphisms within a 48 base pair variable number tandem repeat, and also correlates with substance addiction, for it is believed to be involved in reducing sensitivity to methamphetamines, alcohol, and cocaine. In Israeli and Arab heroin-dependent populations, there was data collected displaying a DRD4 gene polymorphism in exon 3 consisting of seven-repeat alleles not present in non-addicted control groups. This was also observed in a study of heroin-addicted Han Chinese. In a study done with Native American alcoholics, a linkage on chromosome 11 near the DRD4 gene was documented. With these phenotypic liabilities, an individual may be considered to be addicted to a substance after passing a threshold of which there is no diagnostic or solid definition. Dependence is often continued because of temporary positive effects with the denial of the more permanent, negative pharmaceutical effects. There have been documented significant relationships between drug and alcohol dependence and certain genetic factors, with the same genetic correlation to smoking, displaying a significant cohesion between different substance use disorders. Individuals addicted to substances may, therefore, be genetically predisposed to the situation and are then pushed past threshold by environmental stimuli.
6. Social-cultural impact
We have discovered that the nature of addiction is not solely based on free will to use, or an individual's conscious choice to use, but may have deeper influences. The nature of drug addiction is three-fold: biological, psychological, and social. Although humans may be biologically and psychologically predisposed to drug use and addiction, they may often be driven towards that state by social and cultural influences. To what extent environmental stimuli affect a person's vulnerability to addiction is unknown and may be varying. However, we cannot ignore the great impact of environmental and mental stimuli in the progression towards addiction. It has been found that certain environmental variables breed higher vulnerability [21]. Family dysfunction and disruption, low social class rearing, poor parental monitoring, and rampant social drug-use exposure may greatly contribute to an individual's movement from substance abuse predisposition to addiction. Both acute and chronic stresses have been linked with substance abuse as well, with acute stress being one of the main influences of relapse in rehabilitated drug addicts. The widespread availability of drugs in certain areas also may affect susceptibility [22]. This is exceptionally notable in low socioeconomic areas in which overcrowding and poverty have been associated statistically with increased substance abuse. In addition, repeated exposure to successful high-status role models who use substances, whether these role models are figures in the media, peers or older siblings, is likely to influence children and adolescents. Similarly, the perception that smoking, drinking or drug use is standard practice among peers also serves to promote substance abuse.
When examining drug addiction through this triple-perspective, we are forced as a global society to re-evaluate the criminalization of drug use and addiction throughout world. In general, social drug policies have been conservative and unyielding. Most often, addicts are left to feed their addiction through illegal means of acquiring drugs. As a result of conservative influence in national politics, a "tough on drugs" philosophy that stresses zero tolerance, law enforcement, and abstinence has been adopted. This philosophy neglects the need for medical and psychological treatment of substance addiction.
Columbia University's National Center on Addiction and Substance Abuse report over 75% of state penitentiary inmates require drug abuse treatment, but the disconcerting fact is that under 20% of those individuals actually are provided with proper treatment programs [23]. If treatment is provided, it is often times extremely short-term and non-intensive, and even less frequently offered to jail inmates. In addition, the Bureau of Justice Statistics stated that only 1 in 10 state prison inmates were provided drug abuse treatment in 1997, down from the 1 in 4 inmates offered treatment in 1991. This is astonishingly low, considering the correctional institution holds more substance abusers than any other national institution. Also commonly noted are the incredibly high comorbidity rates between mental illness and drug addiction within the prison system. It is vital to view substance addiction as a medical condition when dealing with criminal charges, making sure that addicts are provided with treatment for the root of their affliction rather than simply punishing the active symptoms of addiction.
7. Conclusion
Drug use and addiction seem to have been a part of mammalian society since ancient times. Researchers have evidence and reason to believe that the evolution of mammalian brains and psychotropic plants might be related to each other, connected by ancient drug use. Regardless of the possible co-evolution of drugs and mammalian brains, abuse of drugs inevitably causes long-term disadvantages. Drug addiction could be extremely detrimental for any individual, not only because of the various health problems involved, but also due to the fact that it abolishes negative emotions, such as pain, which in turn shuts off basic defense mechanisms against potential threats. While the origins for drug addiction may indeed be genetically founded, abuse is most likely caused by a combination of both external and internal stimuli. Although a person may be pre-disposed to addiction, environmental and emotional stimuli may act as a catalyst towards the state of actual substance addiction. It is suggested that the motivation towards drug abuse comes from reward systems within the mammalian brain causing an initial "like" for a substance and leading to the insatiable "want" that correlates with abuse. Although there has been a distinction made between a possibility of a reward-based abuse and a salience-based abuse, it may be possible to see a combined effort of the two proposed systems working towards eventual drug addiction.
More research spanning the evolutionary history of mammalian brains might give us a greater awareness of the physiological wiring of the mammalian brain. For example, is there a combined influence of the salience and reward systems? Are these systems in fact hard-wired, indicating a hard-wired and possibly genetic underlying origin of liability to drug abuse? What is the true reason all humans are vulnerable to drug abuse? Are these tendencies towards drug abuse preventable or simply treatable? These and other questions may, in turn, allow us a deeper understanding of how to effectively prevent and treat drug abuse without simply placing a bandage over it by relieving the superficial symptoms accompanying it. Essentially, we must investigate what may universally cause this internal affliction before we can move on to examine external environmental stimuli that may be associated with individual cases.
==== Refs
United Nations Office on Drugs and Crime
Interpol
United Nations Information Office
National Institute on Drug Abuse
Falk JL Drug abuse as an adjunctive behavior Drug Alcohol Depend 1998 52 91 98 9800138 10.1016/S0376-8716(98)00084-2
Lende DH Smith EO Evolution meets biopsychosociality: an analysis of addictive behavior Addiction 2002 97 447 458 11964060 10.1046/j.1360-0443.2002.00022.x
Panksepp J Knutson B Burgdorf J The role of brain emotional systems in addictions: a neuro-evolutionary perspective and new 'self-report' animal model Addiction 2002 97 459 469 11964061 10.1046/j.1360-0443.2002.00025.x
Buss DM The evolution of happiness Am Psychol 2000 55 15 23 11392858 10.1037//0003-066X.55.1.15
Cravchik A Goldman D Neurochemical individuality: genetic diversity among human dopamine and serotonin receptors and transporters Arch Gen Psychiat 2000 57 1105 1117 11115324 10.1001/archpsyc.57.12.1105
Everitt BJ Dickinson A Robbins TW The neuropsychological basis of addictive behaviour Brain Res Rev 2001 36 129 138 11690609 10.1016/S0165-0173(01)00088-1
Nesse RM Psychoactive drug use in evolutionary perspective Science 1997 278 63 66 9311928 10.1126/science.278.5335.63
Sullivan RJ Hagen EH Psychotropic substance-seeking: evolutionary pathology or adaptation? Addiction 2002 97 389 400 11964056 10.1046/j.1360-0443.2002.00024.x
Nesse RM Emotional disorders in evolutionary perspective Br J Med Psychol 1998 71 397 415 9875953
Nesse RM Proximate and evolutionary studies of anxiety, stress, and depression: synergy at the interface Neurosci Biobehav Rev 1999 23 895 903 10580304 10.1016/S0149-7634(99)00023-8
Nash WP Information gating: an evolutionary model of personality function and dysfunction Psychiatry 1998 61 46 60 9595595
Cooper ML Frone MR Russell M Mudar P Drinking to regulate positive and negative emotions: a motivational model of alcohol use J Pers Soc Psychol 1995 69 990 1005 7473043 10.1037//0022-3514.69.5.990
Nesse RM An evolutionary perspective on substance abuse Ethol Sociobiol 1994 15 339 348 10.1016/0162-3095(94)90007-8
Blum K Cull JG Braverman ER Comings DE Reward deficiency syndrome Am Sci 1996 84 132 145
Woods SC Ramsay DS Pavlovian influences over food and drug intake Behav Brain Res 2000 110 175 182 10802313 10.1016/S0166-4328(99)00194-1
Vanyukov MM Tarter RE Genetic studies of substance abuse Drug Alcohol Depend 2000 59 101 123 10891624 10.1016/S0376-8716(99)00109-X
Volkow ND The reality of comorbidity: depression and drug abuse Biol Psychiatry 2004 56 714 717 15556111 10.1016/j.biopsych.2004.07.007
Hawkins JD Catalano R Risk and Protective Factors for Alcohol and Other Drug Problems in Adolescence and Early Adulthood: Implications for Substance Use Prevention Psych Bulletin 1992 112 64 105 10.1037//0033-2909.112.1.64
Freudenberg N Jails, Prisons, and the Health of Urban Populations: A Review of the Impact of the Correctional System on Community Health J Urban Health 2001 78 214 235 11419576
| 15987511 | PMC1174878 | CC BY | 2021-01-04 16:36:49 | no | Harm Reduct J. 2005 Jun 29; 2:8 | utf-8 | Harm Reduct J | 2,005 | 10.1186/1477-7517-2-8 | oa_comm |
==== Front
Virol JVirology Journal1743-422XBioMed Central London 1743-422X-2-531598518210.1186/1743-422X-2-53ResearchActively replicating West Nile virus is resistant to cytoplasmic delivery of siRNA Geiss Brian J [email protected] Theodore C [email protected] Michael S [email protected] Departments of Medicine, Washington University School of Medicine, 660 South Euclid Avenue, Box 8051, St. Louis, MO 63110, USA2 Molecular Microbiology, Washington University School of Medicine, 660 South Euclid Avenue, Box 8051, St. Louis, MO 63110, USA3 Pathology & Immunology, Washington University School of Medicine, 660 South Euclid Avenue, Box 8051, St. Louis, MO 63110, USA4 Department of Microbiology, University of Pennsylvania, Philadelphia, PA, 19104, USA2005 28 6 2005 2 53 53 28 5 2005 28 6 2005 Copyright © 2005 Geiss et al; licensee BioMed Central Ltd.2005Geiss et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
West Nile virus is an emerging human pathogen for which specific antiviral therapy has not been developed. Recent studies have suggested that RNA interference (RNAi) has therapeutic potential as a sequence specific inhibitor of viral infection. Here, we examine the ability of exogenous small interfering RNAs (siRNAs) to block the replication of West Nile virus in human cells.
Results
WNV replication and infection was greatly reduced when siRNA were introduced by cytoplasmic-targeted transfection prior to but not after the establishment of viral replication. WNV appeared to evade rather than actively block the RNAi machinery, as sequence-specific reduction in protein expression of a heterologous transgene was still observed in WNV-infected cells. However, sequence-specific decreases in WNV RNA were observed in cells undergoing active viral replication when siRNA was transfected by an alternate method, electroporation.
Conclusion
Our results suggest that actively replicating WNV RNA may not be exposed to the cytoplasmic RNAi machinery. Thus, conventional lipid-based siRNA delivery systems may not be adequate for therapy against enveloped RNA viruses that replicate in specialized membrane compartments.
==== Body
Background
West Nile virus (WNV) is a significant human and veterinary mosquito-borne pathogen that has rapidly spread across North America. Humans develop a febrile illness and a small subset progress to meningitis or encephalitis syndromes [1]. Currently, no specific therapy or vaccine has been approved for treatment or prophylaxis of WNV infection in humans.
WNV is an enveloped virus with an 11-kilobase positive strand RNA genome. It is translated directly from the genomic RNA as a single polyprotein and cleaved by cellular and viral proteases into ten mature proteins, three structural (C, M, and E) and seven non-structural (NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5) proteins [2,3]. Virus entry occurs by endocytosis after the E protein interacts with cellular receptor(s). Genomic viral RNA traffics to the endoplasmic reticulum (ER), where WNV protein translation and RNA replication occur [4]. The positive strand genomic WNV RNA that associates with the ER is competent for translation and transcription of negative strand RNA. WNV and related flaviviruses induce ER membrane proliferation and reorganization, and replicating viral RNA has been observed at these membranous structures [5-7]. Disruption of WNV protein translation and/or RNA replication blocks the viral lifecycle and aborts infection.
RNA interference (RNAi) is a cellular process that specifically degrades RNA within the cytoplasm of cells in a sequence-specific manner [8]. RNAi occurs in plants [9], nematodes [10], parasites [11,12], insects [13], and mammalian cells [14,15] and is believed to function as a regulator of cellular gene expression and possibly as an innate defense against RNA viruses [16]. RNAi uses double stranded RNA (dsRNA) to target and degrade sequence-specific single-stranded RNA. The cytoplasmic ribonuclease DICER recognizes and cleaves long dsRNA molecules into 21 to 30 base pair small interfering RNA (siRNA) molecules; these associate with the RNA Induced Silencing Complex (RISC) to target and degrade complementary single-stranded RNA molecules [8].
RNAi has been used as a method to transiently disrupt various gene products to study their function [14,15,17-20]. Many mammalian viruses appear susceptible to treatment with exogenous siRNA. Cells that express virus-specific siRNA are resistant to infection by WNV [21], poliovirus [22,23], influenza A [21,24], HIV [25] and hepatitis C [26,27]in vitro. Administration of siRNAs in vivo has modestly reduced hepatitis B antigen production [28,29] and influenza A virus infection [30,31]. The sequence specific activity of siRNA against viruses has led to great interest in its potential as a new class of antiviral therapy. Nonetheless, there may be limitations with this approach as in vivo RNAi has not been demonstrated as effective as post-exposure therapy.
Previously, we demonstrated that transgenic expression of a sequence-specific siRNA prior to infection could efficiently inhibit WNV replication [21]. However, for a WNV-specific siRNA to be effective as a post-exposure therapeutic, it would need to inhibit infection in cells that are actively replicating WNV RNA. In this study, we evaluated the efficacy of siRNA against WNV that has already initiated active replication. Although cytoplasm-directed transfection of cells with siRNA prior to infection efficiently blocked WNV infection, administration after infection had little efficacy. Unlike plant viruses that encode active suppressors of RNA interference [32-34], WNV did not appear to actively inhibit the RNAi response, but rather avoided degradation by replicating in a manner that was inaccessible to the RNAi machinery.
Results
In vitro generated siRNA inhibits WNV infection in cells
We have previously demonstrated that plasmid expressed hairpin siRNA efficiently inhibited infection of WNV in mouse and human cell lines [21]. Because a therapeutic application of exogenously delivered rather than plasmid-expressed siRNA may be more clinically relevant, we assessed the inhibitory activity of in vitro transcribed hairpin siRNA against WNV infection. A 21-nucleotide region of the WNV capsid gene (nucleotides 312–332; Cap) was initially targeted, as this region is conserved among all WNV strains and lacks homology to known cellular genes. To demonstrate the specificity of Cap siRNA, a hairpin siRNA that targets the Influenza A M2 gene (nucleotides 18–38, M2) [21] and a mutated version of Cap siRNA (Cap Mut) that had 4 changes were also designed (Table 1). Our in vitro transcription strategy employed partially duplexed oligonucleotides containing a double stranded T7 promoter sequence (Fig 1A).
Table 1 Small interfering RNA
Name Virus Start Nucleotide Target Sequence
Cap WNV Lineage I 312 gaacaaacaaacagcgatgaa
Cap-Mut WNV Lineage I 312 gaagaaagaaagaccgatgaa
M2 Influenza A M2 18 ggtcgaaacgcctatcagaaa
3110 WNV Lineage I 3110 gggcagttctgggtgaagt
3317 WNV Lineage I 3317 ctacggtcaccctgagtga
4119 WNV Lineage I 4119 gaggagcaagtctgctatgc
4823 WNV Lineage I 4823 gtgtcaaggaggatcgact
5039 WNV Lineage I 5039 gacggtgatgtgattgggct
5497 WNV Lineage I 5497 gcagcaagaggttacattt
6337 WNV Lineage II 6337 gttgaagtcatcacgaagt
6349 WNV Lineage I 6349 gtggaagtcatcacgaagc
6915 WNV Lineage I 6915 caacgagatgggttggcta
7353 WNV Lineage I 7353 gaagaacgctgtagtggat
7693 WNV Lineage I 7693 ggacgcaccttgggagaggt
8892 WNV Lineage I 8892 ggtcaacagcaatgcagct
8898 WNV Lineage I 8898 cagcaatgcagctttgggt
9095 WNV Lineage I 9095 gaagcagagccatttggtt
9607 WNV Lineage I 9607 gggaaaggacccaaagtca
10355 WNV Lineage I 10355 gagagatatgaagacacaac
siRNA were generated against 19–21 nucleotide sequences corresponding to the target region of different parts of the New York 1999 WNV genome. Sequences were chosen using the SciTools RNAi design program and compared against the GenBank database to exclude sequences that may affect cellular genes.
Figure 1 WNV is susceptible to siRNA pretreatment. A. Scheme for generation of small hairpin siRNAs. B. Cap siRNA specifically inhibits WNV RNA accumulation. Huh7.5 cells were mock transfected or transfected with Cap siRNA or Cap Mut siRNA. Eighteen hours later cells were infected with WNV at MOI 0.1. Forty-eight hours later cells were collected and total RNA was recovered. WNV RNA was measured by quantitative real time RT-PCR. The results are an average of three independent experiments and the error bars indicate standard error of the mean. C. Capsid siRNA specifically inhibits WNV E protein expression. Huh7.5 cells were mock transfected or transfected with M2 siRNA, Cap siRNA, or Cap Mut siRNA as described above. Forty-eight hours after infection, cells were collected and processed for flow cytometry using anti-WNV envelope protein antibody E1. The results are one representative example of three independent experiments. D. Inhibitory activity of different WNV-specific siRNA. Huh7.5 cells transfected with the indicated siRNA, infected with WNV, and then analyzed for viral antigen as described in Materials and Methods. The fold inhibition was calculated after dividing the percentage of antigen positive cells from mock-transfected cells by the percentage of antigen positive cells from siRNA transfected cells. The results are one representative example of two independent experiments.
Human Huh7.5 cells were used because they were efficiently transfected with siRNA and infected with WNV. Huh-7.5 were transfected with Cap or Cap-Mut siRNA, infected with a New York strain of WNV at 18 hours after transfection, and analyzed 48 hours post-infection for levels of viral RNA by RT-PCR (Fig 1B). Pretreatment of Huh7.5 cells with Cap siRNA resulted in approximately 1 log reduction of WNV RNA, whereas pretreatment of Huh7.5 cells with Cap-Mut siRNA showed no significant reduction of WNV RNA. To confirm that RNAi also decreased WNV antigen production, siRNA-transfected Huh7.5 cells were examined for WNV envelope protein expression at 48 hours post infection. Approximately 70% of mock or TKO treated Huh-7.5 cells were positive for WNV antigen, levels comparable to that observed in cells transfected with either M2 (57% positive) or Cap-Mut (59% positive) siRNAs (Fig 1C). In contrast, less than 3% of Cap siRNA transfected cells stained positive for WNV E antigen. Thus, in vitro generated sequence-specific hairpin siRNA efficiently and specifically blocked WNV RNA and antigen production in mammalian cells.
To demonstrate that siRNAs targeting different regions of the WNV genome could inhibit infection, multiple siRNAs were designed spanning the nonstructural genes of WNV (Fig 1D). Two siRNAs (5497 and 6349) targeted to the nonstructural proteins reduced WNV envelope expression by at least 4-fold. Despite using an siRNA prediction algorithm, many of the siRNAs demonstrated little ability to inhibit envelope protein expression, possibly due to secondary structure in the WNV genomic RNA. Interestingly, treatment with combinations of siRNA did not show appreciably greater inhibition than treatment with either siRNA alone (data not shown).
Timing of siRNA treatment affects effectiveness against WNV
siRNA therapy in a clinical setting likely would require treatment after WNV infection has occurred. Because of this, we assessed the ability of siRNA to block WNV RNA before and after infection (Fig 2A). Huh7.5 cells were transfected with siRNAs 18 hours before infection or at 10 hours after infection. Cells were not transfected at very early times after infection because the TKO transfection reagent interfered with the ability of WNV virus to infect cells in a time-dependent manner (B. Geiss and M. Diamond, unpublished observation). By ~10 hours post-infection virus had entered cells and were no longer affected by the TKO reagent. Total RNA was harvested 48 hours post-infection and analyzed for genomic WNV RNA content. As expected, pretreatment of cells with Cap siRNA, but not Cap Mut siRNA, resulted in a 10-fold reduction in WNV RNA levels. Strikingly, addition of Cap siRNA 10 hours after infection resulted in no reduction of WNV RNA. The lack of inhibitory effect of RNAi at late times after infection was not due to the emergence of resistant mutants: sequence analysis of multiple viral isolates at 48 hours post-infection from cells that had been transfected with Cap or 6349 siRNA demonstrated no mutations in the targeted viral sequences (data not shown). Thus, WNV, in contrast to poliovirus [35], did not appear to mutate to evade siRNA-triggered degradation.
Figure 2 WNV becomes resistant to RNAi after infection. A. Huh7.5 cells were mock transfected or transfected with Cap or Cap Mut siRNA at the indicated times before or after WNV infection. Forty-eight hours after infection cells were harvested and WNV RNA levels were determined by quantitative real-time RT-PCR. The results are an average of three independent experiments and error bars indicate standard error of the mean. B. Induction of RNAi resistance by an attenuated lineage II WNV. Inset. Genomic structure of the attenuated lineage II WNV, which includes an IRES-controlled GFP insertion in the 3' UTR. 6337 denotes the target region of the lineage II specific siRNA. Attenuated WNV becomes resistant to siRNA after infection is established. Huh7.5 cells were transfected with Cap Mut, 6349, or 6337 siRNA at the indicated times prior to or after infection. Forty-eight hours after infection total RNA was collected and viral RNA was assessed as in Fig 1.
The rate of viral replication does not affect RNAi resistance
The establishment of siRNA resistance could in part, be due to the ability of a rapidly replicating WNV to saturate the RNAi degradation machinery. To test if the replication rate affected siRNA resistance, we used an attenuated lineage II WNV that contains a GFP marker gene inserted into the 3' UTR and replicates more slowly than wild-type lineage I or II WNV [36]. Because the nucleotide sequence of the lineage II WNV was different than the lineage I WNV, a new sequence-specific siRNA was designed (6337) that targeted the analogous region on NS3 as siRNA 6349. Huh7.5 cells were transfected with Cap-Mut, 6349, or 6337 at 18 hours prior to or 10 hours after infection with the attenuated lineage II WNV, and viral RNA content was determined at 48 hours post-infection (Fig 2B). As expected, pretreatment with either Cap Mut or the lineage I-specific 6349 siRNA did not inhibit replication, whereas pretreatment with 6337 siRNA strongly blocked replication (~ 60-fold). In contrast, treatment with any of the three siRNAs 10 hours after infection demonstrated no inhibitory effect. Thus, a WNV strain with a lower replication rate similarly resisted the inhibitory effects of RNAi soon after replication was established.
The mode of introduction of siRNA affects the ability to establish RNAi
Based on the timing experiments, the establishment of resistance to RNAi correlated with the onset of WNV RNA replication [37]. Shortly after infection, flaviviruses induce vesicular membrane proliferation that becomes the site of viral RNA replication [6,7,38]. Because the lipid-based transfection reagent targets nucleic acids to the cytoplasm (Mirus Corp, personal communication), the presence of additional membranes between the viral RNA and the cytoplasm could prevent the siRNA from reaching the actively replicating WNV RNA complex. Electroporation, in contrast, transiently opens pores in cellular membranes [39] allowing nucleic acids and cytoplasmic components to cross membranous structures such as the nucleus, endoplasmic reticulum, and potentially the membranous vesicles induced by WNV.
To determine if the mode of delivery of siRNA affected RNAi resistance, we tested whether siRNA could inhibit replication of a persistently replicating subgenomic lineage I WNV replicon; this cell line (Huh7.5-Rep) expresses the non-structural but lacks the structural proteins of WNV (Fig 3A). Huh7.5-Rep cells were transfected with 6337 and 6349 siRNAs using a lipid-based reagent or by electroporation, cultured, and assayed for reduction of NS3 antigen expression 72 hours later (Fig 3B). Replicon-expressing cells that were transfected with siRNA using the lipid-based reagent showed no significant reduction in viral protein or RNA. In contrast, using electroporation, NS3 protein levels were reduced by approximately 10-fold by 6349 siRNA, but not by the lineage II-specific 6337 siRNA. The reduction of NS3 antigen levels correlated with ~7.5-fold decreases in replicon RNA levels in the presence of 6349 (Fig 3C).
Figure 3 Mode of siRNA introduction influences WNV RNAi susceptibility. A. Huh7.5-Rep cells. (Top) Diagram of the genetic structure of the pWN5'Pur replicon. (Bottom) Flow cytometric analysis of Huh7.5 cells that express the pWN5'Pur replicon. Only non-structural proteins (e.g., NS1 and NS3 but not E) are expressed. B. siRNA treatment of Huh7.5-Rep cells. Huh7.5-Rep cells were mock-transfected, transfected with TKO reagent complexed with 6337 or 6349 siRNAs, or electroporated with 6337 or 6349 siRNAs. Three days later, cells were processed for viral NS3 protein expression by flow cytometry using anti-NS3 antibody (right). Fold inhibition of NS3 antigen production was determined using the formula (% NS3 positive mock electroporated / % NS3 positive siRNA electroporated). C. RNA analysis of Huh7.5-Rep cells electroporated with siRNA. Huh7.5-Rep cells were electroporated with 6349 or 7353 siRNA as described in Materials and Methods. Three days later, total cellular RNA was collected and viral RNA was assessed. Fold inhibition was determined by dividing the amount of viral RNA in mock electroporated samples to the amount of viral RNA in siRNA electroporated samples. The results are an average of three independent experiments and error bars indicate standard error of the mean.
Cellular localization of siRNA
The preceding data suggested that lipid-mediated transfection delivered siRNA into the cytoplasm, whereas electroporation at least transiently exposed replicating viral RNA to the RNAi response. However, it remained unclear whether the amount of siRNA delivered or the localization determined its ability to inhibit WNV RNA. To assess the relative amount and distribution of siRNA after transfection or electroporation, Huh7.5 cells were transfected or electroporated with Cy5-labeled Cap siRNA and analyzed 18 hours later for Cy5 expression by flow cytometry and localization by fluorescence microscopy (Fig 4). Although all transfected cells were positive for Cy5 fluorescence, the signal was significantly higher in lipid-transfected cells than in electroporated cells (Fig 4A, geometric mean fluorescence intensity 4370 compared to 766). Microscopic analysis of lipid-transfected cells showed Cy5 signal primarily in the cytoplasm, with exclusion of Cy5 from the nucleus (Fig 4B, middle panels). In contrast, Cy5 signal was observed diffusely throughout the cell after electroporation (Fig 4B, right panels), suggesting that electroporation effectively delivered siRNA across intracellular membranes. Taken together, our data suggests that the cellular localization of siRNA appears more important than the absolute amount of siRNA delivered into the cell in determining its effectiveness against actively replicating WNV RNA.
Figure 4 Localization of siRNA. Huh7.5 cells were transfected with Cy5 labeled Cap siRNA using the lipid TKO reagent (middle panels) or by electroporation (right panels). Eighteen hours later, cells were collected and analyzed for Cy5 fluorescence by (A) flow cytometry or (B) fluorescence microscopy as described in Materials and Methods. Arrows denote the position of the nucleus. The gain in the Cy5 micrograph of electroporated cells was increased to compensate for lower levels of intracellular siRNA as compared to lipid-transfected samples.
RNAi against other mRNAs is intact in WNV-infected cells
Although the mode and location of siRNA introduction could affect the sensitivity of actively replicating WNV to RNAi, we speculated that WNV could additionally evade RNAi by directly inhibiting one or more steps of the RNAi pathway. Targeted inhibition of the RNAi pathway has been observed in plant viruses, and has been recently reported with several mammalian viruses including LaCrosse virus, adenovirus, and influenza A virus [32-34,40-43]. To determine if WNV replication directly attenuated the RNAi response, we tested the efficiency of siRNA-mediated inhibition of Influenza A virus M2 protein expression in cells that actively replicated WNV genomes (Fig 5). Mock-infected Huh7.5 cells, WNV-infected Huh7.5 cells (10 hours post-infection), and Huh7.5-Rep cells were co-transfected with an M2 expression plasmid and either Cap or M2 siRNA. Twenty-four hours later, cells were analyzed for M2 expression by flow cytometry. As expected, transfection of Cap siRNA did not significantly affect the expression of M2 in Huh7.5, Huh7.5-Rep, or WNV infected Huh7.5 cells. However, transfection of M2 siRNA effectively reduced the expression of the M2 protein in all cell types, including those that actively replicated WNV RNA. Thus, WNV replication per se did not affect the establishment of RNAi of a heterologous gene.
Figure 5 RNAi is active in WNV infected cells. RNAi of influenza M2 gene in cells that replicate WNV RNA. Huh7.5, Huh7.5-Rep, and WNV infected Huh7.5 cells (8 hours post infection) were transfected with pCM2 and Cap or M2 siRNA as described in Materials and Methods. 24 hours later cells were processed by flow cytometry for M2 expression using antibody 14C2. The percentage of M2 inhibition was calculated according to the following formula: (1 – (% M2 expression of siRNA-transfected cells / % M2 expression in cells transfected with transfection vehicle only) × 100). The results are an average of three independent experiments and error bars indicate standard error of the mean.
Viral translation is necessary for WNV RNAi resistance
The experiments above suggest that resistance to RNAi by WNV occurs in the setting of ongoing viral replication. During the de novo infection of a cell, translation of the input positive viral RNA strand is required before replication occurs [2]. To more finely dissect the kinetics of RNAi resistance with respect to the initiation of RNA replication, we added puromycin, a reversible inhibitor of protein chain elongation. Because puromycin inhibits cellular and viral protein translation, we first assessed how it independently affected the establishment of RNAi. Cap or Cap Mut siRNA were transfected into cells in the presence of puromycin. Four hours later, cells were infected with WNV for an additional four hours, and then free virus and puromycin were removed by serial washing of infected monolayers. Forty-eight hours after initial infection, cells were analyzed for WNV envelope protein expression (Fig 6A). When translation of full-length viral protein was reduced by puromycin, Cap siRNA more effectively inhibited WNV antigen production (90-fold versus 10-fold reduction), suggesting that cellular translation was not necessary for priming the RNAi response and that delay of onset of WNV translation enhanced the efficiency of siRNA-mediated inhibition. Importantly, the inhibition was sequence-specific as no significant decrease in viral antigen expression was observed with the Cap Mut siRNA.
Figure 6 WNV RNAi resistance is dependent of viral translation early in infection. A. Puromycin does not interfere with RNAi. Huh7.5 cells were mock-treated or treated with 6 μg/ml puromycin and transfected with Cap or Cap Mut siRNA for 4 hours. Cells were washed twice, the puromycin replaced, and cells were infected with WNV at MOI 0.1. Four hours later cells were washed twice and replaced with medium that lacked puromycin. Forty-eight hours after infection, cells were collected and analyzed for WNV envelope protein expression by flow cytometry. Fold inhibition was calculated as described above. The results are an average of three independent experiments and error bars indicate standard error of the mean. B. Puromycin time course. Huh7.5 cells were infected with WNV (MOI = 0.01), and puromycin was added at the indicated times before or after infection. At 9 hours post-infection cells were washed and transfected with Cap or Cap Mut siRNA. Forty-eight hours later, WNV envelope protein expression was assessed by flow cytometry. Fold inhibition was calculated as described in Fig 3B. The results are an average of three independent experiments and error bars indicate standard error of the mean.
To define the kinetics of RNAi resistance around the time of initial viral replication, a puromycin time course was performed (Fig 6B). Cells were infected with WNV at MOI 0.01 and puromycin was added to Huh7.5 cells at different times (-1, 0, 1, 2, 3, 4, 5, 6, 7, or 8 hours) before or after infection. At 9 hours after WNV infection, all cells were transfected with Cap or Cap Mut siRNA and puromycin was removed from the medium. Cells treated with puromycin from -1 to 1 hours post-infection were greatly protected against WNV infection. However, significant attenuation of RNAi emerged when puromycin was added just four hours after infection. These results suggest that the induction of RNAi resistance by WNV depends on translation of viral polyprotein and occurs as early as 4 hours after infection.
Discussion
In this paper we examine the ability of exogenous siRNA to inhibit the replication of WNV. Developing strategies for specific inhibition of WNV is an important goal as no current therapy exists for infected individuals. siRNA has been proposed as a potential therapy against several viruses, and we have previously demonstrated that plasmid based RNAi is effective against WNV in vitro [21]. Here, we tested the ability of exogenously generated siRNA to inhibit WNV infection, as this reagent may be more practical for clinical use because there is little possibility of adverse integration into a patient's genome. Using a conventional lipid-based delivery system that targets siRNA to the cytoplasm, we confirmed that pretreatment of cells prevented infection. However, resistance to RNAi was observed when siRNA was delivered after viral translation and replication had commenced. In contrast, when siRNA was delivered by electroporation, a technique that allows macromolecules to pass across intracellular membranes, it reduced viral replication in a sequence-specific manner even if active replication was already underway. The data in this manuscript provide a first description of flavivirus resistance to RNAi during infection, and suggests a possible mechanism: WNV resists exogenously-introduced siRNA by replicating in a compartment that is sequestered behind cellular membranes.
Poliovirus is a positive strand RNA virus that replicates its genome in the cytoplasm of infected cells [44] and although susceptible to siRNA treatment, may relieve the selective pressure from siRNA by accumulating mutations in the targeted region [22,23,35]. In contrast, despite sequencing multiple independent isolates, we were unable to identify any mutations in siRNA-targeted regions in WNV-infected or replicon-expressing cells that were exposed to inhibitory siRNA. Also with poliovirus, some of the RNAi resistance could be overcome by administration of multiple inhibitory siRNA to disparate regions of the genome [35]. However, this was not observed with WNV, as simultaneous delivery of multiple inhibitory siRNA did not affect the resistance to RNAi in WNV-infected or replicon expressing cells. Thus, unlike poliovirus, WNV does not appear evade RNAi by mutating its target sequences.
WNV polyprotein translation and RNA replication within hours of infection [45]. Treatment of cells with the protein chain elongation inhibitor puromycin confirmed that establishment of RNAi resistance depended on translation of the infectious viral RNA, and that this occurred within the first four hours of infection. Electroporation of siRNA into cells expressing actively replicating WNV replicons aborted replication, suggesting both a mechanism and a means to overcome WNV-induced RNAi resistance. Nonetheless, it is possible that the method of delivery independently affects the ability of the siRNA to prime the RNAi response. The route of delivery differs between TKO transfection and electroporation (endosome versus direct transfer across membranes), and a proportion of TKO transfected siRNA may remain in endosomes for extended periods of time after transfection. However, even though five-fold more siRNA was detected in cells transfected by the TKO method, no inhibition was observed in cells that had ongoing replication of WNV RNA. We favor an alternative explanation in which WNV replication complexes are physically sequestered in a de novo generated specialized membranous compartment that is inaccessible to the cytoplasmic RNAi machinery. However, if siRNA gains access to these compartments (e.g., by electroporation) the RNAi machinery can be primed for sequence-specific degradation of viral RNA. Consistent with this, several studies have indicated that the reorganization and proliferation of endoplasmic reticulum membranes induced by flaviviruses is essential for efficient replication [6,7,38]. Uchil and Satchidanandam [46] proposed a model of flavivirus RNA replication in which viral dsRNA is enclosed within a double membrane structure; such a model could explain our findings. When siRNA is introduced by transfection prior to WNV infection, the cytoplasmic RNAi machinery becomes primed, and efficiently degrades infectious viral RNA after nucleocapsid penetration but before translation. In contrast, when siRNA is introduced by lipid-based transfection several hours after infection, replicating viral RNA is sequestered from the cytoplasm where the RNAi response is primed, allowing near-normal levels of replication to occur. During electroporation, however, siRNAs may be delivered across membranes and into the lumen of the viral replication compartment. How the Dicer and RISC components gain access into the replication compartment remains unknown. Although some cytoplasmic proteins may translocate across membranes during electroporation, the large size of the RNAi machinery may limit transport across membrane structures. We speculate that a small amount of Dicer and RISC gains access to the lumen of the replication compartment during its formation, and become activated when siRNA are delivered via electroporation. Clearly, additional experiments are necessary to confirm the precise mechanism.
Because treatment with siRNA in vivo would occur after an infection has been established, post-infection administration of siRNA in cell culture may reasonably predict the therapeutic utility of siRNA against individual viruses. Although many recent reports, including our own [21-27], have documented that pretreatment of cells with siRNA effectively aborts infection, few studies have examined the effects of siRNA treatment on established virus infection in vitro or in vivo. For example, siRNA administration into mice 5 hours after Influenza A infection only modestly reduced viral titers [31]. Several groups have recently demonstrated that electroporation of hepatitis C (HCV)-specific siRNA reduced HCV RNA replication in cells expressing subgenomic replicon [26,27,47-50], results that are consistent with ours. In contrast, one study reported that siRNA transfection with oligofectamine, a lipid-based reagent, modestly reduced HCV protein expression and RNA replication in HCV-replicon expressing cells [50]. The disparity among results with lipid-based transfection systems may be reagent-based, as oligofectamine is reported to deliver a fraction of the siRNA across membranes (Invitrogen, personal communication) and thus, may transport small amounts of siRNA into the HCV replication compartment.
Conclusion
The data presented here suggests that actively replicating WNV avoids the RNAi response by replicating in a manner that is inaccessible to cytoplasm-targeted delivery of siRNA. Consistent with this, we observed little therapeutic effect of siRNA against WNV in vivo in mice (B. Geiss, M. Diamond, unpublished observation). No protection against WNV was observed when mice were treated with siRNA 24 hours after infection [51]. This lack of siRNA-mediated therapeutic effect in vivo correlates with the induction of siRNA resistance that we observe in vitro. Future studies will address the role of flavivirus-induced membrane reorganization in RNAi resistance, and determine whether this mechanism is a common feature of other positive strand enveloped RNA viruses. Such information may inform the development of alternate delivery systems that allow siRNA to efficiently cross intracellular membranes and inhibit actively replicating enveloped viruses.
Materials and methods
Cells, viruses, and plasmids
Baby hamster kidney cells (BHK21-15 [52]) and human Huh-7.5 hepatoma cells (gift from C. Rice, New York, NY [53]) were cultured in Dulbecco's Modified Eagle Medium with 10% fetal bovine serum as previously described [52]. The lineage I (3000.0259, New York 2000) and the lineage II WNV strains have been described previously [54-56].
The lineage I WNV subgenomic replicon plasmid pWN5'Pur was generated from a genomic clone of the New York 1999 strain (plasmids pWN-AB1 and pWN-CG) provided by R. Kinney (Centers for Disease Control, Fort Collins, CO). pWN5'Pur was generated by deleting WNV nucleotides 181–2379 and fusing the first 31 amino acids of the capsid protein followed by the FMDV 2A autocleavage peptide [57] and the puromycin N-acetyl transferase (PAC) gene [58]. The EMCV IRES [53] was placed downstream of the PAC stop codon, so that translation of the WNV structural proteins begins at nucleotide 2380 (Methionine 794). The lineage II WNV genomic clone containing a Not I restriction site or an IRES-driven GFP have been described [36,56].
DNA template for replicon RNA transcription was prepared by linearization of pWN5'Pur with Xba I restriction endonuclease followed by phenol:chloroform extraction and ethanol precipitation. Replicon RNA was generated using the Amplicap T7 High Yield Message Maker kit (Epicenter Technologies, Madison WI). T7 RNA transcripts were electroporated into Huh7.5 cells as described below to generate Huh7.5-Rep cells. Huh7.5-Rep cells were stably selected with 5 μg/ml puromycin (Sigma-Aldrich, St. Louis MO). Reverse transcriptase PCR was performed as previously described [59,60] using primers 3026F (5'TGACTCGAAGATCATTGGAA) and 4496R (5'ATCCATATCTTCCAAGGTGC). Plasmid pCAGGS M2 was described previously [21].
siRNA production, RNA and DNA transfection
siRNA were generated in vitro by run-off transcription from a partially double-stranded oligonucleotide template. Oligonucleotides that contained the T7 RNA polymerase promoter (5'AAATTTAATACGACTCACTATA) were annealed to a 75-mer oligonucleotide, which contained an antisense T7 RNA polymerase promoter sequence, 19–21 nucleotides corresponding to the target sequence (Table 1), a 10 nucleotide loop region, 19–21 nucleotides complementary to the target sequence, and two adenine residues (5'AA (sense 21) AACCAGAAGA (antisense 21) TATAGTGAGTCGTATTAAATTT). Targeted sequences were chosen using the SciTools RNAi design program (Integrated DNA Technology, Coralville, IA) and compared against the GenBank database to exclude sequences that may affect cellular genes. Polyacrylamide gel electrophoresis (PAGE)-purified oligonucleotides were purchased from Integrated DNA Technologies (Coralville, IA).
siRNA were transcribed using the MegaShortScript T7 Transcription Kit (Ambion, Austin, TX) according to the manufacturers recommendations with the exception that 200 additional units of T7 RNA Polymerase (Ambion) were included in each 20 μl reaction. RNA transcription reactions were carried out at 37°C for 1.5 hours, treated with DNAse I for 15 min, extracted with phenol and chloroform, desalted over ChromaSpin TE-10 columns (BD Biosciences, Palo Alto, CA), and stored at -80°C. siRNA were quantified by PAGE gel electrophoresis and UV spectroscopy. Cy5 labeled siRNA was generated by adding 1 mM Cy5-UTP (Amersham Biosciences, Piscataway, NJ) to transcription reactions. Huh7.5 cells were transfected at various times before or after infection with 1 μg siRNA using 5 μl Trans-IT TKO reagent (Mirus Corp., Madison, WI) according to the manufacturer's instructions. Electroporations were performed using a BTX ElectroSquarePorator as described [53]. In all electroporation experiments, 5 × 106 cells were electroporated in the presence of 50 μg siRNA. TKO transfected cells were washed twice with fresh DMEM media before infection with WNV. Viral antigen expression in WNV-infected cells was analyzed by flow cytometry (FACSCalibur, Becton-Dickinson) using the anti-WNV envelope E1 monoclonal antibody [61]. Monoclonal antibodies against NS1, (9NS1; K. Chung and M. Diamond, unpublished data) and NS3, (clone E1E6; R. Beatty and E. Harris, unpublished data) were used to detect viral antigen in cells that expressed WNV replicons. Goat-anti-mouse IgG -FITC (Sigma-Aldrich, St. Louis MO) was used to detect primary antibodies. Plasmid DNA transfections were performed using Trans-IT LT1 reagent (Mirus Corp., Madison WI) at a ratio of 8 μl transfection reagent / 1 μg plasmid DNA according to the manufacturers recommendation. For experiments involving co-transfection of plasmid and siRNA, plasmid DNA and siRNA were separately complexed with the appropriate transfection reagent, mixed together and incubated at room temperature for 15 minutes, and added to cells.
Quantitative real-time reverse transcriptase PCR
WNV infected samples were collected at 48 hours post-infection, and total RNA was isolated using RNEasy RNA extraction columns (Qiagen, Valencia, CA) according to the manufacturer's instructions. Real-time reverse transcriptase PCR and quantitation of WNV transcripts was performed as previously described [62]. Quantitation of Lineage I replicon RNA was performed using a primer-probe set directed towards the 3' UTR WNV genome (forward primer 5'AGAGTGCAGTCTGCGATAGTGC; probe 5' Fam ACAAAGGCAAACCAACGCCCCA TAMRA; reverse primer 5'CCTTTCGCCCTGGTTAACA). Quantitation of Lineage II genome was performed using a primer set directed towards the 3' UTR of the Lineage II WNV genome (forward primer 5'AGAGTGCAGTCTGCGATAGTGC; probe 5' FAM ACAAAGGCAAAACATCGCCCCA TAMRA; reverse primer 5'CCCTTCTCCCTGGTTAACA).
Fluorescence microscopy
Cy5-Cap transfected or electroporated Huh7.5 cells were plated onto Lab-Tek glass slides (Nalge Nunc, Naperville IL) and incubated for 18 hours at 37°C. Cells were fixed in cold 4% paraformaldehyde, washed extensively, permeabilized with 0.5% Triton X-100, and mounted in Prolong Gold Plus DAPI mounting reagent (Molecular Probes, Eugene OR). Slides were visualized and digitally captured using a Zeiss Axioskop microscope (Zeiss Microimaging, Thornwood, NY).
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
BJG designed and constructed the subgenomic WNV replicons, designed and performed all experiments, and helped draft the manuscript. TCP designed, constructed, and tested the WNV-GFP clone and critically reviewed the manuscript. MSD and BJG designed the study, and MSD helped draft and critically review the manuscript.
Acknowledgements
We thank Richard Kinney for the WNV plasmids pWN-AB1 and pWN-CG, Andy Pekosz and Matt McCown for the plasmid pCAGGS M2 and for monoclonal antibody 14C2, and Robert Beatty and Eva Harris for monoclonal antibody E1E6. We would also like to thank Elizabeth Moulton for helping to set up the siRNA production system, Keril Blight for use of the electroporation apparatus, Andy Pekosz and Robert Doms for critical review of this manuscript and helpful discussions, and the Blight, Olivo, Leib, Pekosz, Diamond, Klein, and Morrison laboratories for helpful discussions. This work was supported by a grant from the NIH (1 U01 AI53870). BG was supported by a Ruth L. Kirschstein National Research Service Award, 5 T32 AI07172-25,
==== Refs
Petersen LR Marfin AA West Nile virus: a primer for the clinician Ann Intern Med 2002 137 173 179 12160365
Lindenbach Flaviviridae: The Viruses and Their Replication Fields Virology 2001 1 4th Philadelphia, Lippincott, Williams, and Wilkins 991 1041
Chambers TJ McCourt DW Rice CM Production of yellow fever virus proteins in infected cells: identification of discrete polyprotein species and analysis of cleavage kinetics using region-specific polyclonal antisera Virology 1990 177 159 174 2353452 10.1016/0042-6822(90)90470-C
Mackenzie JM Jones MK Young PR Immunolocalization of the dengue virus nonstructural glycoprotein NS1 suggests a role in viral RNA replication Virology 1996 220 232 240 8659120 10.1006/viro.1996.0307
Mackenzie JM Westaway EG Assembly and maturation of the flavivirus Kunjin virus appear to occur in the rough endoplasmic reticulum and along the secretory pathway, respectively J Virol 2001 75 10787 10799 11602720 10.1128/JVI.75.22.10787-10799.2001
Westaway EG Khromykh AA Mackenzie JM Nascent flavivirus RNA colocalized in situ with double-stranded RNA in stable replication complexes Virology 1999 258 108 117 10329573 10.1006/viro.1999.9683
Westaway EG Mackenzie JM Kenney MT Jones MK Khromykh AA Ultrastructure of Kunjin virus-infected cells: colocalization of NS1 and NS3 with double-stranded RNA, and of NS2B with NS3, in virus-induced membrane structures J Virol 1997 71 6650 6661 9261387
Meister G Landthaler M Dorsett Y Tuschl T Sequence-specific inhibition of microRNA- and siRNA-induced RNA silencing Rna 2004 10 544 550 14970398 10.1261/rna.5235104
Que Q Jorgensen RA Homology-based control of gene expression patterns in transgenic petunia flowers Dev Genet 1998 22 100 109 9499584 10.1002/(SICI)1520-6408(1998)22:1<100::AID-DVG10>3.0.CO;2-D
Fire A Xu S Montgomery MK Kostas SA Driver SE Mello CC Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans Nature 1998 391 806 811 9486653 10.1038/35888
Ngo H Tschudi C Gull K Ullu E Double-stranded RNA induces mRNA degradation in Trypanosoma brucei Proc Natl Acad Sci U S A 1998 95 14687 14692 9843950 10.1073/pnas.95.25.14687
Ullu E Tschudi C Chakraborty T RNA interference in protozoan parasites Cell Microbiol 2004 6 509 519 15104593 10.1111/j.1462-5822.2004.00399.x
Hoa NT Keene KM Olson KE Zheng L Characterization of RNA interference in an Anopheles gambiae cell line Insect Biochem Mol Biol 2003 33 949 957 12915186 10.1016/S0965-1748(03)00101-2
Paddison PJ Caudy AA Bernstein E Hannon GJ Conklin DS Short hairpin RNAs (shRNAs) induce sequence-specific silencing in mammalian cells Genes Dev 2002 16 948 958 11959843 10.1101/gad.981002
Paddison PJ Caudy AA Hannon GJ Stable suppression of gene expression by RNAi in mammalian cells Proc Natl Acad Sci U S A 2002 99 1443 1448 11818553 10.1073/pnas.032652399
Waterhouse PM Wang MB Lough T Gene silencing as an adaptive defence against viruses Nature 2001 411 834 842 11459066 10.1038/35081168
Elbashir SM Lendeckel W Tuschl T RNA interference is mediated by 21- and 22-nucleotide RNAs Genes Dev 2001 15 188 200 11157775 10.1101/gad.862301
Elbashir SM Harborth J Lendeckel W Yalcin A Weber K Tuschl T Duplexes of 21-nucleotide RNAs mediate RNA interference in cultured mammalian cells Nature 2001 411 494 498 11373684 10.1038/35078107
Garrus JE von Schwedler UK Pornillos OW Morham SG Zavitz KH Wang HE Wettstein DA Stray KM Cote M Rich RL Myszka DG Sundquist WI Tsg101 and the vacuolar protein sorting pathway are essential for HIV-1 budding Cell 2001 107 55 65 11595185 10.1016/S0092-8674(01)00506-2
Sui G Soohoo C Affar el B Gay F Shi Y Forrester WC A DNA vector-based RNAi technology to suppress gene expression in mammalian cells Proc Natl Acad Sci U S A 2002 99 5515 5520 11960009 10.1073/pnas.082117599
McCown M Diamond MS Pekosz A The utility of siRNA transcripts produced by RNA polymerase i in down regulating viral gene expression and replication of negative- and positive-strand RNA viruses Virology 2003 313 514 524 12954218 10.1016/S0042-6822(03)00341-6
Gitlin L Karelsky S Andino R Short interfering RNA confers intracellular antiviral immunity in human cells Nature 2002 418 430 434 12087357 10.1038/nature00873
Saleh MC Van Rij RP Andino R RNA silencing in viral infections: insights from poliovirus Virus Res 2004 102 11 17 15068875 10.1016/j.virusres.2004.01.010
Ge Q McManus MT Nguyen T Shen CH Sharp PA Eisen HN Chen J RNA interference of influenza virus production by directly targeting mRNA for degradation and indirectly inhibiting all viral RNA transcription Proc Natl Acad Sci U S A 2003 100 2718 2723 12594334 10.1073/pnas.0437841100
Jacque JM Triques K Stevenson M Modulation of HIV-1 replication by RNA interference Nature 2002 418 435 438 12087358 10.1038/nature00896
Wilson JA Jayasena S Khvorova A Sabatinos S Rodrigue-Gervais IG Arya S Sarangi F Harris-Brandts M Beaulieu S Richardson CD RNA interference blocks gene expression and RNA synthesis from hepatitis C replicons propagated in human liver cells Proc Natl Acad Sci U S A 2003 100 2783 2788 12594341 10.1073/pnas.252758799
Randall G Grakoui A Rice CM Clearance of replicating hepatitis C virus replicon RNAs in cell culture by small interfering RNAs Proc Natl Acad Sci U S A 2003 100 235 240 12518066 10.1073/pnas.0235524100
McCaffrey AP Nakai H Pandey K Huang Z Salazar FH Xu H Wieland SF Marion PL Kay MA Inhibition of hepatitis B virus in mice by RNA interference Nat Biotechnol 2003 21 639 644 12740585 10.1038/nbt824
Giladi H Ketzinel-Gilad M Rivkin L Felig Y Nussbaum O Galun E Small interfering RNA inhibits hepatitis B virus replication in mice Mol Ther 2003 8 769 776 14599810 10.1016/S1525-0016(03)00244-2
Tompkins SM Lo CY Tumpey TM Epstein SL Protection against lethal influenza virus challenge by RNA interference in vivo Proc Natl Acad Sci U S A 2004 101 8682 8686 15173583 10.1073/pnas.0402630101
Ge Q Filip L Bai A Nguyen T Eisen HN Chen J Inhibition of influenza virus production in virus-infected mice by RNA interference Proc Natl Acad Sci U S A 2004 101 8676 8681 15173599 10.1073/pnas.0402486101
Chen J Li WX Xie D Peng JR Ding SW Viral virulence protein suppresses RNA silencing-mediated defense but upregulates the role of microrna in host gene expression Plant Cell 2004 16 1302 1313 15100397 10.1105/tpc.018986
Kasschau KD Xie Z Allen E Llave C Chapman EJ Krizan KA Carrington JC P1/HC-Pro, a viral suppressor of RNA silencing, interferes with Arabidopsis development and miRNA unction Dev Cell 2003 4 205 217 12586064 10.1016/S1534-5807(03)00025-X
Qu F Ren T Morris TJ The coat protein of turnip crinkle virus suppresses posttranscriptional gene silencing at an early initiation step J Virol 2003 77 511 522 12477856 10.1128/JVI.77.1.511-522.2003
Gitlin L Stone JK Andino R Poliovirus Escape from RNA Interference: Short Interfering RNA-Target Recognition and Implications for Therapeutic Approaches J Virol 2005 79 1027 1035 15613331 10.1128/JVI.79.2.1027-1035.2005
Pierson TC Diamond MS Ahmed AA Valentine LE Davis CW Samuel MA Hanna SL Puffer BA Doms RW An infectious West Nile virus that expresses a GFP reporter gene Virology 2005 334 28 40 15749120 10.1016/j.virol.2005.01.021
Takegami T Hotta S Synthesis and localization of Japanese encephalitis virus RNAs in the infected cells Microbiol Immunol 1990 34 849 857 1963921
Mackenzie JM Jones MK Westaway EG Markers for trans-Golgi membranes and the intermediate compartment localize to induced membranes with distinct replication functions in flavivirus-infected cells J Virol 1999 73 9555 9567 10516064
Weaver JC Electroporation: a general phenomenon for manipulating cells and tissues J Cell Biochem 1993 51 426 435 8496245
Joost Haasnoot PC Cupac D Berkhout B Inhibition of virus replication by RNA interference J Biomed Sci 2003 10 607 616 14576463 10.1159/000073526
Li WX Li H Lu R Li F Dus M Atkinson P Brydon EW Johnson KL Garcia-Sastre A Ball LA Palese P Ding SW Interferon antagonist proteins of influenza and vaccinia viruses are suppressors of RNA silencing Proc Natl Acad Sci U S A 2004 101 1350 1355 14745017 10.1073/pnas.0308308100
Lu S Cullen BR Adenovirus VA1 noncoding RNA can inhibit small interfering RNA and MicroRNA biogenesis J Virol 2004 78 12868 12876 15542639 10.1128/JVI.78.23.12868-12876.2004
Soldan SS Plassmeyer ML Matukonis MK Gonzalez-Scarano F La Crosse Virus Nonstructural Protein NSs Counteracts the Effects of Short Interfering RNA J Virol 2005 79 234 244 15596819 10.1128/JVI.79.1.234-244.2005
Baltimore D Eggers HJ Franklin RM Tamm I Poliovirus-induced RNA polymerase and the effects of virus-specific inhibitors on its production Proc Natl Acad Sci U S A 1963 49 843 849 13966259
Lindenbach BD Rice CM trans-Complementation of yellow fever virus NS1 reveals a role in early RNA replication J Virol 1997 71 9608 9617 9371625
Uchil PD Satchidanandam V Architecture of the flaviviral replication complex. Protease, nuclease, and detergents reveal encasement within double-layered membrane compartments J Biol Chem 2003 278 24388 24398 12700232 10.1074/jbc.M301717200
Sen A Steele R Ghosh AK Basu A Ray R Ray RB Inhibition of hepatitis C virus protein expression by RNA interference Virus Res 2003 96 27 35 12951263 10.1016/S0168-1702(03)00170-9
Zhang J Yamada O Sakamoto T Yoshida H Iwai T Matsushita Y Shimamura H Araki H Shimotohno K Down-regulation of viral replication by adenoviral-mediated expression of siRNA against cellular cofactors for hepatitis C virus Virology 2004 320 135 143 15003869 10.1016/j.virol.2003.11.023
Kronke J Kittler R Buchholz F Windisch MP Pietschmann T Bartenschlager R Frese M Alternative approaches for efficient inhibition of hepatitis C virus RNA replication by small interfering RNAs J Virol 2004 78 3436 3446 15016866 10.1128/JVI.78.7.3436-3446.2004
Takigawa Y Nagano-Fujii M Deng L Hidajat R Tanaka M Mizuta H Hotta H Suppression of hepatitis C virus replicon by RNA interference directed against the NS3 and NS5B regions of the viral genome Microbiol Immunol 2004 48 591 598 15322339
Bai F Wang T Pal U Bao F Gould LH Fikrig E Use of RNA interference to prevent lethal murine west nile virus infection J Infect Dis 2005 191 1148 1154 15747251 10.1086/428507
Diamond MS Roberts TG Edgil D Lu B Ernst J Harris E Modulation of Dengue virus infection in human cells by alpha, beta, and gamma interferons J Virol 2000 74 4957 4966 10799569 10.1128/JVI.74.11.4957-4966.2000
Blight KJ Kolykhalov AA Rice CM Efficient initiation of HCV RNA replication in cell culture Science 2000 290 1972 1974 11110665 10.1126/science.290.5498.1972
Ebel GD Dupuis AP Ngo K Nicholas D Kauffman E Jones SA Young D Maffei J Shi PY Bernard K Kramer LD Partial genetic characterization of West Nile virus strains, New York State, 2000 Emerg Infect Dis 2001 7 650 653 11585527
Wengler G Gross HJ Studies on virus-specific nucleic acids synthesized in vertebrate and mosquito cells infected with flaviviruses Virology 1978 89 423 437 568848 10.1016/0042-6822(78)90185-X
Yamshchikov V Mishin V Cominelli F A new strategy in design of +RNA virus infectious clones enabling their stable propagation in E. coli Virology 2001 281 272 280 11277699 10.1006/viro.2000.0793
de Felipe P Skipping the co-expression problem: the new 2A "CHYSEL" technology Genet Vaccines Ther 2004 2 13 15363111 10.1186/1479-0556-2-13
Vara J Malpartida F Hopwood DA Jimenez A Cloning and expression of a puromycin N-acetyl transferase gene from Streptomyces alboniger in Streptomyces lividans and Escherichia coli Gene 1985 33 197 206 3888783 10.1016/0378-1119(85)90094-0
Diamond MS Shrestha B Marri A Mahan D Engle M B cells and antibody play critical roles in the immediate defense of disseminated infection by West Nile encephalitis virus J Virol 2003 77 2578 2586 12551996 10.1128/JVI.77.4.2578-2586.2003
Lanciotti RS Kerst AJ Nasci RS Godsey MS Mitchell CJ Savage HM Komar N Panella NA Allen BC Volpe KE Davis BS Roehrig JT Rapid detection of west nile virus from human clinical specimens, field-collected mosquitoes, and avian samples by a TaqMan reverse transcriptase-PCR assay J Clin Microbiol 2000 38 4066 4071 11060069
Oliphant T Engle M Nybakken GE Doane C Johnson S Huang L Gorlatov S Mehlhop E Marri A Chung KM Ebel GD Kramer LD Fremont DH Diamond MS Development of a humanized monoclonal antibody with therapeutic potential against West Nile virus Nat Med 2005 11 522 530 15852016 10.1038/nm1240
Diamond MS Shrestha B Mehlhop E Sitati E Engle M Innate and adaptive immune responses determine protection against disseminated infection by West Nile encephalitis virus Viral Immunol 2003 16 259 278 14583143 10.1089/088282403322396082
| 15985182 | PMC1174879 | CC BY | 2021-01-04 16:38:58 | no | Virol J. 2005 Jun 28; 2:53 | utf-8 | Virol J | 2,005 | 10.1186/1743-422X-2-53 | oa_comm |
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1608950310.1371/journal.pbio.0030257Research ArticleBiophysicsImmunologyMolecular Biology/Structural BiologyNeuroscienceBiochemistryIn VitroStructure of a Pheromone Receptor-Associated MHC Molecule with an Open and Empty Groove Crystal Structure of the MHC Ib Molecule M10.5Olson Rich
1
Huey-Tubman Kathryn E
1
2
Dulac Catherine
3
Bjorkman Pamela J [email protected]
1
2
1 Division of Biology, California Institute of Technology, Pasadena, California, United States of America,2 Howard Hughes Medical Institute, California Institute of Technology, Pasadena, California, United States of America,3 Department of Molecular and Cellular Biology, Howard Hughes Medical Institute, Harvard University, Cambridge, Massachusetts, United States of AmericaSali Andrej Academic EditorUniversity of California, San FranciscoUnited States of America8 2005 12 7 2005 12 7 2005 3 8 e2576 4 2005 18 5 2005 Copyright: © 2005 Olson et al.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.
Sniffing Out the Structure of a Pheromone Receptor-Associated MHC Molecule
Neurons in the murine vomeronasal organ (VNO) express a family of class Ib major histocompatibility complex (MHC) proteins (M10s) that interact with the V2R class of VNO receptors. This interaction may play a direct role in the detection of pheromonal cues that initiate reproductive and territorial behaviors. The crystal structure of M10.5, an M10 family member, is similar to that of classical MHC molecules. However, the M10.5 counterpart of the MHC peptide-binding groove is open and unoccupied, revealing the first structure of an empty class I MHC molecule. Similar to empty MHC molecules, but unlike peptide-filled MHC proteins and non-peptide–binding MHC homologs, M10.5 is thermally unstable, suggesting that its groove is normally occupied. However, M10.5 does not bind endogenous peptides when expressed in mammalian cells or when offered a mixture of class I–binding peptides. The F pocket side of the M10.5 groove is open, suggesting that ligands larger than 8–10-mer class I–binding peptides could fit by extending out of the groove. Moreover, variable residues point up from the groove helices, rather than toward the groove as in classical MHC structures. These data suggest that M10s are unlikely to provide specific recognition of class I MHC–binding peptides, but are consistent with binding to other ligands, including proteins such as the V2Rs.
MHC-like protein M10.5 is expressed in the vomeronasal organ. The structure does not bind endogenous class I-binding peptides, but is thought to interact with a larger V2R pheromone receptor.
==== Body
Introduction
In most mammals, chemical communication between conspecific animals is involved in initiation of reproductive and territorial behaviors. The detection of these species- and gender-specific chemical cues, also called pheromones, is thought to involve receptors of the vomeronasal organ (VNO), a small neuronal epithelium located between the nasal cavity and the palate [1]. Although neurons of the main olfactory epithelium involved in odorant detection ultimately project to cognitive regions of the brain, vomeronasal neurons send inputs via the accessory olfactory bulb to specialized centers of the hypothalamus and amygdala, where they elicit neuroendocrine responses and behaviors such as oestrous synchronization, aggression, and sex discrimination [1–3].
Pheromone receptors belong to the ubiquitous family of G protein–coupled receptors (GPCRs), but are unrelated in sequence to main olfactory epithelium receptors that detect volatile odorants [4,5]. Mouse pheromone receptors can be divided into two subtypes, V1R and V2R, each of which is expressed in the dendritic tips of bipolar neurons in spatially distinct regions of the VNO. The human orthologs of most of these genes appear to be pseudogenes [1]. Mouse V1R receptors are found in the apical VNO domain, are thought to signal through the G-protein α-subunit Gαi2, and exhibit sequence similarity to the T2R family of bitter taste receptors [6]. V2R receptors, in contrast, are found in the basal VNO domain, are likely to signal via the Gαo molecule, and are related in sequence to metabotropic glutamate (mGluRs), GABAB (γ-aminobutyric acid-B), and calcium sensing receptors. V1R and V2R receptor family members, like all G protein–coupled receptors, contain seven putative transmembrane helices, but, in addition, V2R members include a large N-terminal extracellular domain.
Recently, it was shown that the V2R class of pheromone receptors specifically interacts with members of the mouse M1 and M10 families [7] of major histocompatibility complex (MHC) class Ib proteins [8], which do not appear to have human orthologs [9]. Classical class I MHC molecules, which exhibit high polymorphism in mice, humans, and other mammals, present peptides derived from cytoplasmic proteins to T cells during immune surveillance, and are expressed on most or all nucleated cells [10]. The less polymorphic non-classical class Ib molecules are expressed on a more limited subset of cells and are involved in a variety of functions, including presentation of hydrophobic peptides (e.g., by Qa-2), presentation of formylated peptides by H2-M3, and lipid presentation by CD1 proteins [11]. The non-classical M1 and M10 proteins are expressed exclusively in the VNO and appear to facilitate cell surface expression of V2Rs. Male mice that are genetically deficient in the class I MHC–associated β2-microglobulin (β2m) light chain show no surface expression of V2R pheromone receptors in the dendritic terminals of VNO sensory neurons and lack aggressive behavior toward intruder males [8]. Facilitation of V2R surface expression appears to involve M10 binding to a V2R because immunopurification of VNO receptors identified a multimolecular complex formed between M10, β2m, and a V2R protein. A given neuron generally expresses only one of the roughly 150 different subtypes of V2R receptor and one of the nine known VNO–specific MHC class Ib proteins (six M10 and three M1 family genes) [8,12]. Genetic analysis of individual neurons suggests specificity in binding between different V2R and M10 proteins [8].
Homologous receptors to the V2R proteins, such as the GABAB receptors and the umami and sweet taste receptors, also require accessory proteins for effective surface expression [13–15]. However, the nature of M10s as MHC molecules together with the association between the MHC and mating behavior in rodents [16] offers the intriguing possibility that M10 proteins play a direct role in mate or species recognition. Mice tend to choose mates with disparate MHC haplotypes [17], and pregnant mice have a higher incidence of spontaneous abortion if exposed to the scent of a male with a different MHC haplotype (the Bruce effect) [18]. These functions, as well as other species-specific indicators, may be mediated via M10-V2R complexes, perhaps by binding peptides or other accessory molecules in the canonical peptide-binding groove that exists in classical class I MHC structures [10]. Peptide binding during heavy and light chain assembly in the endoplasmic reticulum (ER), as occurs for classical class I MHC molecules, seems unlikely for M10 proteins because in situ hybridization assays demonstrate that TAP1 and TAP2, which are required for transport of cytoplasmic-derived peptides into the ER, are not expressed in the VNO [8]. However, a recent study shows that V2R–containing neurons are activated by nonameric class I MHC–binding peptides [19], suggesting involvement of exogenously acquired MHC–binding peptides, and perhaps M10 proteins, in individual mate or species recognition.
To determine the peptide-binding capability and structural characteristics of V2R receptor-associated MHC molecules, we expressed, characterized, and solved the crystal structure of the ectodomain of M10.5, an M10 family member. The M10.5 structure reveals an open conformation of the α1–α2 domain helices that contains no ordered peptidic or non-peptidic occupant, which represents the first example, to our knowledge, of the structure of an empty class I MHC molecule. Thermal stability studies suggest that the M10.5 groove is normally occupied; however M10.5 does not associate with the sorts of peptides that normally bind to class I MHC molecules. These results suggest a new and divergent function for the M10 family of proteins.
Results
The M10.5 Structure
A soluble form of M10.5 was expressed together with human β2m in baculovirus-infected insect cells, and soluble M10.5–β2m complexes were purified from cell supernatants. Similar efforts to express soluble M10.5 together with mouse β2m in insect cells or Chinese hamster ovary (CHO) cells were unsuccessful (data not shown), perhaps related to the observation that mouse class I MHC molecules form stronger heterodimeric complexes with human, as compared with mouse, β2m [20,21]. M10.5 was crystallized in space group P212121 with five molecules in the asymmetric unit. Most crystals diffracted weakly to 3.5–4.0 Å resolution, but an incomplete dataset from a rare crystal that diffracted beyond 3.5 Å was combined with a more complete dataset to 4.0 Å (Table 1). The structure was solved by molecular replacement using the mouse class Ib MHC molecule Qa-2 [22] as a search model. Refinement using non-crystallographic symmetry (NCS) restraints yielded a final model (R
cryst = 26.6%; R
free = 30.3%) (Table 1). Although data to 3.0 Å were included in the refinement, the high-resolution data are incomplete, thus the effective resolution of the structure is 3.2 Å. Two loops comprising residues 145–150 (α2 domain) and 194–197 (α3 domain) are missing in electron-density maps and are not included in our model.
Table 1 Crystallographic Data and Refinement Statistics
The overall structure of M10.5 resembles the structures of classical class I MHC molecules. A BLAST search [23] identifies mouse H-2Dd as the most closely related classical class I MHC molecule (approximately 50% amino acid identity) for which a structure is available [24], thus we have used H-2Dd for comparisons. As in other class I structures, the first 180 residues of the M10.5 heavy chain form the α1–α2 domain superdomain, which is composed of an eight-stranded anti-parallel β-sheet platform topped by two anti-parallel α-helices. The following approximately 90 residues form the α3 domain, which resembles an immunoglobulin constant region domain (Figure 1A and 1B). The non-covalently attached β2m light chain, also resembling an immunoglobulin constant region, contacts both the underside of the α1–α2 platform and one β-sheet of the α3 domain in an orientation consistent with previously solved mouse and human class I MHC structures, indicating that pairing human β2m with the mouse M10.5 heavy chain does not disrupt the overall M10.5 structure.
Figure 1 The Structure of M10.5
(A) Ribbon diagram of M10.5 (side view). The heavy chain is blue, the β2m light chain is green, disulfide bonds are yellow, and ordered carbohydrate attached to Asn223 is shown in ball-and-stick representation. Two disordered loops in the heavy chain are shown as dashed blue lines.
(B) Top view of the α1–α2 platform overlaid with an Fo-Fc annealed omit electron-density map contoured at 3.5 σ. The map was calculated for one of five molecules in the asymmetric unit using NCS restraints in the annealing process. Residues 55–84 and 137–174, shown in stick representation, were omitted from the structure factor calculation. Electron density is absent for residues 145–150 (dashed line), indicating that they are disordered. No significant electron density is observed in the groove between the α1 and α2 helices.
(C) Stereo view of the superposition of the α1–α2 platforms from M10.5, H-2Dd [24], FcRn [32], and HFE [33]. Structures were aligned using residues classified as platform β-sheet residues (see Materials and Methods). The cleft between α1 and α2 helices is significantly narrower in FcRn and HFE than in M10.5 and H-2Dd.
Ordered N-linked carbohydrate is observed attached to Asn223 within the loop joining the third and fourth β-strand in the α3 domain. The carbohydrate occludes the counterpart of the region in class I MHC molecules that is the major determinant for binding the T cell co-receptor CD8 [25], likely preventing M10.5 from participating in CD8+ T cell–mediated immunological responses. Ordered carbohydrate is not observed at either of the two other predicted N-linked glycosylation sites (Asn62 and Asn198), but the quality of the electron-density map is poor in these regions.
The α1–α2 Platform of M10.5 Contains an Open, but Apparently Empty, Groove
A large groove between the α1 and α2 domain helices forms the peptide-binding site in classical class I MHC molecules [10]. Structural studies of class I MHC molecules and homologs show a correlation between the degree of separation of the α1–α2 domain helices and the ability to bind peptides or other small molecules [26]. The helices are separated by approximately 18 Å in the center of the grooves of classical peptide-binding class I MHC molecules such as H-2Dd [24], and class I MHC-related proteins that bind other small molecule ligands also contain open grooves with separated α1 and α2 domain helices [27–31]. In contrast, the neonatal Fc receptor (FcRn) [32], the hemochromatosis protein HFE [33], and MIC-A [34], MHC homologs that do not bind small molecule ligands, have collapsed grooves with a smaller separation between the α1–α2 domain helices [32–34].
A superposition of the α1–α2 platforms of M10.5, H-2Dd, FcRn, and HFE illustrates the variation in groove size (Figure 1C) and demonstrates that M10.5 has an open groove more similar to the peptide-binding classical class I MHC molecules than the non–peptide-binding homologs. The M10.5 α1–α2 domain helices superimpose well with the analogous H-2Dd helices (RMS [root mean square] deviation of 1.22 Å for 164 of 181 α-carbon atoms), with the largest differences located in the region of the α2 domain helix immediately preceding six residues that are disordered in M10.5 (residues 145 to 150). The overall similarity of the M10.5 and H-2Dd α1–α2 platforms, which are contained within class I heavy chains that are paired with human (M10.5) or mouse (H-2Dd) β2m light chains, demonstrates that the observed structure of the M10.5 α1–α2 platform does not result from an artifactual change induced by pairing with human, rather than mouse, β2m. The more open character of the M10.5 groove as compared with the grooves of non–peptide-binding MHC homologs is seen when the calculated accessible surface areas are compared: approximately 730 Å2 for M10.5 compared with approximately 760 Å2 for typical class I MHC molecules [30,33], approximately 690 Å2 for H-2Dd, and approximately 415 Å2 and approximately 235 Å2 for HFE and FcRn, respectively [33] (see Materials and Methods). Thus the M10.5 groove can be classified as “open” and capable of binding to a ligand.
The α1–α2 domain groove is occupied by a peptide or mixture of peptides in all class I MHC structures solved to date, and continuous electron-density representing peptide(s) is always seen in the α1–α2 groove [35]. Indeed, attempts to crystallize an empty version of H-2Kb, which was expressed in cells lacking peptide-loading machinery and purified in the absence of a binding peptide, resulted in a crystal structure that revealed a peptide derived from the cell growth media [36]. M10.5 omit electron-density maps in which the α1 and α2 helices were removed from structure factor calculations return clear density for the omitted region (Figure 1B), indicating that the maps are of sufficient quality to detect bound molecules of the size of an 8–10 residue peptide. However, M10.5 electron-density maps show no ordered density corresponding to a peptide: Annealed Fo-Fc maps calculated using NCS constraints or tight restraints do not show continuous electron density within the groove (Figure S1), and maps calculated before or after 5-fold real space averaging also failed to reveal unbroken density in the cleft (data not shown). The possibility that a minor peak near the center of the groove in averaged maps represents a portion of a bound polyethylene glycol molecule is discussed in the caption of Figure 1. We conclude that the M10.5 groove does not contain a single defined peptidic or non-peptidic occupant or a mixture of compounds with a similar conformation.
Comparison of M10.5 and Classical Class I MHC Grooves
Crystallographic studies of classical class I MHC–peptide complexes have defined six pockets (A–F) within the peptide-binding groove that interact with various portions of bound peptides (Figure 2A) [37–39]. The A and F pockets at either end of the peptide-binding groove are largely conserved and interact with the N- and C-terminus, respectively, of the bound peptide [37,39]. The B, C, D, and E pockets contain residues that vary between alleles, resulting in different allele-specific peptide-binding preferences.
Figure 2 Groove Surface Characteristics
(A) Electrostatic surface representations calculated using GRASP [42] of the M10.5 (left) and H-2Dd (right) grooves. Positions of the six pockets are labeled A–F in yellow. In H-2Dd, Arg62 and Glu163 create a bridge over the peptide-binding groove. In M10.5, Glu63 and Arg167 define a similar feature. The H-2Dd groove is constricted between Asn70 and Arg155. In M10.5, Ala70 and Gly155 lead to a wider groove and continuous D and E pockets.
(B) Comparison of residues in the A and F pockets of classical class I MHC molecules and M10.5. All alanine-peptide (yellow, atoms color-coded according to atom type) is derived from the H-2Dd structure [24]. Left: In class I MHC proteins, a cluster of four tyrosine residues in the A pocket (Tyr7, Tyr59, Tyr159, and Tyr171) form hydrogen bonds with backbone atoms in the peptide N-terminus (residues are listed in single letter code for M10.5 and H-2Dd, respectively). In M10.5, Tyr7 is replaced by a threonine residue and Tyr171 is replaced with a cysteine. Five of the six M10 proteins have an additional tyrosine at position 33 that could potentially replace one of the two missing tyrosine residues. M10.5 has a serine instead of a glycine at position 26, which could participate in additional hydrogen-binding interactions. Right: The F pocket of class I MHC molecules is blocked on one end by Thr80, Tyr84, and Lys146. The M10.5 pocket is open on this end (see panels A and C) due to the substitution of a glutamate residue for Tyr84 and the disorder of the loop containing residue 146 (Asp in M10.5). Lys142 in M10.5 is missing sidechain density and has been modeled as an alanine, but most likely would not further occlude the M10.5 F pocket.
(C) Molecular surface representations of M10.5 and H-2Dd with an all-alanine peptide (yellow) derived from the H-2Dd structure [24] superimposed in the M10.5 groove. M10.5 atoms that came within 2.5 Å of the peptide trace were considered clashes and are colored red. Four additional alanine residues (green) were added to the C-terminus of the H-2Dd–binding peptide to illustrate that peptides binding in the M10.5 groove could extend out of the F pocket side.
The A and F pocket regions of the M10.5 groove contain substitutions that prevent the interactions that anchor peptide termini into the groove of a classical class I MHC molecule. Two of the four conserved tyrosines in the MHC class I A pocket (tyrosines 7, 59, 159, and 171), which hydrogen bond directly or through a water molecule with main chain atoms of the peptide N-terminal residue [37,39], are substituted in M10.5 as Thr7 and Cys171 (Figure 2B). Substitutions in other M10s also eliminate one or two of the four A pocket tyrosines [8,12]. It remains possible, however, that M10.5 and other M10s could bind a peptide N-terminus in a non-classical manner, utilizing residues including Thr7, Ser26, and Tyr33 as hydrogen-bond donors. Residues that typically anchor the C-terminus of a peptide in the class I MHC F pocket are also different in M10.5. In H-2Dd, the ninth and tenth residue of the bound peptide (P9 and P10) form hydrogen-bonding and van der Waals contacts with Val76, Asp77, Tyr84, Thr143, Lys146, and Trp147 (Figure 2B). In M10.5 and most other M10s, these residues are replaced by non-conservative substitutions, and the last two are missing entirely from the M10.5 electron-density map. Thus potential interactions between M10 F pocket groove residues and peptides would have to occur in a non-classical manner.
An interesting feature of possible functional relevance is the relatively open nature of the F pocket side of the M10.5 groove (Figure 2C), suggesting that potential small molecule or peptide occupants could extend out of this side of the groove. Homology models of other M10 proteins constructed using the M10.5 structure and the Swiss-model Protein modeling Server [40] suggest that this open character is a general feature of M10 proteins (data not shown). In class I MHC molecules, Tyr84 and Lys146 occlude the F pocket side of the groove, but the tyrosine to glutamate substitution at position 84 in M10.5 (Figure 2B) contributes to the more open M10.5 groove. In addition, M10.5 Asp146 is part of the disordered 145–150 region of the α2 domain helix, thus it does not contribute to closing the F pocket side of the groove. The disorder of these residues may be related to a need to maintain flexibility in the F pocket end of the M10.5 groove to allow binding of ligands that extend out of the M10.5 groove.
In order to make additional comparisons between the grooves of M10.5 and peptide-binding class I MHC molecules, we used a previously described algorithm [30,33] to identify residues that comprise the grooves of M10.5 and H-2Dd. Residues that line the M10.5 and H-2Dd grooves were defined as having more than 5.0 Å2 accessible surface area to a 1.4 Å probe and less than 5.0 Å2 to a 5.0 Å probe (Table 2, Figure 3A and 3B), resulting in 33 groove residues in M10.5 and 22 groove residues in H-2Dd. The additional groove residues in M10.5 that are buried in H-2Dd are divided into two clusters, one in the area surrounding the A and B pockets, and the other around the F pocket (see Figure 2A). The M10.5 A pocket is larger than its H-2Dd counterpart due to the substitution of class I MHC residues Tyr7 and Tyr171 with less-bulky threonine and cysteine residues, respectively, and a different location of the sidechain of M10.5 Arg167 as compared to H-2Dd Trp167. Arg167 was postulated to likely block peptide binding in M10.5 [8] by analogy to Arg167 in FcRn, which partially occludes the A-B pocket [32]. However, the sidechain of Arg167 in M10.5 is partially disordered, suggesting mobility that could result in a rearrangement to allow a groove occupant to access the open pocket area below the arginine sidechain. Other differences that create a larger M10.5 A pocket compared to H-2Dd involve several large aromatic residues lining the H-2Dd groove that are replaced by smaller non-aromatic residues in M10.5. These include H-2Dd Tyr7, Phe74, Trp97, Phe116, Trp167, and Tyr171, which are substituted as Thr7, Ala74, Glu97, Leu116, Arg167, and Cys171 in M10.5. Another difference between the M10.5 and H-2Dd grooves is a considerable enlargement of the M10.5 D pocket, which is continuous with the E pocket. The enlargement is caused by the replacement of H-2Dd residues Asn70 and Arg155 with alanine and glycine residues, respectively, removing the constriction that sandwiches the peptide residue at position 4 (glycine) in the H-2Dd structure [24].
Figure 3 Groove Residues in M10.5 and H-2Dd
Stereo view of residues lining the α1–α2 groove in M10.5 (A) and H-2Dd (B). Groove residues are defined as in Table 2.
Table 2 Groove Residues
The chemical character of the groove of MHC–related protein can sometimes reveal the nature of its ligand. For example, the largely hydrophobic grooves of CD1 [30] and the class Ib MHC molecules Qa-2 [22] and HLA-E [41] allow binding of lipids (CD1) and hydrophobic peptides (the class Ib proteins). By contrast, the M10.5 groove, like the grooves of H-2Dd and other classical class I MHC molecules, contains a mixture of polar and non-polar residues (Table 2). A surface potential map generated by GRASP [42] shows that the A, C, E, and F pockets of M10.5 are slightly acidic whereas the B and D pockets are uncharged (see Figure 2A). The sidechain of Arg9 (a buried valine in H-2Dd) points into the center of the M10.5 groove and forms a salt-bridge with the sidechain of Glu97. A second salt-bridge between Glu63 and Arg167 bridges the A and B pockets. The groove in H-2Dd is primarily acidic, with a similar salt bridge formed between Arg62 and Glu163 over the A–B pocket boundary (Figure 2A). Peptides bound to H-2Dd are accommodated under the salt bridge [24], thus the salt bridge in the M10.5 groove would not necessarily prevent binding of a peptide or other small molecule.
Sequence Conservation and Receptor Binding
Like classical class I MHC molecules, variability within M10 proteins is mainly localized to the α1–α2 platform, with the α3 domain being more constant [8,12]. As previously predicted [8] and now confirmed by the M10.5 structure, amino acids with the highest degree of sequence variability within the M10 and M1 families cluster on the top face of the α1 and α2 domain helices. Indeed, three of the six disordered residues (145–150) in the M10.5 α2 domain helix belong to the top 10% of variable residues within the M10 α1 and α2 domains. By contrast, M10 residues that point into the groove show less variability (Figure 4A). This pattern of variability is opposite to the pattern in classical MHC class I molecules, in which residues that display the greatest sequence variability point towards the peptide-binding groove, resulting in allelic specificity for binding peptides, whereas residues on the tops of the α1 and α2 domain helices are more conserved [43,44] (Figure 4B).
Figure 4 Sequence Conservation in M10 and H-2 Loci
(A) Ribbon diagram of the M10.5 α1–α2 platform with positions of residues that are 100% identical in the nine M10 and M1 families colored blue; residues that are among the top 10% most variable are colored red. Variability is determined by the number of amino acids at a given position divided by the frequency of the most common allele at that position. Residues 145–150, which are disordered, are designated by a dashed line.
(B) Ribbon diagram of the H-2Dd α1–α2 domain. Conserved (blue) and highly variable (red) residues are derived from an alignment of 19 H-2D alleles [80].
M10.5 Is Thermally Unstable
Classical class I MHC molecules and UL18 are thermally unstable in the absence of bound peptide [21,45,46]. In contrast, non-peptide binding class I MHC homologs, such as FcRn and the hemochromatosis protein HFE, do not show reduced thermal stability, presumably due to structural rearrangements that close the counterparts of their peptide-binding grooves [32,33,47]. The M10.5 groove is open, but apparently unoccupied, in the crystal structure, raising the question of whether M10.5 is stable in the absence of a groove occupant.
To determine the thermal stability of M10.5, we monitored heat-induced unfolding by recording the circular dichroism (CD) signal at 223 nm as a function of increasing temperature. Two unfolding transitions, an upward-sloping transition with a Tm of 43 °C and a downward-sloping transition with a Tm of 64 °C, are evident in the melting curve of insect cell-derived M10.5 (Figure 5A). The M10.5 melting curve and derived Tm values are similar to previously reported results derived from an empty form of the mouse class I MHC molecule H-2Kd complexed with human β2m [21,46]. In these studies, the first transition (Tm = 43–45 °C) represented the unfolding of the empty H-2Kd heavy chain, and the second transition (Tm = 64 °C) represented the independent unfolding of β2m subsequent to heavy chain denaturation [21,46] (Figure 5B). By analogy, we interpret the first and second transitions in the M10.5 melting curve as resulting from M10.5 and β2m denaturations, respectively. In contrast to the low thermal stability of M10.5, the peptide-filled form of the H-2Kd heavy chain melts with a Tm of approximately 57 °C [21] (Figure 5B) and FcRn melts with a Tm of approximately 62 °C at pH 6 [47]. The low thermal stability of purified M10.5 suggests that a ligand or binding partner not present in the purified preparation is necessary to stabilize the native conformation at 37 °C in vivo.
Figure 5 Thermal Denaturation Profiles
Thermal denaturation of M10.5 (A) compared with empty and peptide-filled H-2Kd (B) (modified from Figure 3A in Fahnestock et al. [21]). The CD signal at 223 nm was monitored as a function of temperature for insect cell-derived M10.5 and empty and peptide-filled versions of soluble H-2Kd produced in CHO cells. T
ms for the melting of the M10.5 heavy chain and β2m light chain (marked with arrows) were derived by estimating the half-point of the ellipticity change between the beginning and end of each transition. The M10.5 denaturation profile is similar to the profile obtained from empty H-2Kd, suggesting that the M10.5 groove is not occupied. The possibility that polyethylene glycol, a component in the crystallization solutions, might bind to and stabilize M10.5 was investigated by repeating the unfolding experiment in the presence of 20% polyethylene glycol 1000, with no significant changes to the thermal stability profile (data not shown).
M10.5 Does Not Associate with Endogenous Peptides
The empty groove in the M10.5 structure could result from expression in invertebrate cells, which do not possess the machinery to load peptides into class I MHC molecules [48,49]. When classical class I MHC molecules are expressed in vertebrate cells or purified from native sources, mixtures of short peptides derived from cytoplasmic proteins can be eluted [50]. Sequence analyses of eluted peptides revealed that any given class I allele can bind a wide variety of short (8- to 10-mer) peptides that conform to a particular allele-specific peptide-binding motif involving preferred residues at the peptide C-terminus and an internal position (usually position 2 or 5/6) [50]. To determine whether M10.5 binds endogenous or exogenous peptides with these or other characteristics, we expressed the M10.5 ectodomain in CHO cells, mammalian cells that support peptide loading, and examined acid eluates from purified M10.5 using methods previously used to identify peptides eluted from class I MHC molecules [51].
Purified M10.5-β2m heterodimers expressed in CHO and in insect cells were treated with acetic acid to dissociate potential peptide material. Insect cell-derived M10.5 was analyzed before and after incubation with a mixture of short (8- to 9-mer) peptides. As it would be impossible to test all possible 8- to 9-mer peptides, we prepared a mixture of peptides including those reported to activate V2R-expressing neurons [19] in order to see if any of these class I MHC-binding peptides bind to M10.5. We used soluble versions of two other β2m-binding class I MHC–like proteins expressed in CHO cells as controls: UL18, a viral class I MHC homolog that binds endogenous peptides resembling class I MHC–binding peptides [52], and FcRn, a class I MHC homolog that does not bind peptides or contain a non-peptidic groove occupant [32,47].
Low molecular weight acid eluates derived from M10.5 and the control proteins were sequenced by Edman degradation and analyzed by MALDI-TOF (matrix-assisted laser desorption ionization time-of-flight) mass spectrometry. Acid eluates from CHO–derived M10.5 and both peptide-treated and untreated insect cell-derived M10.5 resemble eluates from FcRn, the non–peptide-binding class I homolog, rather than eluates from UL18, a peptide-binding protein (Table 3). With the exception of cycle 1, which typically shows a high background for FcRn and other proteins that do not bind peptides [52], the total yield of the amino acids from each cycle of pool sequencing of the M10.5 acid eluates remains nearly constant and not significantly above background. By contrast, the UL18 acid eluate shows the presence of a mixture of peptides with characteristics similar to those previously reported [52]. MALDI-TOF analysis of the acid eluates indicated molecules with masses consistent with peptides in the UL18 sample, but not in the negative control or M10.5 samples (data not shown). Thus M10.5 does not appear to bind any form of peptide, including N-terminally blocked peptides, or other small molecule ligand when expressed in CHO or insect cells. In addition, purified M10.5 does not bind any of a collection of class I MHC–binding peptides, as would have been expected if M10 proteins were the molecules responsible for binding the class I MHC–binding peptides reported to activate VNO neurons [19].
Table 3 Amino Acid Yields from Acid Eluates
Computer Modeling of Potential M10.5 Groove Occupants
Given the open character of the M10.5 groove and the thermal instability of M10.5-β2m heterodimers, we believe that the groove is likely to be a binding site for a peptide, protein, or a non-peptidic small molecule ligand. Here we use computer modeling to evaluate which potential ligands can fit into the M10.5 groove.
We first consider whether the M10.5 groove can accommodate a class I MHC–binding peptide by superimposing a polyalanine version of an H-2Dd–binding peptide [24] or a Qa-2–binding peptide [22] on the M10.5 structure (see Figure 2C). The H-2Dd–binding peptide fits into most regions of the M10.5 groove, but clashes with three residues in the A and B pockets (see Table 2). The Qa-2 peptide adopts a sharper upward bend near residue P2 (data not shown), and consequently, two of these clashes (Arg9 and Phe99) do not occur. The third clash involves Arg167 with the N-terminus of the peptide, but arginine sidechains are often flexible and might be able to move out of the way of the N-terminus of a bound peptide. Several residues in the vicinity of Arg167 have smaller sidechains in M10.5 than in H-2Dd, which could open up additional space for alternative conformations of a peptide terminus. These include Thr7 (Tyr7 in H-2Dd) and Cys171 (Tyr171 in H-2Dd), as well as Arg167 (Trp167 in H-2Dd) if its sidechain can adopt an alternate conformation. We conclude that the M10.5 groove can accommodate a peptide that adopts a class I MHC–binding conformation, but that differences between the A and F pocket regions of M10.5 and classical class I MHC molecules (see Figure 2B) would require a peptide bound in the M10.5 groove to be anchored differently than a class I MHC–binding peptide.
We next considered whether longer peptides or an extended region of a protein could fit into the M10.5 groove by extending the polyalanine version of the H-2Dd–binding peptide by four alanine residues at its C-terminus. The resulting peptide was fit into the M10.5 groove with the extra residues extending out of the F pocket side of the groove (Figure 2C). The relatively open F pocket end of the M10.5 groove is able to accommodate the extra residues, suggesting that M10 proteins could bind extended regions of other proteins. Candidate M10–binding proteins include the V2Rs and/or the major urinary proteins (MUPs), which are thought to deliver small-molecule compounds to chemosensory receptor neurons [1,12]. Although the nature of the interaction is unknown, biochemical data demonstrate an association between M10 and V2R proteins [8], thus the open M10.5 groove may form a binding site for a piece of a V2R. MUPs, however, are not known to associate with M10 proteins, and a surface plasmon resonance-based binding experiment did not reveal an interaction between immobilized M10.5 and MUPs in fractionated mouse urine (data not shown; see Materials and Methods).
It is also possible that pheromones bind within the M10 groove and either are presented directly to V2R receptors or initiate a conformational change within the M10 molecule that results in the activation of the receptor. Because there are no known V2R–activating ligands, we examined pheromones known to activate V1R receptors including 2-heptanone, α- or β-farnesene, and 2,5-dimethylpyrazine [53]. Although these molecules fit easily within the M10.5 groove (data not shown), the hydrophobic nature of most pheromones is not complementary to the charged character of the groove (see Figure 2A), which is much larger than a single pheromone molecule. Smaller pheromones might associate with residues within the relatively hydrophobic M10.5 D pocket (Figure 2A), however, the rest of the groove would remain unoccupied and presumably still unstable, and it is not obvious how such an interaction would lead to activation of a V2R if it occurs through a conformational change in an M10 groove.
Discussion
The M10 family of murine class Ib MHC molecules are expressed exclusively in the VNO and appear to act as chaperones to facilitate cell surface expression of the V2R class of pheromone receptors [1,8]. Although the role of M10 molecules in V2R signaling is unclear, direct homologs of proteins typically associated with immunogenic identity such as MHC proteins provide attractive candidates to mediate MHC–disassortative mating preferences [17] and pregnancy-block phenomena [18]. Here we present a biochemical and structural analysis of M10.5, a representative M10 molecule, aimed at providing new insights into the function of M10 proteins and their association with V2Rs. As M10 proteins are related by greater than 60% sequence identity [8,12], above the approximately 30% sequence identity threshold suggesting similar three-dimensional structures [54], the results obtained for M10.5 are relevant to other M10 proteins.
Proteins with an MHC fold generally have open, occupied grooves, as in classical MHC proteins, other class Ib proteins, CD1 [30], Zn-α2-glycoprotein [28], and the protein C receptor [55], or closed, unoccupied grooves, as in FcRn [32], HFE [33], MIC-A [34], Rae-1β [56], and T22 [57]. Thus it is surprising that the M10.5 groove is open, but apparently unoccupied, in the crystal structure. Consistent with the lack of defined extra electron density for a groove occupant, analyses of acid eluates derived from M10.5 expressed in insect cells and in CHO cells failed to reveal peptidic or non-peptidic material (Table 3). In addition, M10.5 did not bind MHC–binding peptides, including peptides reported to activate VNO neurons expressing M10 family members [19]. These results do not rule out a peptide-binding role for M10 proteins because the peptides used for these experiments may not have been optimal for binding to M10.5. However, the fact that M10.5 does not associate with any endogenous peptides when expressed in CHO cells suggests that it either has more stringent criteria for peptide binding than conventional class I proteins, which bind endogenous peptides when expressed in CHO cells [21], or that it does not bind peptides at all. If M10.5 does bind peptides, changes in the M10 A and F pockets from their counterparts in class I MHC grooves (see Figure 2B) suggest that M10.5 and other M10 molecules cannot bind the same sort of peptides that are bound by class I MHC molecules. Thus it is unlikely that the class I MHC–binding peptides reported to activate VNO neurons [19] exert their effects by binding to M10 family members, consistent with the observation that purified M10.5 showed no detectable binding to peptides used in that study.
Although the ectodomains of classical class I MHC molecules can fold in the absence of peptide [21,58], and full-length empty class I proteins can reach the cell surface, they are thermally unstable and are rapidly degraded unless an appropriate binding peptide is added exogenously or the cells are grown at 26 °C [45]. Heat-induced unfolding of recombinant M10.5 reveals a thermal stability similar to empty, rather than peptide-filled, class I MHC molecules (Figure 5A and 5B). It has been assumed that the low thermal stability of empty class I MHC molecules contributes to the apparent inability of empty class I molecules to crystallize (unpublished data). Indeed, the results of molecular dynamics simulations have been used to predict that empty α1–α2 platforms do not adopt a single defined conformation [59]. However, in the case of the empty M10.5 α1–α2 platform, we are able to generate moderately well-ordered crystals, and we see a single conformation for the five copies of M10.5 in the crystallographic asymmetric unit (Figure S1). The empty platforms may be stabilized somewhat by crystal contacts, but different crystal contacts for the five M10.5 molecules do not produce different conformations of the α1–α2 platform, as evidenced by relatively low temperature factors for α1–α2 residues (see Materials and Methods).
The low thermal stability of empty recombinant M10.5 suggests that the groove is normally occupied to stabilize the protein at 37 °C. Although our results cannot be used to identify a physiologically relevant groove occupant, the structural and biochemical results can be used to determine which types of ligands are unlikely. Our analysis of the M10.5 structure and the fact that endogenous peptides were not found associated with M10.5 expressed in CHO cells (Table 3) do not support a model in which M10 proteins are stabilized by class I MHC–binding peptides from either endogenous or exogenous sources. In addition, the M10.5 groove does not have the largely hydrophobic character that would be expected if it were a binding site for hydrophobic pheromones serving as chemical cues in urine, and we observed no detectable binding signals when fractionated mouse urine was injected over purified M10.5 in a surface plasmon resonance-based binding assay (data not shown, see Materials and Methods).
A clue as to the possible nature of an M10 groove occupant comes from the different patterns of variability within the grooves of M10 and classical class I MHC proteins (Figure 4), which likely reflect the different functional roles of vomeronasal versus classical MHC molecules. Variability within the grooves of classical class I MHC molecules, which mainly involves inward-pointing groove residues [43], creates different peptide-binding preferences such that different alleles present different types of peptides that conform to allele-specific peptide-binding motifs [35]. If the grooves of M10 proteins are occupied, the relative conservation of residues that point toward the M10 groove suggests that different M10 proteins bind a more limited set of ligands than the ligands of different class I MHC alleles. The greater variability in the upward-pointing residues on the M10 helices (Figure 4A) suggests allele-specific interactions with other proteins, consistent with the suggestion of specificity in binding between different V2R and M10 proteins [8]. The variability analysis combined with the M10.5 structure suggests a potential interaction site: The disordered six-residue loop within the M10.5 α2 domain (residues 145–150) contains three highly variable residues that could be involved in an interaction with a V2R. As a precedent for a disordered region of an MHC-like structure being at a receptor binding site, a portion of the α2 domain helix that is disordered in the structure of MIC-A (residues 152 to 161, corresponding to approximately the same M10.5 residues) [34] becomes ordered in a structure of MIC-A bound to the NKG2D receptor [60].
The disordered region of the M10.5 α2 domain helix may also be related to a need for flexibility within the F pocket region of the M10 groove, which could allow binding of larger ligands than the 8- to 10-mer peptides bound in the grooves of classical class I MHC molecules. Although the M10.5 groove is about the same size as the grooves of classical class I MHC molecules (approximately 730 Å2 versus approximately 760 Å2), and the A pocket side of both types of grooves is closed, the F pocket side of the M10.5 groove is more open than the counterpart region of a class I molecule (see Figure 2). The closed ends of class I MHC grooves result in the preference for binding short (8- to 10-mer) peptides that do not extend out of either end of the groove [35]. The open F pocket end of the M10.5 groove may allow it to bind larger ligands, perhaps even an extended region of an intact protein, such as a V2R. As a precedent for this type of interaction, the open grooves of class II MHC molecules bind to an extended region with the invariant chain protein during transit through the ER and Golgi (reviewed in [61]). In this case, the invariant chain not only occludes the class II MHC groove to prevent association with ER peptides, it also serves as a chaperone to direct class II proteins to acidic compartments where they acquire peptides derived from exogenous proteins. A related situation could occur for M10 proteins, such that an extended region within the ectodomain of a V2R would bind within the groove of an M10 protein during transit through the ER and Golgi. An interaction in which an extended loop from a V2R protein binds to an M10 groove would explain why peptides or other small molecule ligands were not found in recombinant M10.5 molecules expressed in the absence of V2R proteins and why the empty recombinant molecules are thermally unstable. In this hypothesized scenario, newly synthesized M10 and V2R proteins would be stabilized through mutual interactions with a V2R loop in the M10 groove, enabling the M10 to escort the V2R to the cell surface, rationalizing the observation that M10 proteins are required for cell surface expression of V2Rs [8].
Materials and Methods
M10.5 expression and purification
A construct encoding soluble M10.5 (corresponding to the ectodomain with the preceding hydrophobic leader sequence and an upstream insect Kozak sequence (
CCTATAAAT) plus a C-terminal Factor Xa site and a 6xHis tag) was subcloned into the BamHI site of the dicistronic baculovirus transfer vector pAcUW31 (BD Biosciences Clontech, Mountain View, California, United States). The ectodomain was truncated after residue Gly299 (Gly275 using numbering corresponding to mature class I MHC heavy chains). cDNA encoding human β2m plus its hydrophobic leader sequence was inserted into the BglII site of the transfer vector. Recombinant baculovirus was generated by co-transfection of the transfer vector with linearized viral DNA (Baculogold; BD Biosciences Clontech) and supernatants were harvested from baculovirus-infected Tn5 (High Five) insect cells. Cell supernatants were buffer exchanged into 20 mM Tris (pH 7.4)/150 mM NaCl, and concentrated to 1 liter using an Amicon RA2000 tangential-flow concentrator (Amicon, Beverly, Massachusetts, United States) with a 10 kDa cutoff membrane (Pall Corporation, East Hills, New York, United States). The resulting solution was adjusted to 50 mM Tris (pH 7.4)/300 mM NaCl/10 mM imidazole/10% glycerol, and loaded onto an 8-ml Ni-NTA column (Qiagen, Valencia, California, United States) equilibrated in 50 mM Tris (pH 7.4)/300 mM NaCl/10% glycerol. The column was washed with a similar buffer containing 40 mM imidazole and the bound M10.5 eluted in 250 mM imidazole. M10.5-containing fractions were pooled, concentrated to 2 ml, and loaded onto a 16/60 Superdex 75 column (Amersham Biosciences Corp., Piscataway, New Jersey, United States) equilibrated in 20 mM Tris (pH 7.4)/150 mM NaCl/1 mM EDTA/1 mM β-mercaptoethanol. SDS-PAGE analysis of the major peak on the gel filtration column revealed two bands migrating with apparent molecular masses of 37 kDa (expected mass 32.6 kDa + carbohydrate), corresponding to the M10.5 ectodomain, and 12 kDa (expected mass 11.7 kDa), corresponding to β2m. The M10.5-β2m peak eluted at a position corresponding to a protein of approximately 47 kDa, suggesting that the protein is monomeric (i.e., a single heterodimer). Fractions containing M10.5 were pooled resulting in a yield of approximately 1.5 mg per liter of insect cell supernatant.
N-terminal sequencing of the M10.5 heavy chain was accomplished by first blotting gel-run protein to a PVDF membrane (polyvinylidene fluoride, Millipore, Billerica, Massachusetts, United States), followed by sequencing using a 492 cLC Procise protein micro-sequencer (Applied Biosystems, Foster City, California, United States). The sequence obtained, SHWLKT, corresponds to the processed mature form of M10.5. In our numbering system, the first amino acid of the processed M10.5 heavy chain is denoted as residue 2 to correspond with the numbering of the processed forms of classical class I MHC molecules.
A vector for expression of M10.5 in CHO cells was constructed by subcloning the analogous region of M10.5 into a derivative of pBJ5-GS, which carries the glutamine synthetase gene as a selectable marker and means of gene amplification in the presence of methionine sulfoximine [62]. The M10.5 expression vector was transfected into CHO cells together with a human β2m expression vector as described [33]. Selection, amplification, maintenance of methionine sulfoximine-resistant cells, and identification of M10.5-expressing cells were done as described [46]. M10.5-β2m heterodimers were purified from CHO cell supernatants as described for the insect cell-derived protein.
Crystallization and data collection
Crystals (space group P212121; a = 124.11 Å, b = 124.71 Å, c = 149.37 Å; five molecules per asymmetric unit) were grown at room temperature utilizing the hanging drop method by combining 1 μl of protein solution (approximately 8 mg/ml of insect cell-derived M10.5) with 1 μl of precipitant containing 0.1 M imidazole (pH 8.0)/20% PEG 1000/0.2 M calcium acetate. Crystals were cryopreserved in liquid nitrogen after soaking in mother liquor containing 10% glycerol as a cryoprotectant. Native datasets were collected at 100 K at beamline 12.3.1 at the Advanced Light Source (Berkeley, California, United States) and beamline 9–2 at the Stanford Synchrotron Radiation Laboratory (Stanford, California, United States). The data were indexed and scaled using the HKL suite of programs (HKL Research, Charlottesville, Virginia, United States) (see Table 1). The statistics reported in Table 1 refer to a single merged native dataset obtained by including frames from both native datasets in the scaling procedure.
Structure solution and refinement
A protein–protein BLAST search indicated that Qa-2 and H-2Dd are the closest-related proteins in sequence to M10.5 for which crystal structures are available. Molecular replacement was carried out using the CCP4 program AMORE [63,64] with an all atom version of Qa-2 (not including the bound peptide) as a search model. Molecular replacement was not successful using each dataset individually, but a solution was obtained when the data were merged into a single native dataset (see Table 1). Five molecules were located in the asymmetric unit giving an R-factor of 47.9% after rigid-body refinement.
The model was refined using all reflections to 3.0 Å, but as the data are only approximately 60% complete between 3.1 Å and 3.0 Å, the effective resolution of the structure is approximately 3.2 Å. Test set reflections (5% of total) were picked using the thin shell method in DATAMAN [65] to reduce the influence of NCS correlations on R-factor calculations. To reduce the possibility of model bias, initial averaged maps were generated using the Qa-2 structure (open groove) or the FcRn structure (closed groove) to generate model structure factors. Maps produced by both methods indicated an open M10.5 groove. The M10.5 model was built using the program O [66] into real space–averaged, annealed composite omit electron-density maps in which 5% of the molecule was omitted at a time. Simulated annealing and grouped B-factor refinement with 5-fold NCS constraints was carried out using CNS [67]. Once the R
free value stopped improving with successive cycles of refinement, the NCS constraints were relaxed to NCS restraints using a weight of 300 kcal/mol·Å2. Parts of the model that significantly deviated between NCS-related molecules or that formed packing contacts were removed from the NCS restraints. The model stereochemistry was checked after each round of refinement using PROCHECK [68] and WHAT_CHECK [69]. The model (R
cryst = 26.6%; R
free = 30.3%) includes 264 out of 274 residues in the M10.5 heavy chain and all 99 residues of β2m (Table 1). Residues 145–150 and 195–198 are not seen in the electron-density map, and 36 sidechains in the molecule 1 M10.5 heavy chain and 13 sidechains in β2m are disordered and were modeled as alanines. The average main chain temperature factors per domain for molecule 1 are 48.3 Å2 (α1–α2), 68.2 Å2 (α3), and 44.4 Å2 (β2m). The temperature factors are the highest in the distal half of the α3 domain, correlating with this region having less continuous electron density than other regions of the maps.
Groove surface area calculations were performed as previously described [30,33] using the GRASP program [42]. Alignments of the M10.5 α1- α2 domains with other α1–α2 platforms were carried out with the CCP4 program Lsqkab [63] using platform β-sheet residues 3–13, 21–28, 34–37, 46–47, 93–103, 111–118, 122–126, and 133–135. Figures were produced using Molscript [70], Raster3D [71], and PyMOL [72].
Thermal stability analyses
An AVIV 62A DS CD spectrometer with a thermoelectric cell holder and a cuvette with a 1-mm path length was used to monitor heat-induced unfolding of insect cell-derived M10.5 using samples containing 15 μM protein in 50 mM phosphate buffer. In one experiment, the protein solution also contained 20% polyethylene glycol 1000. The CD signal was monitored at 223 nm while the temperature was increased from 1 to 100°C in 1 degree increments with an equilibration time of 2 min and an averaging time of 30 s. T
ms were determined by estimating the half-point of the ellipticity change between the pure native and pure denatured states.
Acid elutions and peptide sequencing
Samples of purified M10.5 expressed in CHO and insect cells were analyzed for the presence of bound peptides as previously described [46]. In these experiments, FcRn (or HFE in an independent experiment; data not shown) served as a negative control since it had been previously established by biochemical and crystallographic methods that FcRn and HFE do not associate with endogenous peptides [32,33,47]. UL18 served as the positive control, because it associates with endogenous peptides when expressed in CHO cells [46,52]. A 20-fold molar excess of the H-2Kd–binding peptide SYIPSAEKI [73,74] was added to insect cell M10.5 followed by gel-filtration chromatography to remove unbound peptide. This was repeated twice using a mixture of peptides, each at a 10-fold molar excess compared to M10.5. One mixture contained 13 peptides, including the following peptides used in a study reporting peptide-induced activation of VNO neurons [19]: AAPDNRETF (binds to H-2Db), AAPDARETA (mutated form of first peptide), ASNENMETM (binds to H-2Db), FAPGNYPAL (binds to H-2Db), SYFPEITHI (binds to H-2Kd), SAFPEITHA (mutated form of preceding peptide), SYIPSAEKI (binds to H-2Kd), SFVDTRTLL (binds to H-2Kd), plus five other peptides: ALPHAILRL (binds to UL18) [52], TYCRTRALV (modified from an H-2Kd–binding peptide) [51,74], RGYLYQGL (binds to H-2Kb) [75], FAPGVFPYM (binds to H-2Db) [76], and ovalbumin-derived SIINFEKL (binds to H-2Kb) [77]. The other mixture contained six of the above peptides. Acid elutions and sequencing were performed by established methods [51,78,79]. Briefly, 250 μg of CHO-derived and insect cell-derived soluble M10.5 (peptide-treated and untreated), UL18, and FcRn were concentrated to 100 μl in a Centricon 3 kDa cutoff centrifugal concentrator (Amicon). After addition of 1 ml of 50 mM ammonium acetate (pH 7.5), the proteins were again concentrated to 100 μl. This washing step was repeated once, followed by the addition of 1 ml of 10% acetic acid. The samples were heated to 70 °C for 15 min, and the solutions concentrated to 100 μl. This elution step was repeated once and the resulting 2 ml of filtrate was concentrated by evaporation to 50 μl. Automated Edman degradation was performed on 10 μl using a 492 cLC Procise protein micro-sequencer (Applied Biosystems). Analysis of acid eluates was also performed using a PerSeptive Biosystems Voyager Elite MALDI-TOF mass spectrometer (PerSeptive Biosystems, Framingham, Massachusetts, United States) with delayed extraction and a high sensitivity linear detector.
Surface plasmon resonance binding assay
A Biacore 2000 biosensor system (Pharmacia-LKB Biotechnology, Uppsala, Sweden) was used to monitor interactions between M10.5 and potential ligands in mouse urine. Purified insect cell-derived M10.5 was coupled to two flow cells of a CM5 biosensor chip (Pharmacia-LKB Biotechnology) to coupling densities of 1600 and 4900 resonance units using standard amine coupling chemistry. Purified FcRn was coupled to a third flow cell to 1600 resonance units as a negative control, and a fourth flow cell was mock coupled. Urine was freshly collected from adult male and female C57BL/6 mice, and samples were frozen immediately. Male mouse urine (300–500 μl) was desalted through a G25 protein desalting spin column (Pierce Biotechnology, Rockford, Illinois, United States) prior to loading directly onto a RESOURCE Q (1 ml) high-performance ion exchange column (Amersham Biosciences). The column was then washed in 5 ml of buffer A (10 mM phosphate buffer [pH 7.0]), followed by elution in 10 ml of a linear gradient of buffer A to buffer B (10 mM phosphate buffer [pH 7.0]/300 mM NaCl) using a Vision Workstation HPLC system (Applied Biosystems). The concentration of MUPs in the HPLC fractions were estimated to be approximately 0.5 mg/ml based on Coomassie-stained SDS-PAGE gels and absorbance at 280 nm. Unfractionated urine contained approximately 100-fold higher concentrations of MUPs based on the gel staining. Whole and fractionated urine samples were diluted 10-fold into a buffer containing 50 mM HEPES (pH 7.4)/150 mM NaCl and injected at room temperature over the M10.5-, FcRn-, and mock-coupled flow cells. None of the samples showed any measurable responses aside from nonspecific refractive index changes, and no significant differences were observed between M10.5- and FcRn-coupled cells (data not shown). Although these results rule out binding between M10.5 and protein components in urine (e.g., MUPs, which are approximately 19 kDa), detection of molecules smaller than 1 kDa using a Biacore 2000 is problematic due to noise and buffer effects. As a positive control for the integrity of the coupled M10.5 proteins, a rabbit polyclonal antiserum raised against insect cell-purified M10.5 was injected over all four flow cells giving a high-affinity response only on the M10.5-coupled cells (data not shown).
Supporting Information
Figure S1 Groove Fo-Fc Electron-Density Maps for NCS-Related Molecules
(A) Sigma-A–weighted Fo-Fc maps contoured at 3.0 σ showing electron density within the α1–α2 domain. Electron density is shown for each of the five molecules in the asymmetric unit. There is no density indicative of bound peptide(s) or another type of groove occupant, except for a small peak in 5-fold averaged maps, which may result from approximately six carbons of a weakly interacting polyethylene glycol molecule in a location similar to that seen for a hexaethylene glycol molecule in the groove of Zn-α2-glycoprotein [27], which, like M10.5, was crystallized in the presence of polyethylene glycol. The peak could account for about half of a hexaethylene glycol molecule, but by contrast to the more ordered and stronger polyethylene glycol density in the Zn-α2-glycoprotein groove, the extra electron density within the M10.5 groove is weak and discontinuous. Thus if M10.5 does bind polyethylene glycol, it does not bind it in a single, ordered conformation, as observed in the structure of Zn-α2-glycoprotein [27], and probably does so to approximate a higher-affinity natural ligand.
(B) Real space–averaged Fo-Fc map using the M10.5 model refined with 5-fold NCS constraints.
(2.0 MB JPG).
Click here for additional data file.
Accession Numbers
Uniprot accession numbers (http://www.pir.uniprot.org) for proteins discussed in this paper are as follows: β2m, human (P61769), β2m, mouse (Q91XJ8), C receptor (Q9UNN8), CD1 (P11609), CD8 (P01731), FcRn (P13599), GABAB (Q9WV18), Gαi2 (P08752), Gαo (P18872), H-2Dd (P01900), H-2Kb (P01901), H-2Kd (P01902), H2-M3 (Q31093), HFE (Q6B0J5), HLA-E (P13747), M10.5 (Q85ZW7), MIC-A (Q5C9P8), NKG2D receptor (P26718), Qa-2 (P79567), Rae-1β (O08603), T22 (Q9BCZ1), TAP1 (Q62427), TAP2 (P36371), UL18 (P08560), and Zn-α2-glycoprotein (P25311).
The Research Collaboratory for Structural Bioinformatics (RCSB) Protein Data Bank accession number (http://www.rcsb.org/pdb/) for the M10.5 structure is 1ZS8. The accession numbers for the other proteins discussed in this paper are as follows: FcRn (3FRU), H-2Dd (1BII), and Qa-2 (1K8D).
Protein sequencing analyses carried out by the Protein/Peptide Microanalytical Lab were supported by the Beckman Institute at the California Institute of Technology (Caltech). We thank Kyle Lassila and Heidi Privett for help with CD measurements, Tony Giannetti and Anthony West for help with Biacore experiments, Peter Snow, Inderjit Nangiana, and Cynthia Jones at the Caltech Protein Expression Center for assisting with insect cell expression, Fabio Papes for providing fractionated urine samples, the Ralph M.Parsons Foundation for computational support, and members of the Bjorkman laboratory for critical reading of the manuscript. This work was supported by a Rosalind Alcott Post-doctoral Fellowship administered by Caltech (R.O.), the Howard Hughes Medical Institute (C.D. and P.J.B.), and the NIH/NIDCD (R01 DC003903 (C.D.)).
The Advanced Light Source is supported by the Director, Office of Science, Office of Basic Energy Sciences, Materials Sciences Division, of the U.S. Department of Energy under Contract No. DE-AC03-76SF00098 at Lawrence Berkeley National Laboratory. Portions of this research were carried out at the Stanford Synchrotron Radiation Laboratory, a national user facility operated by Stanford University on behalf of the U.S. Department of Energy, Office of Basic Energy Sciences.
Competing interests. The authors have declared that no competing interests exist.
Author contributions. RAO and PJB conceived and designed the experiments. RAO and KEHT performed the experiments. RAO, KEHT, CD, and PJB analyzed the data. CD and PJB contributed reagents/materials/analysis tools. RAO, CD, and PJB wrote the paper.
Citation: Olson R, Huey-Tubman KE, Dulac C, Bjorkman PJ (2005) Structure of a pheromone receptor-associated MHC molecule with an open and empty groove. PLoS Biol 3(8): e257.
Abbreviations
β2mβ2-microglobulin
CDcircular dichroism
CHOChinese hamster ovary
ERendoplasmic reticulum
MALDI-TOFmatrix-assisted laser desorption ionization time-of-flight
MHCmajor histocompatibility complex
MUPsmajor urinary proteins
NCSnon-crystallographic symmetry
VNOvomeronasal organ
==== Refs
References
Dulac C Torello AT Molecular detection of pheromone signals in mammals: from genes to behaviour Nat Rev Neurosci 2003 4 551 562 12838330
Leypold BG Yu CR Leinders-Zufall T Kim MM Zufall F Altered sexual and social behaviors in trp2 mutant mice Proc Natl Acad Sci U S A 2002 99 6376 6381 11972034
Stowers L Holy TE Meister M Dulac C Koentges G Loss of sex discrimination and male–male aggression in mice deficient for TRP2 Science 2002 295 1493 1500 11823606
Dulac C Axel R A novel family of genes encoding putative pheromone receptors in mammals Cell 1995 83 195 206 7585937
Matsunami H Buck LB A multigene family encoding a diverse array of putative pheromone receptors in mammals Cell 1997 90 775 784 9288756
Chandrashekar J Mueller KL Hoon MA Adler E Feng L T2Rs function as bitter taste receptors Cell 2000 100 703 711 10761935
Jones EP Kumanovics A Yoshino M Fischer Lindahl K Mhc class I and non-class I gene organization in the proximal H2-M region of the mouse Immunogenetics 1999 49 183 195 9914332
Loconto J Papes F Chang E Stowers L Jones EP Functional expression of murine V2R pheromone receptors involves selective association with the M10 and M1 families of MHC class Ib molecules Cell 2003 112 607 618 12628182
Hurt P Walter L Sudbrak R Klages S Müller I The genomic sequence and comparative analysis of the rat major histocompatibility complex Genome Res 2004 14 631 639 15060004
Garcia KC Teyton L Wilson IA Structural basis of T cell recognition Annu Rev Immunol 1999 17 369 397 10358763
Braud VM Allan DS McMichael AJ Functions of nonclassical MHC and non-MHC–encoded class I molecules Curr Opin Immunol 1999 11 100 108 10047540
Ishii T Hirota J Mombaerts P Combinatorial coexpression of neural and immune multigene families in mouse vomeronasal sensory neurons Curr Biol 2003 13 394 400 12620187
Margeta-Mitrovic M Jan YN Jan LY A trafficking checkpoint controls GABA(B) receptor heterodimerization Neuron 2000 27 97 106 10939334
Li X Staszewski L Xu H Durick K Zoller M Human receptors for sweet and umami taste Proc Natl Acad Sci U S A 2002 99 4692 4696 11917125
Nelson G Chandrashekar J Hoon MA Feng L Zhao G An amino-acid taste receptor Nature 2002 416 199 202 11894099
Potts WK Manning CJ Wakeland EK Mating patterns in seminatural populations of mice influenced by MHC genotype Nature 1991 352 619 621 1865924
Yamazaki K Boyse EA Mike V Thaler HT Mathieson BJ Control of mating preferences in mice by genes in the major histocompatibility complex J Exp Med 1976 144 1324 1335 1032893
Bruce HM An exteroceptive block to pregnancy in the mouse Nature 1959 184 105 13805128
Leinders-Zufall T Brennan P Widmayer P S PC Maul-Pavicic A MHC class I peptides as chemosensory signals in the vomeronasal organ Science 2004 306 1033 1037 15528444
Hochman JH Shimizu Y DeMars R Edidin M Specific associations of fluorescent β-2-microglobulin with cell surfaces. The affinity of different H-2 and HLA antigens for β-2-microglobulin J Immunol 1988 140 2322 2329 2450918
Fahnestock ML Tamir I Narhi L Bjorkman PJ Thermal stability comparison of purified empty and peptide-filled forms of a class I MHC molecule Science 1992 258 1658 1662 1360705
He X Tabaczewski P Ho J Stroynowski I Garcia KC Promiscuous antigen presentation by the nonclassical MHC Ib Qa-2 is enabled by a shallow, hydrophobic groove and self-stabilized peptide conformation Structure (Camb) 2001 9 1213 1224 11738047
Altschul SF Gish W Miller W Myers EW Lipman DJ Basic local alignment search tool J Mol Biol 1990 215 403 410 2231712
Achour A Persson K Harris RA Sundback J Sentman CL The crystal structure of H-2Dd MHC class I complexed with the HIV-1–derived peptide P18-I10 at 2.4 Å resolution: implications for T cell and NK cell recognition Immunity 1998 9 199 208 9729040
Kern PS Teng MK Smolyar A Liu JH Liu J Structural basis of CD8 coreceptor function revealed by crystallographic analysis of a murine CD8αα ectodomain fragment in complex with H-2Kb
Immunity 1998 9 519 530 9806638
Maenaka K Jones EY MHC superfamily structure and the immune system Curr Opin Struct Biol 1999 9 745 753 10607669
Delker SL West AP McDermott L Kennedy MW Bjorkman PJ Crystallographic studies of ligand binding by Zn-α2-glycoprotein J Struct Biol 2004 148 205 213 15477100
Sánchez LM Chirino AJ Bjorkman P Crystal structure of human ZAG, a fat-depleting factor related to MHC molecules Science 1999 283 1914 1919 10206894
Gadola SD Zaccai NR Harlos K Shepherd D Castro-Palomino JC Structure of human CD1b with bound ligands at 2.3 Å, a maze for alkyl chains Nat Immunol 2002 3 721 726 12118248
Zeng Z Castano AR Segelke BW Stura EA Peterson PA Crystal structure of mouse CD1: An MHC–like fold with a large hydrophobic binding groove Science 1997 277 339 345 9219685
Wang CR Castano AR Peterson PA Slaughter C Lindahl KF Nonclassical binding of formylated peptide in crystal structure of the MHC class Ib molecule H2-M3 Cell 1995 82 655 664 7664344
Burmeister WP Gastinel LN Simister NE Blum ML Bjorkman PJ Crystal structure at 2.2 Å resolution of the MHC–related neonatal Fc receptor Nature 1994 372 336 343 7969491
Lebrón JA Bennett MJ Vaughn DE Chirino AJ Snow PM Crystal structure of the hemochromatosis protein HFE and characterization of its interaction with transferrin receptor Cell 1998 93 111 123 9546397
Li P Willie ST Bauer S Morris DL Spies T Crystal structure of the MHC class I homolog MIC-A, a γδ T cell ligand Immunity 1999 10 577 584 10367903
Batalia MA Collins EJ Peptide binding by class I and class II MHC molecules Biopolymers 1997 43 281 302 9316393
Apostolopoulos V Yu M Corper AL Li W McKenzie IF Crystal structure of a non-canonical high affinity peptide complexed with MHC class I: a novel use of alternative anchors J Mol Biol 2002 318 1307 1316 12083519
Madden DR Gorga JC Strominger JL Wiley DC The structure of HLA-B27 reveals nonamer self-peptides bound in an extended conformation Nature 1991 353 321 325 1922337
Saper MA Bjorkman PJ Wiley DC Refined structure of the human histocompatibility antigen HLA-A2 at 2.6 Å resolution J Mol Biol 1991 219 277 319 2038058
Fremont DH Matsumura M Stura EA Peterson PA Wilson IA Crystal structures of two viral peptides in complex with murine MHC class I H-2Kb
Science 1992 257 919 927 1323877
Schwede T Kopp J Guex N Peitsch MC SWISS-MODEL: An automated protein homology-modeling server Nucleic Acids Res 2003 31 3381 3385 12824332
O'Callaghan CA Tormo J Willcox BE Braud VM Jakobsen BK Structural features impose tight peptide binding specificity in the nonclassical MHC molecule HLA-E Mol Cell 1998 1 531 541 9660937
Nicholls A Sharp KA Honig B Protein folding and association: Insights from the interfacial and thermodynamic properties of hydrocarbons Proteins 1991 11 281 296 1758883
Bjorkman PJ Saper MA Samraoui B Bennett WS Strominger JL The foreign antigen binding site and T cell recognition regions of class I histocompatibility antigens Nature 1987 329 512 518 2443855
Bjorkman PJ Parham P Structure, function, and diversity of class I major histocompatibility complex molecules Annu Rev Biochem 1990 59 253 288 2115762
Ljunggren HG Stam NJ Ohlen C Neefjes JJ Hoglund P Empty MHC class I molecules come out in the cold Nature 1990 346 476 480 2198471
Chapman TL Bjorkman PJ Characterization of a murine cytomegalovirus class I major histocompatibility complex (MHC) homolog: Comparison to MHC molecules and to the human cytomegalovirus MHC homolog J Virol 1998 72 460 466 9420246
Raghavan M Gastinel LN Bjorkman PJ The class I major histocompatibility complex related Fc receptor shows pH-dependent stability differences correlating with immunoglobulin binding and release Biochemistry 1993 32 8654 8660 8357807
Jackson MR Song ES Yang Y Peterson PA Empty and peptide-containing conformers of class I major histocompatibility complex molecules expressed in Drosophila melanogaster cells Proc Natl Acad Sci U S A 1992 89 12117 12121 1465448
Pamer E Cresswell P Mechanisms of MHC class I–restricted antigen processing Annu Rev Immunol 1998 16 323 358 9597133
Rammensee HG Falk K Rötzschke O Peptides naturally presented by MHC class I molecules Annu Rev Immunol 1993 11 213 244 8476560
Rötzschke O Falk K Deres K Schild H Norda M Isolation and analysis of naturally processed viral peptides as recognized by cytotoxic T cells Nature 1990 348 252 254 1700304
Fahnestock ML Johnson JL Feldman RM Neveu JM Lane WS The MHC class I homolog encoded by human cytomegalovirus binds endogenous peptides Immunity 1995 3 583 590 7584148
Leinders-Zufall T Lane AP Puche AC Ma W Novotny MV Ultrasensitive pheromone detection by mammalian vomeronasal neurons Nature 2000 405 792 796 10866200
Chothia C Lesk AM The relation between the divergence of sequence and structure in proteins EMBO J 1986 5 823 826 3709526
Oganesyan V Oganesyan N Terzyan S Qu D Dauter Z The crystal structure of the endothelial protein C receptor and a bound phospholipid J Biol Chem 2002 277 24851 24854 12034704
Li P McDermott G Strong RK Crystal structures of RAE-1β and its complex with the activating immunoreceptor NKG2D Immunity 2002 16 77 86 11825567
Wingren C Crowley MP Degano M Chien Y Wilson IA Crystal structure of a γδ T cell receptor ligand T22: a truncated MHC–like fold Science 2000 287 310 314 10634787
Springer S Doring K Skipper JC Townsend AR Cerundolo V Fast association rates suggest a conformational change in the MHC class I molecule H-2Db upon peptide binding Biochemistry 1998 37 3001 3012 9485452
Zacharias M Springer S Conformational flexibility of the MHC class I α1–α2 domain in peptide bound and free states: a molecular dynamics simulation study Biophys J 2004 87 2203 2214 15454423
Li P Morris DL Willcox BE Steinle A Spies T Complex structure of the activating immunoreceptor NKG2D and its MHC class I–like ligand MICA Nat Immunol 2001 2 443 451 11323699
Cresswell P Assembly, transport, and function of MHC class II molecules Annu Rev Immunol 1994 12 259 293 8011283
Bebbington CR Hentschel CCG Glover DM The use of vectors based on gene amplification for the expression of cloned genes in mammalian cells DNA cloning: A practical approach 1987 Oxford IRL Press 163 188
Collaborative Computational Project, Number 4 The CCP4 suite: Programs for protein crystallography Acta Crystallogr D Biol Crystallogr 1994 50 760 763 15299374
Navaza J Implementation of molecular replacement in AMoRe Acta Crystallogr D Biol Crystallogr 2001 57 1367 1372 11567147
Kleywegt GJ Jones TA xdlMAPMAN and xdlDATAMAN—Programs for reformatting, analysis and manipulation of biomacromolecular electron-density maps and reflection data sets Acta Crystallogr D Biol Crystallogr 1996 52 826 828 15299647
Jones TA Zou JY Cowan SW Kjeldgaard M Improved methods for building protein models in electron-density maps and the location of errors in these models Acta Crystallogr A 1991 47 110 119 2025413
Brünger AT Adams PD Clore GM DeLano WL Gros P Crystallography and NMR system: A new software suite for macromolecular structure determination Acta Crystallogr D Biol Crystallogr 1998 54 905 921 9757107
Laskowski RA MacArthur MW Moss DS Thornton JM PROCHECK: A program to check the stereochemical quality of protein structures J Appl Cryst 1993 26 283 291
Hooft RW Vriend G Sander C Abola EE Errors in protein structures Nature 1996 381 272 8692262
Kraulis PJ Molscript—a program to produce both detailed and schematic plots of protein structures J Appl Crystallogr 1991 24 946 950
Merritt EA Raster3D version 2.0. A program for photorealistic molecular graphics Acta Crystallogr D Biol Crystallogr 1994 50 869 873 15299354
DeLano WL The PyMOL user's manual 2002 San Carlos (California) Delano Scientific Available: http://www.pymol.org . Accessed 31 May 2005
Romero P Corradin G Luescher IF Maryanski JL H-2Kd –restricted antigenic peptides share a simple binding motif J Exp Med 1991 174 603 612 1714934
Fahnestock ML Johnson JL Feldman RM Tsomides TJ Mayer J Effects of peptide length and composition on binding to an empty class I MHC heterodimer Biochemistry 1994 33 8149 8158 8025120
Shibata K Imarai M van Bleek GM Joyce S Nathenson SG Vesicular stomatitis virus antigenic octapeptide N52–59 is anchored into the groove of the H-2Kb molecule by the side chains of three amino acids and the main-chain atoms of the amino terminus Proc Natl Acad Sci U S A 1992 89 3135 3139 1313583
Zhao R Loftus DJ Appella E Collins EJ Structural evidence of T cell xeno-reactivity in the absence of molecular mimicry J Exp Med 1999 189 359 370 9892618
Rötzschke O Falk K Stevanovic S Jung G Walden P Exact prediction of a natural T cell epitope Eur J Immunol 1991 21 2891 2894 1718764
Van Bleek GM Nathenson SG Isolation of an endogenously processed immunodominant viral peptide from the class I H-2Kb molecule Nature 1990 348 213 216 1700303
Jardetzky TS Lane WS Robinson RA Madden DR Wiley DC Identification of self peptides bound to purified HLA-B27 Nature 1991 353 326 329 1922338
Kabat EA Wu TT Perry H Gottesman K Foeller C Sequences of proteins of immunological interest, 5th ed, Volume 1. NIH Publ. no. 91–3242 1991 Bethesda (Maryland) US Department of Health and Human Services 1137
| 16089503 | PMC1174912 | CC BY | 2021-01-05 08:21:24 | no | PLoS Biol. 2005 Aug 12; 3(8):e257 | utf-8 | PLoS Biol | 2,005 | 10.1371/journal.pbio.0030257 | oa_comm |
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 10.1371/journal.pbio.0030280SynopsisBiophysicsImmunologyMolecular Biology/Structural BiologyNeuroscienceBiochemistryIn VitroSniffing Out the Structure of a Pheromone Receptor-Associated MHC Molecule Synopsis8 2005 12 7 2005 12 7 2005 3 8 e280Copyright: © 2005 Public Library of Science.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.
Structure of a Pheromone Receptor-Associated MHC Molecule with an Open and Empty Groove
==== Body
Although humans are thought to use their sense of smell to some extent when choosing a mate—or to a great extent, if perfume advertisements are to be believed—for many animals odor is centrally important in their mating decisions. Male moths, for example, are attracted over great distances to receptive mates by powerful species- and gender-specific cues known as pheromones.
Mammals, too, produce pheromones, although few mammalian pheromones have been unambiguously identified. These compounds have both long-lasting effects on the hormonal state of the animal receiving the signal and short-term effects on its social behavior. In rodents, pheromones are processed by a special part of the olfactory system called the vomeronasal organ (VNO), which lies between the nasal cavity and the top of the mouth.
In rats and mice, the VNO expresses two large families of genes encoding putative pheromone receptors—the V1Rs and the V2Rs. In 2003, it was discovered that for V2R receptors to be functional, they have to associate with members of the M10 and M1 families of non-classical major histocompatibility complex (MHC) class Ib molecules. Classical MHC class Ia molecules are a huge family of closely related, immunologically important molecules that present small pieces of foreign proteins (peptides) to T lymphocytes to help them recognize invading pathogens. By contrast, the smaller group of non-classical MHC molecules have both immune and non-immune functions. Intriguingly, peptides that bind to classical MHC class Ia molecules have been reported to activate V2R-expressing neurons. Furthermore, mice and rats mate preferentially with animals expressing MHC molecules different from their own.
Given all these pieces of information implicating both classical and non-classical MHC molecules in the social behavior of rodents, Pamela Bjorkman and her colleagues wanted to discover more about the relationship between these two classes of MHC molecules. To do this, they undertook a structural study of M10.5, one of the nine VNO-specific MHC class Ib proteins. They crystallized molecules of M10.5 expressed in insect cells and then used X ray crystallography to investigate how the M10.5 structure compares to that of classical MHC molecules.
The structure of a pheromone receptor-associated molecule, shown here as a ribbon model, suggests that the molecule binds to an unknown compound
Overall, the structure turned out to be very similar but there was one big surprise. MHC class Ia molecules contain a characteristic open groove, which has thus far always been occupied by a peptide in crystal structures. The analogous open groove in the M10.5 was unexpectedly empty. However, experiments showed that the empty M10.5 molecule was thermally unstable, suggesting that the groove is normally occupied.
The researchers tried several approaches to identify the mysterious M10.5 ligand(s). First, they expressed M10.5 in mammalian cells rather than insect cells—insect cells lack the cellular machinery that normally loads peptides into MHC molecules—but still no peptides bound in the M10.5 groove. Then, they provided the M10.5 molecule with a mixture of peptides known to bind to MHC class I molecules. Again, no sign of peptide binding. Finally, the researchers used computer modeling to predict potential M10.5 groove occupants. From this analysis, they concluded that M10.5 and other M10s could bind a more restricted but longer set of peptides than MHC class Ia molecules. One possibility is that the M10.5 groove provides a binding site for V2Rs, but it might also bind pheromones.
Further experiments are now needed to identify the true binding partners of M10.5 and the other MHC class Ib molecules that are expressed in the VNO. Their eventual identification should provide insights into pheromone detection and facilitate the understanding of mating preferences in rodents. As for the mating preferences of humans, researchers will have to look elsewhere to solve that mystery since we do not appear to have M10 proteins, or a VNO!
| 0 | PMC1174913 | CC BY | 2021-01-05 08:21:25 | no | PLoS Biol. 2005 Aug 12; 3(8):e280 | utf-8 | PLoS Biol | 2,005 | 10.1371/journal.pbio.0030280 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar15031589902810.1186/ar1503Research ArticleLevels of gastrin-releasing peptide and substance P in synovial fluid and serum correlate with levels of cytokines in rheumatoid arthritis Grimsholm Ola [email protected]ää-Dahlqvist Solbritt [email protected] Sture [email protected] Department of Integrative Medical Biology, Section for Anatomy, Umeå University, Umeå, Sweden2 Department of Rheumatology, Umeå University Hospital, Umeå, Sweden2005 7 2 2005 7 3 R416 R426 27 9 2004 7 12 2004 21 12 2004 4 1 2005 Copyright © 2005 Grimsholm et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is cited.
It is well known that cytokines are highly involved in the disease process of rheumatoid arthritis (RA). Recently, targeting of neuropeptides has been suggested to have potential therapeutic effects in RA. The aim of this study was to investigate possible interrelations between five neuropeptides (bombesin/gastrin-releasing peptide (BN/GRP), substance P (SP), vasoactive intestinal peptide, calcitonin-gene-related peptide, and neuropeptide Y) and the three cytokines tumour necrosis factor (TNF)-α, IL-6, and monocyte chemoattractant protein-1 in synovial fluid of patients with RA. We also investigated possible interrelations between these neuropeptides and soluble TNF receptor 1 in serum from RA patients. Synovial fluid and sera were collected and assayed with ELISA or RIA. The most interesting findings were correlations between BN/GRP and SP and the cytokines. Thus, in synovial fluid, the concentrations of BN/GRP and SP grouped together with IL-6, and SP also grouped together with TNF-α and monocyte chemoattractant protein-1. BN/GRP and SP concentrations in synovial fluid also grouped together with the erythrocyte sedimentation rate. In the sera, BN/GRP concentrations and soluble TNF receptor 1 concentrations were correlated. These results are of interest because blocking of SP effects has long been discussed in relation to RA treatment and because BN/GRP is known to have trophic and growth-promoting effects and to play a role in inflammation and wound healing. Furthermore, the observations strengthen a suggestion that combination treatment with agents interfering with neuropeptides and cytokines would be efficacious in the treatment of RA. In conclusion, BN/GRP and SP are involved together with cytokines in the neuroimmunomodulation that occurs in the arthritic joint.
cytokinesgastrin-releasing peptiderheumatoid arthritissubstance PTNF-α
==== Body
Introduction
It has long been known that neuropeptides are secreted locally during immune responses [1], and that they are involved in vasodilation and plasma extravasation, that is, neurogenic inflammation [2,3]. It is furthermore known that inflammatory cells can produce neuropeptides [4,5]. The neuropeptide substance P (SP) has been extensively examined in normal joints and in rheumatoid arthritis (RA) during the past two decades. SP has known proinflammatory properties, enhancing the proliferation of rheumatoid synoviocytes [6] and inducing the release of the proinflammatory mediator prostaglandin E2 and destructive enzymes such as collagenase [6]. An increased level of SP has been observed in the synovial fluid from RA patients [7-11].
Another neuropeptide is gastrin-releasing peptide (GRP) [12], the mammalian homologue of bombesin (BN), a tetradecapeptide originally isolated from the skin of the frog Bombina bombina [13]. BN/GRP-like peptides affect several systems in mammals, such as satiety, thermal regulation, nociception, and the activation of the sympatho-adrenomedullary outflow [14,15]. We have recently shown that BN/GRP-like peptides are present in joint fluid in arthritis and that they are increased in amount in RA [11]. Several other neuropeptides, such as vasoactive intestinal peptide (VIP), calcitonin-gene-related peptide (CGRP), and neuropeptide Y (NPY), have also been found in synovial fluid from patients with RA [10,16]. VIP is a potent anti-inflammatory agent [17,18]. It therefore has a beneficial effect in collagen-induced arthritis (CIA) [17], inhibits the production of proinflammatory factors, including tumour necrosis factor (TNF)-α [19,20] and chemokines [21], and inhibits IL-2 production in stimulated murine T lymphocytes [22].
The proinflammatory cytokine TNF-α plays a central role in the pathogenesis of RA [23,24]. Increased concentrations of this cytokine in synovial fluid from RA patients have been reported in several studies (e.g. [25-27]). It has also been shown that TNF-α is expressed locally in joints of patients with RA and in inflamed joints in experimentally induced arthritis [28,29]. Increased concentrations of soluble TNF receptors (sTNFRs) p55 (sTNFR1) and p75 (sTNFR2) have been found in peripheral blood and synovial fluid from patients with RA [30,31]. These receptors can bind TNF-α and counteract its deleterious effects [30]. Serum concentrations of the two receptors have been correlated with RA disease activity [32]. The proinflammatory cytokine IL-6 is also found in elevated concentrations in the synovial fluid of RA patients [33]. The chemokine monocyte chemoattractant protein (MCP)-1 is a cytokine that attracts monocytes to a site of inflammation [34,35]. This chemokine has been suggested to be an important factor for the recruitment of leukocytes to the joints in RA and is also found in elevated concentrations in synovial fluid from RA patients [36-38].
Interactions between neuropeptides and cytokines are well known (e.g. [39-41]), and it has been suggested that regulatory neuroimmune pathways in the joints are important mechanisms modulating the expression of the arthritis [42]. However, there is limited information on the relations between neuropeptides and cytokines in the synovial fluid and serum from patients with RA. What is known is that SP and IL-6 concentrations correlate in synovial fluid from patients with RA [43]. Furthermore, the effect of SP on IL-1, IL-6, and TNF-α production in RA was recently examined [44]. CGRP has been found to reduce the production of IL-6 and MMP-2 in RA synovial cells in vitro [18]. To the best of our knowledge, there is no information at all on the occurrence of a possible relation between BN/GRP and cytokines in arthritic patients.
The lack of information on relations between neuropeptides and cytokines in RA hampers the understanding of whether neuropeptide-modifying agents, in addition to cytokine antagonists such as TNF-α blockers, might be useful in the treatment of RA. New information has led to suggestions that treatments interfering with neuropeptides should be tested in combination with pre-existing treatments [45]. A number of neuropeptide receptor agonists and antagonists produced during recent years have been found to have effects in other situations [46-49]. Actually, using an animal model of RA, collagen-induced arthritis, it was shown that VIP affected the Th1/Th2 balance by inhibiting the synthesis of Th1 cytokines and inducing the synthesis of Th2 cytokines [17]. Also, the results of recent studies in vitro on cells of human origin have suggested that modulation of the effects of neuropeptides may be a new strategy for the treatment of arthritis [50,51]. In the present study, we therefore aimed to investigate the relation between the concentrations of three cytokines (TNFα, IL-6, and MCP-1) and several neuropeptides (BN/GRP, SP, NPY, VIP, and CGRP) in the synovial fluid from patients with RA, stratified for early and long-standing disease. We also studied the concentrations of sTNFR1 in relation to these neuropeptides in serum. We found that BN/GRP and SP showed correlations with the proinflammatory cytokines in the group with long-standing RA.
Materials and methods
Patients and material
Thirty-five consecutive patients (25 women and 10 men) with RA according to the criteria of the American Rheumatism Association [52], having active inflammation of the knee joints, were included in the study. The patients were divided into two groups, depending on their disease duration. Patients were defined as having 'early RA' (n = 7; mean age, 51 years) or 'long-standing RA' (n = 23 to 28; mean age, 59 years). Median time when sampling of blood and synovial fluid for the early RA group was 8.0 months from onset of symptoms (interquartile range = 5.6 months); the median duration from onset of symptoms until diagnosis was 6.0 months. The early RA group were thus those patients who had had RA for less than approximately 12 months, and the patients in the long-standing RA group were those patients whose disease had lasted for more than a year.
The synovial fluid was aspirated according to the method of Dixon and Emery [53]. The aspirates were centrifuged, and the leukocytes were removed. The samples were frozen and stored at -80°C until assay. Blood samples were also collected concurrently from patients whose synovial fluid was examined: patients with early RA (n = 4) or with long-standing RA (n = 22). The sera were frozen and stored at -80°C until assay. The erythrocyte sedimentation rate (ESR) was measured for all the patients.
Synovial fluid (n = 2 to10) and serum (n = 2 to 11) was also obtained from healthy controls (mean age, 39 years). The samples were collected, frozen, and stored in the same way as the material from the RA patients.
The study protocol was approved by the Ethical Committee at the Faculty of Medicine, Umeå University, and the experiments were conducted according to the principles expressed in the Declaration of Helsinki.
Enzyme-linked immunosorbent assay/enzyme immunoassay
The concentration of TNF-α in joint fluid was determined using a Human ELISA TNFα kit (Pierce-Endogen, Rockford, IL, USA) in accordance with the instructions from the manufacturer. The detection limit in this particular assay was cited as 2 pg/ml. The receptor level of sTNFR1 in serum was measured using a Human Quantikine sTNF R1 kit (R&D Systems, Wiesbaden-Nordenstadt, Germany). The minimum detectable level for this assay was 0.77 pg/ml. The concentrations of MCP-1 and IL-6 were determined with Human ELISA kits (Pierce-Endogen). The minimum detectable concentrations for these assays were <10 pg/ml and <1 pg/ml, respectively. The concentrations of the neuropeptides BN/GRP and SP (in serum) and of the neuropeptides CGRP, NPY, and VIP (in joint fluid and serum) were measured using Enzyme Immunoassay kits (Phoenix Pharmaceuticals, Belmont, CA, USA) in accordance with the supplier's assay instructions. The detection limits for the neuropeptides were as follows: BN/GRP, 300 pg/ml; CGRP, 230 pg/ml; NPY, 130 pg/ml; SP, 90 pg/ml; and VIP, 80 pg/ml.
Radio-immunoassay
For determining concentrations of BN/GRP and SP in joint fluid, RIAs were performed. These conform to the analyses described and presented in a previous study [11]. The concentrations of BN/GRP-like material in joint fluid were determined using a 125I-RIA kit (Peninsula Laboratories Inc, San Carlos, CA, USA) in accordance with the manufacturer's instructions. The detection limit in this particular assay was cited as 10 pg/ml. For the SP assay, a 125I-RIA kit (Eurodiagnostica, Malmö, Sweden) was employed. The sensitivity of the assay was 2.5 pg/ml.
Statistics
Differences between the continuous data for concentrations of the investigated substances in the two subject groups (early RA, long-standing RA) were analysed using the Mann–Whitney U test. Correlation analyses were performed with the Spearman rank correlation test. P values <0.05 were considered significant. Because the amounts of synovial fluid and serum from healthy controls varied, no statistical analysis was performed in this case.
Factor analysis, that is, explorative data analysis, was also performed to find patterns among the measured variables. In this case, because of the number of samples available, the pattern for long-standing RA was evaluated. Factor loadings >0.3 were considered. The factor analyses were performed by SPSS with Varimax, with Kaiser normalisation as a rotation method. For the statistical calculations, values below the detection limit were set to 50% of the detection limit and values beyond the highest standard point were set to the highest linearly detectable concentration.
Results
Cytokines in synovial fluid
Concentrations of immunoreactive TNF-α in synovial fluid were detected in most of the patients (approximately 70%). Table 1 shows medians and interquartile ranges of the concentrations found. No significant differences were found between the two disease groups (Mann–Whitney test). The concentrations of immunoreactive IL-6 in synovial fluid were detected in 95% of the patients with RA. Medians and interquartile ranges are shown in Table 1. The patients with early RA had significantly higher concentrations of IL-6 than the other group (P < 0.05) (Fig. 1a). The chemoattractant cytokine MCP-1 was detected in all patients, (see Table 1). No significant differences were found between the two groups for MCP-1. Values found for control patients are shown in Table 1.
Neuropeptides in synovial fluid
The neuropeptides BN/GRP, CGRP, NPY, SP, and VIP were detected in synovial fluid in 70 to 95% of the patients (Table 1). A statistically significant difference between the two groups, patients with early and those with long-standing RA, was found only for BN/GRP (P < 0.05) (Fig. 1b). SP and BN/GRP values in healthy controls were significantly lower than in the two disease groups (P < 0.001) (Fig. 1b,c). The values found for control patients are shown in Table 1.
sTNFR1 in serum
Soluble TNFR1 was measurable in sera from all subjects (Table 2; Fig 2a.) Differences between the two disease groups were close to significance (P < 0.09).
Neuropeptides in serum
The concentrations of the neuropeptides BN/GRP, CGRP, NPY, SP, and VIP in sera were measurable in 85 to 100% of the patients (Table 2). Statistically significant differences between the two groups of patients, those with early and those with long-standing RA, were found for BN/GRP (P < 0.05) (Fig. 2b) and for SP (P < 0.05). The values found for control patients are shown in Table 2.
Correlations between cytokines and neuropeptides in synovial fluid
Spearman rank correlations between the concentrations of cytokines and those of the five neuropeptides were calculated. There was a correlation between IL-6 and BN/GRP-like peptides in long-standing RA but not in early RA (Fig. 3). There was also a correlation between TNF-α and BN/GRP-like peptides in long-standing RA (rs = 0.40; P < 0.05), but not in the early RA group.
Factor analysis performed on the long-standing RA group yielded four factors with eigenvalues >1, explaining 78% of the total variation between the measured variables (Table 3). SP, BN/GRP, and IL-6 grouped together, with the larger loadings on BN/GRP and on IL-6 (Table 3) in the group with long-standing RA. There was also an interrelation between SP, TNF-α, and MCP-1. None of the other three neuropeptides (NPY, CGRP, and VIP) showed any correlation with any of the cytokines. The ESR grouped together with three neuropeptides (BN/GRP, SP, and CGRP), with the largest loadings on SP and on the ESR.
Correlations between sTNFR1 and neuropeptides in serum
Strong correlations on the Spearman rank test (rs = 0.61; P < 0.05) between concentrations of BN/GRP and those of sTNFR1 were found in the group with long-standing RA, but not in that with early RA (Fig. 4). No correlations were found between sTNFR1 concentrations and the concentrations of the other neuropeptides.
Factor analysis performed on the group with long-standing RA yielded three factors with eigenvalues >1, explaining 81% of the total variation between the measured variables (Table 4). It can be seen here that BN/GRP grouped together with sTNFR1 and CGRP. None of the other neuropeptides showed any correlation with sTNFR1. It was also seen that the ESR grouped together with sTNFR1.
Correlations between the neuropeptides in synovial fluid
Spearman rank correlations showed that the concentrations of NPY in synovial fluid correlated with those of SP (rs = 0.41; P < 0.05), VIP (rs = 0.69; P < 0.01), and CGRP (rs = 0.52; P < 0.05). VIP also correlated with CGRP (rs = 0.7; P < 0.001).
Factor analysis showed that all five neuropeptides grouped together, with the larger loadings on VIP, CGRP, and NPY (Table 3).
Correlations between the neuropeptides in serum
Spearman's rank correlations showed that the concentrations of BN/GRP in serum correlated with those of CGRP (P < 0.01; r = 0.70), VIP (P < 0.05; r = 0.51), and SP (P < 0.01; r = 0.58). Concentrations of SP also correlated with those of CGRP (P < 0.01; r = 0.54) and VIP (P < 0.001; r = 0.83). Furthermore, the concentrations of VIP correlated with the those of CGRP (P < 0.01; r = 0.54).
All the neuropeptides grouped together in the factor analysis on serum (Table 4).
Discussion
In the present study, the relations between various neuropeptides and cytokines in joint fluid and serum of RA patients were examined. Although there were, to a certain extent, correlations between the various neuropeptides examined, a major finding was that BN/GRP and SP were the neuropeptides that correlated most clearly with the cytokines, with BN/GRP showing the highest degree of correlation. We stress that the occurrence of possible relations were sought extensively, using both Spearman correlation analyses and factor analysis. We also observed that the concentrations of BN/GRP and SP in synovial fluid grouped together with the ESR. These observations suggest that BN/GRP and SP are connected to inflammation in the joint. Comparison with the values obtained from healthy controls showed clearly that in the patients in the long-standing RA group, there were increases in concentrations of the cytokines examined, as well as in concentrations of SP and BN/GRP. It is well established that TNF-α is undetectable in synovial fluid from healthy controls [23,27].
CGRP, NPY, SP, and VIP have all been shown to affect cytokine production/release from various immune cells [39,40,44,54]. For example, SP can modulate the release of IL-1, IL-6, and TNF-α from human blood monocytes [39]. The stimulatory effect of SP on these proinflammatory cytokines was due to an augmented secretion from the immunocompetent cells [41]. SP and VIP have opposite roles concerning the regulation of TNF-α production, VIP inhibiting the production of TNF-α from, for example, microglia, and SP increasing the TNF-α production from mast cells and other types of inflammatory cells [40,41]. The impact of neuropeptides on cytokine production and release in RA is not well known. However, in a recent study it was shown that CGRP, NPY, SP and VIP could affect the IL-1β, TNF-α, and IL-6 production by peripheral whole blood cells from RA patients [44]. In the present study, we searched for possible correlations of TNFα, IL-6, and MCP-1 with all neuropeptides included in the analyses of the synovial fluid, but no correlations were found concerning VIP, NPY, or CGRP. The situation was the same for correlation matrices between sTNFR1 and these three neuropeptides in serum.
Because IL-6 was found in elevated concentrations in the synovial fluid of RA patients [33,55], the importance of this cytokine in RA and other diseases has been thoroughly investigated. It has been shown that IL-6 regulates acute-phase proteins such as C-reactive protein and haptoglobin [33,55]. Furthermore, a humanised anti-IL-6 receptor antibody has been found to inhibit IL-6 significantly and to improve the signs and symptoms of RA [56]. In the analysis of synovial fluid in the present study, we found that BN/GRP and SP grouped together with IL-6, with the strongest loadings on BN/GRP and on IL-6. This is one of the major findings leading to the interpretation that these two neuropeptides correlate more closely than the others with disease activity in RA. In the synovial fluid of patients with long-standing RA, there was also a correlation between BN/GRP and TNF-α. As it is well known that TNF-α and other cytokines may occur in insufficient concentrations in serum (e.g. [57]), we analysed sTNFR1 in serum and found that the receptor concentrations correlated with concentrations of BN/GRP on the Spearman's rank analysis and that they grouped together on the factor analysis.
It has long been discussed as to whether SP is clinically significant in RA, and accordingly whether SP-receptor antagonists block joint inflammation (e.g. [58,59]). In early studies, it was shown that intra-articular injections of SP increased the severity of the arthritis [60]. This substance was also early shown to facilitate the production and release of proinflammatory mediators such as prostaglandin E2 and destructive enzymes such as collagenase [6]. In more recent studies, intra-articular injection of SP was found to lead to endothelial-cell proliferation [61]. The overall role of SP in RA has recently been extensively discussed and it was concluded that the therapeutic potential of SP receptor (neurokinin-1 receptor) antagonists should not be ignored [45].
SP is present in the nerve fibres of the synovium [62]. However, both up- [63] and down- [64] regulation of SP expression has been noted for arthritic joints. In accord with these observations, up- and down-regulations of neuropeptide expressions occur in relation to the activity of inflammatory disease in other tissues [65,66]. A reason for decreased magnitude of SP in the nerve fibres of joint capsules may be the occurrence of a local release of SP, leading to a depletion of SP-containing nerve fibres [67-69]. There are also diverging observations concerning the occurrence of increases or decreases of SP concentrations in the synovial fluid of arthritic patients (cf. [7,10,11,16,70-72]). Taking into consideration that neuropeptide and cytokine concentrations may vary depending upon the stage of arthritis for the patients examined and that differences in actual neuropeptide concentrations in serum between different patients should be interpreted with caution, an advantage of the present study is that it focuses on intraindividual comparisons of neuropeptide and cytokine concentrations.
It is not clear whether BN/GRP-like peptides are present in joint innervation, or how the release of the peptides into the synovial fluid occurs. An interesting observation is that BN/GRP-like peptides have been detected in human chondrocytes [73]. In any case, BN/GRP-like peptides are well known to have trophic and growth-promoting effects in various parts of the body. Thus, these peptides have paracrine and autocrine trophic effects in certain cancers, such as small-cell lung cancer and colon cancer [74-76], and are likely to have trophic effects on the salivary glands after irradiation [77] and to promote proliferation of fetal chondrocytes [78]. Furthermore, GRP receptors have morphogenetic properties for colon cancer cells, these being mediated via focal adhesion kinase [79]. The expression of BN/GRP-like peptides in the spinal cord is influenced by glucocorticoids [80], which may be of interest for the understanding of the effects of treatments of RA patients, many of whom are treated with glucocorticoids. Concerning inflammation, it has been shown that BN/GRP-like peptides ameliorate burn-induced intestinal inflammation by preventing the sequestration of neutrophils in the injured tissue [81] and that these peptides presumably are mediators of inflammatory reactions during pulmonary inflammation [82]. Furthermore, BN ameliorates colonic damage in experimental colitis [83], which suggests that the peptide may be involved in restoring intestinal damage in inflammatory bowel disease. Treatment with BN/GRP has also been found to have antiulcerogenic and anti-inflammatory actions, related to effects on chronic gastric ulcers and burn-induced and colitis-induced gut injury [84]. Furthermore, topical use of GRP has been found to promote cutaneous wound healing [85].
The present study is relevant because neuropeptides have been suggested to be efficacious antiarthritic drugs [17,50,51]. Thus, VIP was shown to be effective as an anti-inflammatory agent in vivo for collagen-induced arthritic mice [17,86]. Furthermore, a recent study on human cells in vitro showed that VIP can down-regulate the expression and synthesis of chemokines in synovial tissue cells and fibroblast-like synoviocytes from RA patients and that VIP inhibits the production of TNF-α after TNF-α stimulation [51]. The results obtained by Foey and collaborators [50] show that the studied neuropeptide (VIP) is not on its own a useful therapeutic agent in the treatment of RA but may be useful in combination with phosphodiesterase inhibitor. In line with this idea, it has frequently been suggested that various types of combination treatments might be effective in diseases such as RA. It was initially suggested that a combination therapy targeting two cytokines instead of one would be the treatment of choice for RA [87] and other autoimmune diseases [88]. However, in a recent clinical trial it was observed that a combination therapy with a recombinant IL-1 receptor antagonist and polyethylene-glycol-conjugated sTNFRI (etanercept) did not result in an improvement of the clinical factors of RA in comparison with etanercept alone [89]. Alternative combination treatments might in the future include treatments targeting cytokine and neuropeptide. In line with this idea, it was recently suggested that SP-receptor (neurokinin-1 receptor) antagonists might be useful in combination with pre-existing treatments in RA [45]. In a recent study, a combination of common chemotherapy and immunotherapy targeting BN/GRP-receptor with a synthetic BN/GRP antagonist was found to enhance the killing of small-cell lung cancer cells [90].
As discussed above, there is indeed evidence showing that neuropeptides can modify the secretion of proinflammatory cytokines from immunocompetent cells. That includes the situation for peripheral blood cells from patients with RA [44]. As seen in the present study, BN/GRP and SP are the peptides that show correlations with the inflammatory cytokines in RA. Thus, we tentatively suggest that the increased concentrations of BN/GRP and SP may stimulate cytokine production from immunocompetent cells in RA. The findings that neurokinin-1 receptor is not only detectable on immunocompetent cells, but also on rheumatoid synoviocytes [91] and on endothelial cells in synovitis [61], strengthen the suggestion that SP plays an important role in the rheumatoid joint. In future studies, it would therefore be of interest to investigate whether BN/GRP and its receptor are active during the rheumatic process, and whether treatments interfering with the effects of BN/GRP or SP or both, possibly in parallel with medications affecting cytokine effects, are of value in RA.
Conclusion
The neuropeptides that showed correlations with the proinflammatory cytokines in the patient material examined were BN/GRP and SP, particularly BN/GRP. The observations give new insight into further studies of neuropeptide/cytokine interrelations in RA.
Abbreviations
BN/GRP = bombesin/gastrin-releasing peptide; CGRP = calcitonin-gene-related peptide; ELISA = enzyme-linked immunosorbent assay; ESR = erythrocyte sedimentation rate; GRP = gastrin-releasing peptide; IL = interleukin; MCP-1 = monocyte chemoattractant protein-1; NPY = neuropeptide Y; RA = rheumatoid arthritis; RIA = radio-immunoassay; SP = substance P; sTNFR1 = soluble TNF receptor 1; Th = T helper (cells); TNF = tumour necrosis factor; VIP = vasoactive intestinal peptide.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
OG participated in the design of the study, carried out the immunoassays, and performed the statistical analysis. SF participated in the design, coordinated the study, and took part in the statistical analysis. OG and SF drafted the manuscript. SRD was responsible for collection of the samples, participated in the statistical analysis, contributed to the coordination of the study, and took part in discussions concerning the manuscript. All authors read and approved the final manuscript.
Acknowledgements
The authors are grateful to Ms Ulla Hedlund and Ms Lena Jonsson for excellent technical services. We also thank Ms Solveig Jonsson-Wållberg for contributing to the collection of the samples. We are also grateful to Mr Hans Stenlund for help with the statistical analysis. This study was supported by the Faculty of Medicine, Umeå University.
Figures and Tables
Figure 1 Box-whisker plots showing concentrations in synovial fluid for rheumatoid arthritis and healthy controls. (a) IL-6; (b) bombesin/gastrin-releasing peptide; (c) substance P. Boxes show the 25th and 75th percentiles; horizontal lines in boxes show the medians; whiskers show the 10th and 90th percentiles. *P < 0.05.
Figure 2 Box–whisker plots showing concentrations in serum for the patients with rheumatoid arthritis and healthy controls. (a) Soluble tumour necrosis factor receptor 1; (b) bombesin/gastrin-releasing peptide. Boxes show the 25th and 75th percentiles; horizontal lines in boxes show the medians; whiskers show the 10th and 90th percentiles. *P < 0.05.
Figure 3 Scatter plot, with linear regression line, showing correlation between BN/GRP and IL-6 in synovial fluid. Spearman's rank correlation coefficient (rs = 0.38; P < 0.05) measured, using RIA and ELISA, in synovial fluid from patients with long-standing RA. BN/GRP, bombesin/gastrin-releasing peptide; RA, rheumatoid arthritis.
Figure 4 Scatter plot, with linear regression line, showing correlation between BN/GRP and sTNFR1 in serum. Spearman's rank correlation coefficient (rs = 0.61; P < 0.05), measured, using ELISA, in serum from patients with long-standing RA. BN/GRP, bombesin/gastrin-releasing peptide; RA, rheumatoid arthritis; sTNFR1, soluble TNF receptor 1.
Table 1 Cytokine and neuropeptide concentrations in synovial fluid
Substance Early RA (pg/ml) Long-standing RA (pg/ml) P Controls (pg/ml)
Cytokine
TNF-α 8.0a 7.0c ns nm
(1.0–47) (1.0–24)
IL-6 5100b 3100d <0.05 ndf
(4500–8000) (1300–3900)
MCP-1 1200b 1500d ns 620f
(810–1800) (1200–3000)
Neuropeptide
BN/GRP 31a 17c <0.05 ndg
(21–39) (7.0–23)
CGRP 580b 450d ns ndf
(160–1500) (230–960)
NPY 190b 260e ns 180f
(65–320) (170–500)
SP 48a 58c ns ndg
(36–80) (42–77)
VIP 150b 300d ns nm
(40–200) (55–630)
Concentrations were measured by ELISA and RIA in synovial fluid from the knee joints of patients with RA and in healthy control subjects. Values are medians (and Q1–Q3). P relates to long-standing RA versus early RA (Mann–Whitney test). an = 7; bn = 6; cn = 28; dn = 23; en = 25; fn = 2; gn = 10. BN/GRP, bombesin/gastrin-releasing peptide; CGRP, calcitonin-gene-related peptide; MCP-1, monocyte chemoattractant protein-1; nd, not detectable; nm, not measured; NPY, neuropeptide Y; ns, not significant; RA, rheumatoid arthritis; SP, substance P; TNF, tumour necrosis factor; VIP, vasoactive intestinal peptide.
Table 2 Concentrations of sTNFR1 and neuropeptides in serum
Variable Early RA (pg/ml) Long-standing RA (pg/ml) P Controls (pg/ml)
sTNFR1 1500a 1900b ns 1200d
(1300–1700) (1600–2000)
BN/GRP 3000a 1700b <0.05 340e
(2400–3300) (430–2500)
CGRP 1500a 820b ns 310e
(1100–1600) (510–1600)
NPY 610a 900c ns 930f
(370–890) (700–1900)
SP 170a 410b <0.05 180g
(150–300) (320–440)
VIP 250a 280b ns 290g
(210–280) (230–340)
ESR 20a 32b ns nm
(7 – 32) (22 – 47)
Concentrations were measured using ELISA. Values are medians (and Q1–Q3). The ESR is also shown. P relates to long-standing RA versus early RA. an = 4; bn = 22; cn = 21; dsTNFR1 value in healthy controls according to the manufacturer; en = 4; fn = 2; gn = 11. BN/GRP, bombesin/gastrin-releasing peptide; CGRP, calcitonin-gene-related peptide; ESR, erythrocyte sedimentation rate; NPY, neuropeptide Y; nm, not measured; ns, not significant; RA, rheumatoid arthritis; SP, substance P; sTNFR1, soluble tumour necrosis factor receptor 1; VIP, vasoactive intestinal peptide.
Table 3 Factor analysis on cytokines and neuropeptides in synovial fluid
Variable Factor
1 2 3 4
TNF-α 0.751
IL-6 0.913
MCP-1 0.912
BN/GRP -0.436 0.623 0.327
SP 0.409 0.33 0.361 0.671
VIP 0.784
CGRP 0.715 -0.472
NPY 0.893
ESR 0.71
Principal-component analysis for cytokines and neuropeptides in synovial fluid from 23 patients with long-standing RA, and for the ESR. BN/GRP, bombesin/gastrin-releasing peptide; CGRP, calcitonin-gene-related peptide; ESR, erythrocyte sedimentation rate; MCP, monocyte chemoattractant protein-1; NPY, neuropeptide Y; RA, rheumatoid arthritis; SP, substance P; TNF, tumour necrosis factor; VIP, vasoactive intestinal peptide.
Table 4 Factor analysis on sTNFR1 and neuropeptides in serum
Variable Factor
1 2 3
sTNFR1 0.323 0.827
BN/GRP 0.324 0.778
SP 0.907
VIP 0.934
CGRP 0.63 0.575
NPY 0.925
ESR 0.897
Principal-component analysis for sTNFR1 and five neuropeptides in serum from 21 patients with long-standing RA, and for the ESR. BN/GRP, bombesin/gastrin-releasing peptide; CGRP, calcitonin-gene-related peptide; ESR, erythrocyte sedimentation rate; NPY, neuropeptide Y; RA, rheumatoid arthritis; SP, substance P; sTNFR1, soluble tumour necrosis factor receptor 1; VIP, vasoactive intestinal peptide.
==== Refs
Kaltreider HB Ichikawa S Byrd PK Ingram DA Kishiyama JL Sreedharan SP Warnock ML Beck JM Goetzl EJ Upregulation of neuropeptides and neuropeptide receptors in a murine model of immune inflammation in lung parenchyma Am J Respir Cell Mol Biol 1997 16 133 144 9032120
Gamse R Posch M Saria A Jancso G Several mediators appear to interact in neurogenic inflammation Acta Physiol Hung 1987 69 343 354 2444066
Foreman JC Peptides and neurogenic inflammation Br Med Bull 1987 43 386 400 2446695
Metwali A Blum AM Ferraris L Klein JS Fiocchi C Weinstock JV Eosinophils within the healthy or inflamed human intestine produce substance P and vasoactive intestinal peptide J Neuroimmunol 1994 52 69 78 7515901 10.1016/0165-5728(94)90164-3
Maggi CA The effects of tachykinins on inflammatory and immune cells Regul Pept 1997 70 75 90 9272619 10.1016/S0167-0115(97)00029-3
Lotz M Carson DA Vaughan JH Substance P activation of rheumatoid synoviocytes: neural pathway in pathogenesis of arthritis Science 1987 235 893 895 2433770
Cerinic MM Konttinen Y Generini S Cutolo M Neuropeptides and steroid hormones in arthritis Curr Opin Rheumatol 1998 10 220 235 9608325
Alstergren P Appelgren A Appelgren B Kopp S Lundeberg T Theodorsson E Co-variation of neuropeptide Y, calcitonin gene-related peptide, substance P and neurokinin A in joint fluid from patients with temporomandibular joint arthritis Arch Oral Biol 1995 40 127 135 7540832 10.1016/0003-9969(94)00141-W
Appelgren A Appelgren B Eriksson S Kopp S Lundeberg T Nylander M Theodorsson E Neuropeptides in temporomandibular joints with rheumatoid arthritis: a clinical study Scand J Dent Res 1991 99 519 521 1722347
Hernanz A De Miguel E Romera N Perez-Ayala C Gijon J Arnalich F Calcitonin gene-related peptide II, substance P and vasoactive intestinal peptide in plasma and synovial fluid from patients with inflammatory joint disease Br J Rheumatol 1993 32 31 35 7678534
Westermark T Rantapaa-Dahlqvist S Wallberg-Jonsson S Kjorell U Forsgren S Increased content of bombesin/GRP in human synovial fluid in early arthritis: different pattern compared with substance P Clin Exp Rheumatol 2001 19 715 720 11791645
Spindel ER Krane IM Molecular biology of bombesin-like peptides. Comparison of cDNAs encoding human gastrin-releasing peptide, human neuromedin B, and amphibian ranatensin Ann N Y Acad Sci 1988 547 10 20 3071215
Anastasi A Erspamer V Bucci M Isolation and structure of bombesin and alytesin, 2 analogous active peptides from the skin of the European amphibians Bombina and Alytes Experientia 1971 27 166 167 5544731
Pert A Moody TW Pert CB Dewald LA Rivier J Bombesin: receptor distribution in brain and effects on nociception and locomotor activity Brain Res 1980 193 209 220 6247029 10.1016/0006-8993(80)90958-0
Okuma Y Yokotani K Osumi Y Centrally applied bombesin increases nerve activity of both sympathetic and adrenal branch of the splanchnic nerves Jpn J Pharmacol 1995 68 227 230 7563982
Larsson J Ekblom A Henriksson K Lundeberg T Theodorsson E Concentration of substance P, neurokinin A, calcitonin gene-related peptide, neuropeptide Y and vasoactive intestinal polypeptide in synovial fluid from knee joints in patients suffering from rheumatoid arthritis Scand J Rheumatol 1991 20 326 335 1947895
Delgado M Abad C Martinez C Leceta J Gomariz RP Vasoactive intestinal peptide prevents experimental arthritis by downregulating both autoimmune and inflammatory components of the disease Nat Med 2001 7 563 568 11329057 10.1038/87887
Takeba Y Suzuki N Kaneko A Asai T Sakane T Evidence for neural regulation of inflammatory synovial cell functions by secreting calcitonin gene-related peptide and vasoactive intestinal peptide in patients with rheumatoid arthritis Arthritis Rheum 1999 42 2418 2429 10555038 10.1002/1529-0131(199911)42:11<2418::AID-ANR21>3.0.CO;2-7
Delgado M Munoz-Elias EJ Gomariz RP Ganea D VIP and PACAP inhibit IL-12 production in LPS-stimulated macrophages. Subsequent effect on IFNgamma synthesis by T cells J Neuroimmunol 1999 96 167 181 10337915 10.1016/S0165-5728(99)00023-5
Martinez C Delgado M Pozo D Leceta J Calvo JR Ganea D Gomariz RP Vasoactive intestinal peptide and pituitary adenylate cyclase-activating polypeptide modulate endotoxin-induced IL-6 production by murine peritoneal macrophages J Leukoc Biol 1998 63 591 601 9581803
Delgado M Ganea D Inhibition of endotoxin-induced macrophage chemokine production by vasoactive intestinal peptide and pituitary adenylate cyclase-activating polypeptide in vitro and in vivo J Immunol 2001 167 966 975 11441105
Sun L Ganea D Vasoactive intestinal peptide inhibits interleukin (IL)-2 and IL-4 production through different molecular mechanisms in T cells activated via the T cell receptor/CD3 complex J Neuroimmunol 1993 48 59 69 7901234 10.1016/0165-5728(93)90059-8
Brennan FM Maini RN Feldmann M TNF alpha – a pivotal role in rheumatoid arthritis? Br J Rheumatol 1992 31 293 298 1581770
Charles P Elliott MJ Davis D Potter A Kalden JR Antoni C Breedveld FC Smolen JS Eberl G deWoody K Regulation of cytokines, cytokine inhibitors, and acute-phase proteins following anti-TNF-alpha therapy in rheumatoid arthritis J Immunol 1999 163 1521 1528 10415055
Hidaka T Suzuki K Kawakami M Okada M Kataharada K Shinohara T Takamizawa-Matsumoto M Ohsuzu F Dynamic changes in cytokine levels in serum and synovial fluid following filtration leukocytapheresis therapy in patients with rheumatoid arthritis J Clin Apheresis 2001 16 74 81 11746532 10.1002/jca.1016
Manicourt DH Poilvache P Van Egeren A Devogelaer JP Lenz ME Thonar EJ Synovial fluid levels of tumor necrosis factor alpha and oncostatin M correlate with levels of markers of the degradation of crosslinked collagen and cartilage aggrecan in rheumatoid arthritis but not in osteoarthritis Arthritis Rheum 2000 43 281 288 10693867 10.1002/1529-0131(200002)43:2<281::AID-ANR7>3.0.CO;2-7
Saxne T Palladino MA JrHeinegard D Talal N Wollheim FA Detection of tumor necrosis factor alpha but not tumor necrosis factor beta in rheumatoid arthritis synovial fluid and serum Arthritis Rheum 1988 31 1041 1045 3136775
Ulfgren AK Lindblad S Klareskog L Andersson J Andersson U Detection of cytokine producing cells in the synovial membrane from patients with rheumatoid arthritis Ann Rheum Dis 1995 54 654 661 7677442
Marinova-Mutafchieva L Williams RO Mason LJ Mauri C Feldmann M Maini RN Dynamics of proinflammatory cytokine expression in the joints of mice with collagen-induced arthritis (CIA) Clin Exp Immunol 1997 107 507 512 9067525 10.1046/j.1365-2249.1997.2901181.x
Heilig B Wermann M Gallati H Brockhaus M Berke B Egen O Pezzutto A Hunstein W Elevated TNF receptor plasma concentrations in patients with rheumatoid arthritis Clin Investig 1992 70 22 27 1318122 10.1007/BF00422933
Cope AP Aderka D Doherty M Engelmann H Gibbons D Jones AC Brennan FM Maini RN Wallach D Feldmann M Increased levels of soluble tumor necrosis factor receptors in the sera and synovial fluid of patients with rheumatic diseases Arthritis Rheum 1992 35 1160 1169 1329774
Klimiuk PA Sierakowski S Latosiewicz R Cylwik JP Cylwik B Skowronski J Chwiecko J Circulating tumour necrosis factor alpha and soluble tumour necrosis factor receptors in patients with different patterns of rheumatoid synovitis Ann Rheum Dis 2003 62 472 475 12695163 10.1136/ard.62.5.472
Houssiau FA Devogelaer JP Van Damme J de Deuxchaisnes CN Van Snick J Interleukin-6 in synovial fluid and serum of patients with rheumatoid arthritis and other inflammatory arthritides Arthritis Rheum 1988 31 784 788 3260102
Yoshimura T Robinson EA Tanaka S Appella E Leonard EJ Purification and amino acid analysis of two human monocyte chemoattractants produced by phytohemagglutinin-stimulated human blood mononuclear leukocytes J Immunol 1989 142 1956 1962 2921521
Yoshimura T Robinson EA Tanaka S Appella E Kuratsu J Leonard EJ Purification and amino acid analysis of two human glioma-derived monocyte chemoattractants J Exp Med 1989 169 1449 1459 2926329 10.1084/jem.169.4.1449
Gong JH Ratkay LG Waterfield JD Clark-Lewis I An antagonist of monocyte chemoattractant protein 1 (MCP-1) inhibits arthritis in the MRL-lpr mouse model J Exp Med 1997 186 131 137 9207007 10.1084/jem.186.1.131
Harigai M Hara M Yoshimura T Leonard EJ Inoue K Kashiwazaki S Monocyte chemoattractant protein-1 (MCP-1) in inflammatory joint diseases and its involvement in the cytokine network of rheumatoid synovium Clin Immunol Immunopathol 1993 69 83 91 8403545 10.1006/clin.1993.1153
Koch AE Kunkel SL Harlow LA Johnson B Evanoff HL Haines GK Burdick MD Pope RM Strieter RM Enhanced production of monocyte chemoattractant protein-1 in rheumatoid arthritis J Clin Invest 1992 90 772 779 1522232
Lotz M Vaughan JH Carson DA Effect of neuropeptides on production of inflammatory cytokines by human monocytes Science 1988 241 1218 1221 2457950
Kim WK Kan Y Ganea D Hart RP Gozes I Jonakait GM Vasoactive intestinal peptide and pituitary adenylyl cyclase-activating polypeptide inhibit tumor necrosis factor-alpha production in injured spinal cord and in activated microglia via a cAMP-dependent pathway J Neurosci 2000 20 3622 3630 10804204
Cuesta MC Quintero L Pons H Suarez-Roca H Substance P and calcitonin gene-related peptide increase IL-1 beta, IL-6 and TNF alpha secretion from human peripheral blood mononuclear cells Neurochem Int 2002 40 301 306 11792459 10.1016/S0197-0186(01)00094-8
Kane DJ Lockhart J Balint PV Mann C Ferrell W McInnes IB Protective effect of sensory denervation in inflammatory arthritis evidence of regulatory neuroimmune pathways in the arthritic joint Ann Rheum Dis 2005 64 325 327 15155371 10.1136/ard.2004.022277
Arnalich F de Miguel E Perez-Ayala C Martinez M Vazquez JJ Gijon-Banos J Hernanz A Neuropeptides and interleukin-6 in human joint inflammation relationship between intraarticular substance P and interleukin-6 concentrations Neurosci Lett 1994 170 251 254 7520139 10.1016/0304-3940(94)90331-X
Hernanz A Medina S de Miguel E Martin-Mola E Effect of calcitonin gene-related peptide, neuropeptide Y, substance P, and vasoactive intestinal peptide on interleukin-1beta, interleukin-6 and tumor necrosis factor-alpha production by peripheral whole blood cells from rheumatoid arthritis and osteoarthritis patients Regul Pept 2003 115 19 24 12873794 10.1016/S0167-0115(03)00127-7
Keeble JE Brain SD A role for substance P in arthritis? Neurosci Lett 2004 361 176 179 15135922 10.1016/j.neulet.2003.12.020
Kramer MS Cutler N Feighner J Shrivastava R Carman J Sramek JJ Reines SA Liu G Snavely D Wyatt-Knowles E Distinct mechanism for antidepressant activity by blockade of central substance P receptors Science 1998 281 1640 1645 9733503 10.1126/science.281.5383.1640
de Wit R Herrstedt J Rapoport B Carides AD Guoguang-Ma J Elmer M Schmidt C Evans JK Horgan KJ The oral NK(1) antagonist, aprepitant, given with standard antiemetics provides protection against nausea and vomiting over multiple cycles of cisplatin-based chemotherapy: a combined analysis of two randomised, placebo-controlled phase III clinical trials Eur J Cancer 2004 40 403 410 14746859 10.1016/j.ejca.2003.08.028
Kidd BL Inglis JJ Vetsika K Hood VC De Felipe C Bester H Hunt SP Cruwys SC Inhibition of inflammation and hyperalgesia in NK-1 receptor knock-out mice Neuroreport 2003 14 2189 2192 14625445 10.1097/00001756-200312020-00011
Stucchi AF Shofer S Leeman S Materne O Beer E McClung J Shebani K Moore F O'Brien M Becker JM NK-1 antagonist reduces colonic inflammation and oxidative stress in dextran sulfate-induced colitis in rats Am J Physiol Gastrointest Liver Physiol 2000 279 G1298 1306 11093954
Foey AD Field S Ahmed S Jain A Feldmann M Brennan FM Williams R Impact of VIP and cAMP on the regulation of TNF-alpha and IL-10 production: implications for rheumatoid arthritis Arthritis Res Ther 2003 5 R317 328 14680506 10.1186/ar999
Juarranz MG Santiago B Torroba M Gutierrez-Canas I Palao G Galindo M Abad C Martinez C Leceta J Pablos JL Vasoactive intestinal peptide modulates proinflammatory mediator synthesis in osteoarthritic and rheumatoid synovial cells Rheumatology (Oxford) 2004 43 416 422 14657510
Arnett FC Edworthy SM Bloch DA McShane DJ Fries JF Cooper NS Healey LA Kaplan SR Liang MH Luthra HS The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis Arthritis Rheum 1988 31 315 324 3358796
Dixon A Emery P Local injection therapy in rheumatic disease 1992 4 Basle: Eular Publishers
Hernanz A Tato E De la Fuente M de Miguel E Arnalich F Differential effects of gastrin-releasing peptide, neuropeptide Y, somatostatin and vasoactive intestinal peptide on interleukin-1 beta, interleukin-6 and tumor necrosis factor-alpha production by whole blood cells from healthy young and old subjects J Neuroimmunol 1996 71 25 30 8982099 10.1016/S0165-5728(96)00118-X
Swaak AJ van Rooyen A Nieuwenhuis E Aarden LA Interleukin-6 (IL-6) in synovial fluid and serum of patients with rheumatic diseases Scand J Rheumatol 1988 17 469 474 3266031
Nishimoto N Yoshizaki K Miyasaka N Yamamoto K Kawai S Takeuchi T Hashimoto J Azuma J Kishimoto T Treatment of rheumatoid arthritis with humanized anti-interleukin-6 receptor antibody: a multicenter, double-blind, placebo-controlled trial Arthritis Rheum 2004 50 1761 1769 15188351 10.1002/art.20303
Whiteside TL Cytokine measurements and interpretation of cytokine assays in human disease J Clin Immunol 1994 14 327 339 7883860 10.1007/BF01546317
Lam FY Ferrell WR Mediators of substance P-induced inflammation in the rat knee joint Agents Actions 1990 31 298 307 1707585
Lam FY Ferrell WR Inhibition of carrageenan induced inflammation in the rat knee joint by substance P antagonist Ann Rheum Dis 1989 48 928 932 2480755
Levine JD Clark R Devor M Helms C Moskowitz MA Basbaum AI Intraneuronal substance P contributes to the severity of experimental arthritis Science 1984 226 547 549 6208609
Seegers HC Avery PS McWilliams DF Haywood L Walsh DA Combined effect of bradykinin B2 and neurokinin-1 receptor activation on endothelial cell proliferation in acute synovitis FASEB J 2004 18 762 764 14766794
Iwasaki A Inoue K Hukuda S Distribution of neuropeptide-containing nerve fibers in the synovium and adjacent bone of the rat knee joint Clin Exp Rheumatol 1995 13 173 178 7544709
Garrett NE Kidd BL Cruwys SC Tomlinson DR Changes in preprotachykinin mRNA expression and substance P levels in dorsal root ganglia of monoarthritic rats: comparison with changes in synovial substance P levels Brain Res 1995 675 203 207 7540928 10.1016/0006-8993(95)00066-Y
Konttinen YT Hukkanen M Segerberg M Rees R Kemppinen P Sorsa T Saari H Polak JM Santavirta S Relationship between neuropeptide immunoreactive nerves and inflammatory cells in adjuvant arthritic rats Scand J Rheumatol 1992 21 55 59 1373904
Watanabe T Kubota Y Muto T Substance P containing nerve fibers in rectal mucosa of ulcerative colitis Dis Colon Rectum 1997 40 718 725 9194468
Todorovic V Janic B Koko V Micev M Nikolic JA Ratkovic M Leposavic G Jankovic T Knezevic-Usaj S Milicevic Z Colonic vasoactive intestinal polypeptide (VIP) in ulcerative colitis – a radioimmunoassay and immunohistochemical study Hepatogastroenterology 1996 43 483 488 8799381
Gronblad M Konttinen YT Korkala O Liesi P Hukkanen M Polak JM Neuropeptides in synovium of patients with rheumatoid arthritis and osteoarthritis J Rheumatol 1988 15 1807 1810 3068364
Mapp PI Kidd BL Gibson SJ Terry JM Revell PA Ibrahim NB Blake DR Polak JM Substance P-, calcitonin gene-related peptide- and C-flanking peptide of neuropeptide Y-immunoreactive fibres are present in normal synovium but depleted in patients with rheumatoid arthritis Neuroscience 1990 37 143 153 1700840 10.1016/0306-4522(90)90199-E
Pereira da Silva JA Carmo-Fonseca M Peptide containing nerves in human synovium: immunohistochemical evidence for decreased innervation in rheumatoid arthritis J Rheumatol 1990 17 1592 1599 1707460
Matucci-Cerinic M Lombardi A Leoncini G Pignone A Sacerdoti L Spillantini MG Partsch G Neutral endopeptidase (3.4.24.11) in plasma and synovial fluid of patients with rheumatoid arthritis. A marker of disease activity or a regulator of pain and inflammation? Rheumatol Int 1993 13 1 4 8390712 10.1007/BF00290326
Matucci-Cerinic M Partsch G Marabini S Cagnoni M High levels of substance P in rheumatoid arthritis synovial fluid. Lack of substance P production by synoviocytes in vitro Clin Exp Rheumatol 1991 9 440 441 1718650
Marshall KW Chiu B Inman RD Substance P and arthritis: analysis of plasma and synovial fluid levels Arthritis Rheum 1990 33 87 90 1689161
Iannone F Lapadula G Neuropeptides and human osteoarthritis J Rheumatol 1998 25 386 388 9489840
Carney DN Cuttitta F Moody TW Minna JD Selective stimulation of small cell lung cancer clonal growth by bombesin and gastrin-releasing peptide Cancer Res 1987 47 821 825 3026617
Cuttitta F Carney DN Mulshine J Moody TW Fedorko J Fischler A Minna JD Bombesin-like peptides can function as autocrine growth factors in human small-cell lung cancer Nature 1985 316 823 826 2993906 10.1038/316823a0
Saurin JC Nemoz-Gaillard E Sordat B Cuber JC Coy DH Abello J Chayvialle JA Bombesin stimulates adhesion, spreading, lamellipodia formation, and proliferation in the human colon carcinoma Isreco1 cell line Cancer Res 1999 59 962 967 10029091
Aalto Y Kjorell U Henriksson R Franzen L Forsgren S Bombesin-like peptide is present in duct cells in salivary glands: studies on normal and irradiated animals Neuropeptides 1997 31 167 173 9179870 10.1016/S0143-4179(97)90086-X
Hill DJ McDonald TJ Mitogenic action of gastrin-releasing polypeptide on isolated epiphyseal growth plate chondrocytes from the ovine fetus Endocrinology 1992 130 2811 2819 1572294 10.1210/en.130.5.2811
Matkowskyj KA Keller K Glover S Kornberg L Tran-Son-Tay R Benya RV Expression of GRP and its receptor in well-differentiated colon cancer cells correlates with the presence of focal adhesion kinase phosphorylated at tyrosines 397 and 407 J Histochem Cytochem 2003 51 1041 1048 12871985
Westermark T Isaksson T Holmberg P Kjorell U Rantapaa-Dahlqvist S Forsgren S Dexamethasone treatment after adrenalectomy increases bombesin content in the rat spinal cord [abstract] Acta Physiol Scand 1999 167 A8 10571562 10.1046/j.1365-201x.1999.0600h.x
Alican I Unluer EE Yegen C Yegen BC Bombesin improves burn-induced intestinal injury in the rat Peptides 2000 21 1265 1269 11035214 10.1016/S0196-9781(00)00268-0
Lemaire I Jones S Khan MF Bombesin-like peptides in alveolar macrophage: increased release in pulmonary inflammation and fibrosis Neuropeptides 1991 20 63 72 1791927 10.1016/0143-4179(91)90041-G
Gulluoglu BM Kurtel H Gulluoglu MG Aktan AO Yegen BC Dizdaroglu F Yalin R Bombesin ameliorates colonic damage in experimental colitis Dig Dis Sci 1999 44 1531 1538 10492129 10.1023/A:1026646523284
Yegen BC Bombesin-like peptides: candidates as diagnostic and therapeutic tools Curr Pharm Des 2003 9 1013 1022 12678868 10.2174/1381612033455134
Yamaguchi Y Hosokawa K Nakatani Y Sano S Yoshikawa K Itami S Gastrin-releasing peptide, a bombesin-like neuropeptide, promotes cutaneous wound healing Dermatol Surg 2002 28 314 319 11966788 10.1046/j.1524-4725.2002.99279.x
Williams RO Therapeutic effect of vasoactive intestinal peptide in collagen-induced arthritis Arthritis Rheum 2002 46 271 273 11817602 10.1002/1529-0131(200201)46:1<271::AID-ART10039>3.0.CO;2-C
Palmblad K Cytokines and cytokine-directed intervention in experimental arthritis PhD thesis 2001 Karolinska Institute, Department of Woman and Child Health
Özenci V Immune regulation in multiple sclerosis: Cytokines and metalloproteinases PhD thesis 2000 Karolinska Institute, Division of Neurology
Genovese MC Cohen S Moreland L Lium D Robbins S Newmark R Bekker P Combination therapy with etanercept and anakinra in the treatment of patients with rheumatoid arthritis who have been treated unsuccessfully with methotrexate Arthritis Rheum 2004 50 1412 1419 15146410 10.1002/art.20221
Zhou J Chen J Mokotoff M Zhong R Shultz LD Ball ED Bombesin/gastrin-releasing peptide receptor: a potential target for antibody-mediated therapy of small cell lung cancer Clin Cancer Res 2003 9 4953 4960 14581370
Sakai K Matsuno H Tsuji H Tohyama M Substance P receptor (NK1) gene expression in synovial tissue in rheumatoid arthritis and osteoarthritis Scand J Rheumatol 1998 27 135 141 9572640 10.1080/030097498441010
| 15899028 | PMC1174935 | CC BY | 2021-01-04 16:02:36 | no | Arthritis Res Ther. 2005 Feb 7; 7(3):R416-R426 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1503 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar16891589903110.1186/ar1689Research ArticleExpression of cytokine mRNA and protein in joints and lymphoid organs during the course of rat antigen-induced arthritis Pohlers Dirk [email protected] Angela [email protected] Eberhard [email protected] Carsten B [email protected] Ernesta [email protected] Frank [email protected]äuer Rolf [email protected] Raimund W [email protected] Experimental Rheumatology Unit, Friedrich Schiller University Jena, Jena, Germany2 EUCODIS GmbH, Vienna, Austria3 Pfizer GmbH, Karlsruhe, Germany4 Swiss Institute for Asthma and Allergy Research (SIAF), Davos, Switzerland5 Institute of Clinical Immunology and Transfusion Medicine, University of Leipzig, Leipzig, Germany6 Institute of Pathology, Friedrich Schiller University Jena, Jena, Germany2005 17 2 2005 7 3 R445 R457 4 11 2004 3 12 2004 4 1 2005 11 1 2005 Copyright © 2005 Pohlers et al.;licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is cited.
Cytokine expression was assessed during antigen-induced arthritis (AIA) in synovial membrane (SM), inguinal lymph node (LN), and spleen using competitive RT-PCR and sandwich ELISA. In the SM, early elevations of IL-1β and IL-6 mRNA (by 6 hours; 450- and 200-fold, respectively) correlated with the joint swelling; a 6-fold increase in tumor necrosis factor α (TNFα) was not significant. Not only IL-2 and IFN-γ (which increased 10,000-fold and 200-fold, respectively), but also IL-5 and IL-10, increased acutely (6 hours – day 1; 3-fold and 35-fold, respectively) in the SM. In general, the protein levels in the SM for IL-1β, IL-6, TNFα, IFN-γ, IL-4, and IL-10 (increase from 4-fold to 15-fold) matched the course of mRNA expression. In the inguinal LN, there were early mRNA elevations of IL-6 (a 2.5-fold increase by 6 hours, which correlated positively with the joint swelling) and IL-2 (4-fold by 6 hours), as well as later rises of IL-4 and IL-5 (2.5- and 4-fold, respectively, by day 3). No significant elevations of the corresponding proteins in this tissue were observed, except for IL-1β (by day 6) and IL-10 (by day 1). In the spleen, there were significant mRNA elevations at 6 hours of IL-1β (1.5-fold), IL-6 (4-fold; positively correlated with the joint swelling), IFN-γ (3-fold), and IL-2 (7- to 10-fold). IL-5 and IL-10 (2- and 3-fold, respectively) peaked from 6 hours to day 3 in the spleen. Increases of the corresponding proteins were significant in comparison with day 0 only in the case of IL-2 (day 6). By day 6 (transition to the chronic phase), the mRNA for cytokines declined to or below prearthritis levels in all the tissues studied except for IL-1β in the SM and IL-6 in the spleen. AIA is thus characterized by four phenomena: early synovial activation of macrophages, T helper (Th)1-like, and Th2-like cells; late, well-segregated Th2-like responses in the inguinal LN; late, overlapping Th1-like/Th2-like peaks in the spleen; and chronic elevation of synovial IL-1β mRNA and spleen IL-6 mRNA.
==== Body
Introduction
CD4+ T helper (Th) cells and macrophages infiltrate the synovial membrane (SM) during the course of rheumatoid arthritis (RA) [1-3]. Both cell types, when activated, appear to play a central role in promoting and maintaining the disease process [4,5], for example by producing certain sets of cytokines that influence the quality and extent of the inflammatory process [6]. Cytokines, in turn, can elicit the production of tissue-degrading enzymes, a mechanism eventually involved in tissue destruction and loss of articular function [5,7].
Many studies indicate that Th cells differentiate into functionally polarized effector subpopulations, producing either Th1- or Th2-like cytokines [8,9], although this concept has recently been re-evaluated [10] in a report that focused attention on the specific role and effects of individual cytokines. The Th1/Th2 subpopulations nevertheless appear differentially involved in several human and experimental immunological disorders, exerting either proinflammatory or regulatory functions [11]. Thus far, however, the evidence as to the expression of these cytokines in human RA is relatively limited and/or contradictory [12,13] and does not provide information on the time course and organ distribution of the cytokine profiles. Indeed, an extensive study of longitudinal cytokine profiles in RA is complicated by the influence of disease phases and/or treatments [14], most of which affect proportions and functions of lymphocytes and macrophages (reviewed in [1,2]). Experimental models of arthritis are therefore well suitable for learning about the sequence of cell activation in well-characterized phases of the disorders.
Antigen-induced arthritis (AIA) in the rat [15], a severe knee monoarticular arthritis induced by intra-articular administration of methylated bovine serum albumin (mBSA) after systemic immunization, is a suitable arthritis model inasmuch as CD4+ T cells and macrophages infiltrate the SM [16] and its course consists of clearly discernible phases. The acute phase progresses within approximately 1 week into chronicity, that is, a status of low-grade inflammation with moderate joint destruction and bone formation [15]. Because of its prominently local character, AIA is a unique model for the analysis of cytokine patterns in locally driven immune responses, and also a useful counterpart for comparison with more systemic models of arthritis, such as adjuvant- and collagen-induced arthritis [17].
To elucidate the sequence and the interplay of cytokine gene activation in AIA, therefore, the mRNA expression of monokines and of Th1-like and Th2-like cytokines was analysed in the SM, regional lymph node (LN), and spleen of diseased rats by means of semiquantitative, competitive RT-PCR. To assess local translation into protein, the cytokine levels were also measured by ELISA. The analysis was carried out in mBSA-immunized animals before induction of disease (day 0), at 6 hours after injection of the arthritogen, and on days 1, 3, and 6 of the disease, the latter time point marking the transition to the chronic phase. For the sake of clarity, individual cytokines are assigned to monokine- and Th1-/Th2-like patterns, according to widely accepted schemes [8] and/or their prevalent cellular source in arthritis [2,18].
Materials and methods
Animals
Induction of arthritis
Female Lewis rats (140–190 g, age 7 to 10 weeks, Charles River Laboratories, Sulzfeld, Germany, or Animal Research Facility, Friedrich Schiller University) were immunized 21 and 14 days before induction of AIA with 2 ml (total volume) of a suspension containing 1 ml each of mBSA dissolved in PBS (500 μg/ml; Sigma, Deisenhofen, Germany), and complete Freund's adjuvant (2 mg/ml Mycobacterium tuberculosis; R37 Ra; Difco, Detroit, MI, USA) by multiple subcutaneous injections into both flanks of the animals. On day 0, AIA was induced by intra-articular injection of 100 μg mBSA in 50 μl of PBS into the right knee joint; the left knee received 50 μl of PBS and served as control. All animal studies were approved by the governmental commission for animal protection.
Scoring of arthritis
The disease course was monitored by repeated assessment of the bilateral swelling of the knee joint using a caliper (Kroeplin Längenmesstechnik, Schlüchtern, Germany). The swelling was expressed as the difference between the arthritic (right) and the control, unaffected joint (left).
Assessment of cytokine mRNA expression
Tissue sampling and preparation
Samples of knee joint SM, inguinal LN, and spleen were obtained from rats killed at five time points: day 0, 6 hours, day 1, day 3, and day 6 of AIA (four to six rats per time point). After sacrifice, tissue pieces (approximately 2 to 5 mm3) were excised, snap-frozen in 500 μl guanidinium thiocyanate buffer, and kept at -70°C until processing.
Semiquantitative RT-PCR
Frozen tissues were homogenized, mRNA was isolated, cDNA was prepared and competitive RT-PCR was performed as previously described [19]. Quantification was not done until all cDNAs had been adjusted to equal β-actin mRNA content using semiquantitative RT-PCR with a competitor fragment, which contained the sequences corresponding to the primers [20]. From the present data, there was no experimental indication for the regulation of β-actin under arthritis conditions. In addition, the consistency between values found for mRNA (normalized to β-actin) and for protein (normalized to total protein) in the SM suggests little or no regulation of β-actin in this system. An amount of 2.5 × 10-12 to 2.5 × 10-19 moles, corresponding to 1.5 × 1012 to 1.5 × 105 molecules and to a dilution from 10-3 to 10-10 of the competitor fragment, was added to each PCR assay as an internal standard. Relative quantification of specific cDNAs was carried out as previously described [19]. The highest dilution at which the competitor fragment was still detectable for each particular cytokine PCR was arbitrarily defined as 1 unit; the dilution at which cDNA was detectable and the density of the band of the resulting PCR product in agarose gels were used to express the results as multiples of 1 unit. To guarantee reproducibility of the results, PCR was performed in duplicates, which yielded comparable results.
Assessment of cytokine protein expression
Tissue sampling and preparation
In an independent AIA study, samples of knee-joint SM, inguinal LNs, and spleen were obtained from rats killed at six time points: day 0, 6 hours, day 1, day 3, and day 6 of AIA. Cytokine protein analysis was performed on five to six rats per time point. The tissue pieces were snap-frozen in 250 to 1000 μl PBS–EDTA (0.9% NaCl, 30 mM KCl, 70 mM Na2HPO4, 10 mM KH2HPO4, 10 mM ethylenediaminetetraacetic acid) containing a proteinase inhibitor cocktail (Complete®; Roche Diagnostics, Mannheim, Germany), and kept at -70°C until processing. Immediately after thawing, tissue pieces were homogenized using an Ultra Turrax and centrifuged Subsequently the supernatants were divided into aliquots and kept at -70°C.
Sandwich ELISA
Concentrations of IL-1β, IL-6, tumor necrosis factor (TNF) α, IFN-γ, IL-4, and IL-10 were determined by sandwich ELISA using the monoclonal antibodies MAB501, BAF501 and recombinant rat IL-1β for IL-1β (R&DSystems, Wiesbaden, Germany) or the respective BD OptEIA Sets for all other cytokines (BD Pharmingen, Heidelberg, Germany) in accordance with the manufacturer's recommendations.
Data were normalized to total protein levels as determined using the BCA-Assay (Pierce, Rockville, IL, USA) and expressed as ng cytokine/mg total protein.
Statistics
The nonparametric Mann–Whitney (U) test was applied to analyze differences among groups for all parameters examined. For each cytokine and time point, the levels of mRNA and protein expression were compared with baseline levels (day 0) and with the respective preceding time point. Correlations between cytokine mRNA levels and the severity of joint swelling in individual animals were assessed by means of the Spearman rank correlation test. In both cases, differences were considered statistically significant for P ≤ 0.05.
Results
Clinical parameters
In both experimental series, arthritis typically developed within 6 hours of intra-articular injection of the arthritogen mBSA and reached a peak on day 1 (Fig. 1); swelling started to decrease on day 3. However, a significantly lower joint swelling in the arthritis series used for determination of protein (data not shown) allowed only a qualitative comparison of mRNA and protein levels in the SM and the other organs.
Generally, the following phases could be distinguished: preclinical (day 0); acute (6 hours to day 3); transition to chronicity (day 6). It should be considered that animals undergoing cytokine mRNA and protein analysis before induction of AIA (day 0) were under the influence of systemic immunization with mBSA (see Materials and methods and the next paragraph for details).
Cytokine protein levels in the prearthritis phase
For both mRNA and protein, all subsequent data are presented as fold changes in relation to the cytokine expression on day 0 (i.e. after immunization, but before induction of arthritis). Whereas mRNA data are expressed as relative units and are therefore not comparable among different cytokines at any time point, cytokine protein concentrations are expressed as ng/mg total protein and are therefore suitable for comparison among cytokines. Quantitative data for day 0 of AIA are presented in Table 1.
The relative protein expression in the various organs followed nearly identical patterns and quantitatively ranked as follows:
SM: IL-1β > TNF-α > IFN-γ > IL-6 > IL-10 > IL-4
Ing LN: IL-1β > IL-2 > IFN-γ > TNF-α > IL-10 > IL-6 > IL-4
Spleen: IL-1β > IL-2 > TNF-α > IFN-γ > IL-6 > IL-10 > IL-4
In addition, the concentrations of all the cytokines studied showed the highest values in the SM (approximately 2.5 to 340 ng/mg total protein), followed by those in the inguinal LN (approximately 0.02 to 9.8 ng/mg) and spleen (approximately 0.01 to 1.0 ng/mg) on day 0 and also throughout the course of AIA (Table 1, in conjunction with Fig. 3 and Table 3, underlining the predominantly local character of AIA.
Cytokine mRNA and protein expression in the synovial membrane
IL-1β. IL-1β mRNA increased sharply by 6 hours after induction of arthritis (Fig. 2a). By days 1 and 3, the expression of this mRNA declined significantly, approaching but still significantly exceeding 'prearthritis' levels by day 6. The mRNA levels correlated positively with the degree of joint swelling in individual animals (P = 0.05; Table 2). In general, IL-1β protein levels in the SM matched the mRNA course, with a peak 6 hours after induction (Fig. 3a). In contrast to the mRNA, the protein fell significantly below prearthritis levels already by day 1 and remained at this level until day 6.
IL-6. Like IL-1β, IL-6 mRNA levels rose significantly by 6 hours (Fig. 2b), and declined significantly thereafter. On day 6 the levels of this mRNA did not significantly differ from those in the prearthritis phase. IL-6 mRNA levels correlated positively with the degree of joint swelling in individual animals (P = 0.03; Table 2). Protein expression followed the course of mRNA expression; that is, after an initial peak at 6 hours (Fig. 3b) IL-6 dropped below prearthritis levels on day 1 and thereafter.
TNF-α. TNF-α mRNA levels did not significantly change throughout the disease (although they numerically rose above levels at immunization (Fig. 2c)). However, there was a significant rise of protein at 6 hours after induction (Fig. 3c), followed by a reduction to below prearthritis values on day 1 and thereafter.
IL-2. IL-2 mRNA expression underwent a massive elevation at 6 hours, declined significantly on days 1 and 3, and disappeared by day 6 (Fig. 2d). Because of the limited quantity of SM tissue, the protein levels were not determined.
IFN-γ. IFN-γ mRNA levels were significantly increased at 6 hours and day 1 and dropped significantly by day 3 (Fig. 2e). The protein increased comcomitantly at 6 hours (Fig. 3e) but had dropped to below prearthritis levels already by day 1.
IL-4. Whereas IL-4 mRNA was not detected (Fig. 2f), IL-4 protein was expressed at detectable but low levels. This cytokine increased significantly by 6 hours after induction of AIA (Fig. 3f) and then decreased to below prearthritis values by day 1.
IL-5. IL-5 mRNA peaked moderately, but significantly, by 6 hours (Fig. 2g). The protein levels were not determined.
IL-10. IL-10 mRNA was notably increased at 6 hours and day 1 and then showed significant decreases on days 3 and 6 (Fig. 2h). The protein increased 6 hours after induction of AIA, followed by a significant decrease on day 1 and a drop to below prearthritis values on days 3 and day 6 (Fig. 3h).
Cytokine mRNA and protein expression in the inguinal lymph node
IL-1β. Neither IL-1β mRNA (Fig. 4a) nor IL-1β protein (Table 3, top) showed major changes throughout the course of AIA (with the exception of a minor, but significant, 2-fold increase of protein by day 6 in comparison with day 0).
IL-6. IL-6 mRNA peaked significantly above prearthritis levels by 6 hours after induction of AIA (Fig. 4b), but returned to baseline levels by day 1. On day 3, that is, at the end of the acute peak of AIA, the levels of IL-6 mRNA, although not significantly altered on a group basis, correlated positively with the degree of joint swelling in individual animals (P = 0.02; Table 2). No peak of protein concentrations at 6 hours was detected, but protein, too, declined significantly by day 6 in comparison with baseline levels (Table 3, top).
TNF-α. TNF-α mRNA maintained immunization levels throughout the acute phase of AIA but dropped significantly by day 6 (Fig. 4c), that is, at the transition to chronicity (Fig. 1); a parallel time course was observed for the protein, though the differences did not reach significance (Table 3, top).
IL-2. IL-2 mRNA rose significantly above immunization levels by 6 hours after the induction of AIA and gradually declined to prearthritis levels thereafter (Fig. 4d). The protein showed a similar time course, though the differences did not reach significance (Table 3, top).
IFN-γ. The levels of IFN-γ mRNA approximately doubled throughout the acute phase of AIA in comparison with prearthritis levels, though the increase was not statistically significant (Fig. 4e). The protein showed a similar time course, again without reaching significance (Table 3, top).
IL-4. IL-4 mRNA expression underwent a significant burst on day 3 of AIA, that is, at the end of the acute phase of the joint disease (Fig. 4f). The protein levels, however, despite a limited increase on day 1, showed no significant changes (Table 3, top).
IL-5. IL-5 mRNA expression paralleled that of IL-4 mRNA, also reaching a significant peak on day 3 (Fig. 4g); the expression of IL-4 and IL-5 appeared biphasic, however, as indirectly documented by a drop on day 1 in comparison with prearthritis levels (Fig. 4f,g). Protein levels for IL-5 were not determined.
IL-10. IL-10 mRNA was not detected in this organ at any time point of the disease (Fig. 4h), but IL-10 protein was detected at all time points, showing a significant increase on day 1 and a decrease to prearthritis levels on day 6 (Table 3, top).
Cytokine mRNA and protein expression in the spleen
IL-1β. IL-1β mRNA levels peaked modestly and only initially, by 6 hours (Fig. 5a). Protein levels were unchanged during the course of arthritis (Table 3, bottom).
IL-6. IL-6 underwent a progressive elevation that reached plateau levels between days 3 and 6 of AIA (Fig. 5b), when the clinical signs of synovitis were already decreasing (Fig. 1). The levels of IL-6 mRNA expression correlated positively with the degree of joint swelling (P = 0.01; Table 2). Although an elevation of protein levels was not detected between days 0 and 3, a significant increase on day 6 in comparison with day 3 partially reflected the data for the mRNA (Table 3, bottom).
TNFα. TNF-α mRNA progressively increased to a peak on day 3; however, the large variability among animals likely impeded statistical significance (Fig. 5c). After an initial significant drop at 6 hours (P = 0.05 in comparison with day 0), the protein levels reached prearthritis levels on day 1 and were slightly, but not significantly, elevated on day 6 (Table 3, bottom).
IL-2. IL-2 mRNA showed a biphasic elevation, at 6 hours in correspondence with the beginning of synovitis, and on day 3 (Fig. 5d), coincident with the late acute phase of the disease (Fig. 1). IL-2 protein remained nearly unchanged throughout AIA, with a slight, but significant, increase on day 6 in comparison with day 0 and day 3 (Table 3, bottom).
IFN-γ. IFN-γ mRNA underwent a moderate, plateau-like elevation until day 3 of AIA (Fig. 5e), that is, throughout ascending phase and acute peak of AIA (Fig. 1). As with IL-2 protein, a slight, but significant, increase of IFN protein was seen on day 6 in comparison with day 3 (Table 3, bottom).
IL-4. IL-4 mRNA was not detected throughout the course of AIA (Fig. 5f). Protein was detectable throughout all phases, though without any significant changes (Table 3, bottom).
IL-5. IL-5 mRNA underwent a gradual, moderate rise until day 3 (Fig. 5g); for this cytokine, a negative correlation with the severity of arthritis was observed on day 6 (P < 0.001; Table 2), although the group as such showed no significant IL-5 elevation on this date (Fig. 5g). Protein levels for IL-5 were not determined.
IL-10. IL-10 mRNA was significantly elevated in the spleen, but only within the acute phase of AIA (6 hours) (Fig. 5h). As with IL-5, IL-10 showed a negative correlation with the degree of joint swelling on day 6 (P = 0.04; Table 2). There were no significant changes of IL-10 protein levels throughout the course of AIA (Table 3, bottom).
There was no significant correlation between the clinical time course and the protein levels for any cytokine in any organ at any time point.
Discussion
The present study documents that, in AIA, elevation of IL-1β in the SM as well as elevations of IL-6 mRNA in SM, inguinal LN, and spleen correlate positively with the disease severity and that the rise of Th1-like cytokines in the SM is massive in this model, but that this rise clearly overlaps with early Th2-like responses, as has also been shown by immunohistochemistry in the respective mouse model [21].
Local compartment
At early stages of AIA, the elevation of mRNA for IL-1β and IL-6 is very prominent at the primary site of pathology (SM 450- to 200-fold >> LN and spleen 1.5- to 10-fold; Figs 2, 4, and 5); this is very consistent with the prominently local character of AIA, induced by direct intra-articular injection of the arthritogen mBSA [15]. The levels of gene activation for these cytokines in the SM correlate positively with the clinical severity of AIA, as has also been reported for IL-1β and IL-6 protein in murine or rabbit AIA [5,22]. In the systemic rat adjuvant arthritis, in contrast, IL-1β is far more elevated in the LN than in the SM [19], in line with the different mode of induction of the disease, which favors spread of the arthritogen to the regional LNs [23]. Both the pathology and the sequence of macrophage immigration in the inflamed SM are well characterized in AIA [[24,25], and our own observation]. However, the early profiles of IL-1β and IL-6 mRNA do not match these kinetics. Similarly, there is no obvious relationship with the distribution of macrophage subpopulations, as identified by ED1, ED2, or ED3 markers [26]. It must be considered that the normal SM in the rat contains a number of resident macrophages [27] that, once activated, could be responsible for the early production of monokines in AIA; these monokines may initiate the inflammatory response and promote further cell immigration [28].
AIA synovitis is accompanied by a nonsignificant 6-fold local elevation of TNF-α message (Fig. 2c), but a significantly increased protein production (5-fold; peak level approximately 400 ng/mg total protein; Fig. 3c). In comparison with other cytokines such as IL-6 (200-fold increase of mRNA; 4-fold increase of protein; peak level approximately 80 ng/mg total protein), TNF-α reached higher protein levels despite lower fold changes of mRNA, possibly because of a more efficient translation mechanism. Such discrepancies between mRNA and protein expression have also been found in another study, using streptococcal-cell-wall-induced arthritis [29]: a 100,000-fold mRNA increase for IL-6 resulted in only a 1000-fold protein elevation, whereas the increases for TNF-α were 3.5-fold and 2-fold for mRNA and protein, respectively.
In the SM, there were significant elevations of both IFN-γ and IL-2 mRNA in the acute phase (Figs 2d,e; also confirmed for IFN-γ protein in Fig. 3e), thus indicating a complete Th1-like response at a local level. The burst of these cytokines is early and short-lasting, representing therefore a marker of very acute disease, and supporting the concept that anti-IL-2- or anti-IFN-γ treatments are of potential therapeutic use if performed at the outbreak of disease [30,31]. The marked elevation of both IFN-γ and IL-2 mRNA at the primary site of pathology, in contrast to the modest elevation of these lymphokines at a regional and systemic level, is consistent with the pre-eminently local character of AIA. Notably, this profile is practically the opposite to that of systemic adjuvant arthritis, in which early IFN-γ mRNA elevation is prominent in the regional LN draining the injection site of the arthritogen, but very modest in the SM [19]. Strong Th1-like responses, therefore, appear limited to the site of antigen injection, occurring only upon massive exposure of the local immune system to the antigen. Whether the early IL-2 mRNA peak in the SM of AIA rats also contributes to the development of regulatory T cells remains to be determined [32].
Overlapping the Th1-like response, there were significant elevations of the Th2 cytokines IL-4 (protein only; Fig. 3f), IL-5 (mRNA; Fig. 2g; protein not determined), and IL-10 (Figs 2h and 3h). Of particular interest, IL-10 in the SM peaks and then progressively declines until chronicity ensues (Figs 2h and 3h). The role of IL-10 in arthritis clearly seems a protective one, as indicated by studies on the amelioration of collagen-induced arthritis by administration of IL-10 or its augmentation in IL-10 knockout mice [33,34]. IL-10 is a Th2-like cytokine produced not only by Th1 and Th2 cells, but also (and perhaps predominantly) by macrophages, probably as an autocrine factor of immune regulation (reviewed in [8]). The very early rise of IL-10 in the SM supports this view, because in this organ it coincides with massive macrophage activation (Figs 2 and 3), which is probably due to locally injected, arthritogenic mBSA.
The acute rise of IL-4 in the SM was detected only with regard to protein, possibly due to the extremely sensitive ELISA-Kit (detection limit 0.1 pg/ml) capable of detecting very low amounts of this cytokine in the SM (2.5 ng/mg total protein). The early expression of IL-4 seems to be important to counteract the dramatic inflammatory response in acute arthritis, as is shown by a protective effect of IL-4 administration in the induction phase of CIA [35]. However, lower acute responses but disease-promoting effects in the chronic phase of CIA have been reported in IL-4-/- mice [36], showing a phase-dependent, dual role of this cytokine.
The acute rise of IL-5 mRNA in the SM could be IL-4-dependent, based on the fact that IL-4 is a driving force in many Th2-like responses [8]. Of note, early IL-5 rises have also been observed in adjuvant arthritis [19], and, more importantly, in Biozzi mice susceptible (but not in mice resistant) to collagen-induced arthritis [37]. Timed IL-5- or anti-IL-5 treatments are therefore needed to address the proinflammatory or anti-inflammatory role of IL-5 in the acute phase of AIA.
Systemic compartments
In spite of its prominently local character, AIA is accompanied by a peak of IL-6 mRNA in the inguinal LN at 6 hours (Fig. 4b), and an elevation of this mRNA in the spleen (6 hours, day 3, and day 6; Fig. 5b), in the latter case significantly correlated with the severity of disease (Table 2). This rise is maintained throughout the chronic phase, similarly to IL-6 protein in the serum of AIA [38] and adjuvant arthritis rats [39], and in analogy to findings in the synovial fluid of RA patients [40]. Whereas the acute rise of LN/spleen IL-6 is consistent with the acute-phase response typical of early AIA [41], the contribution of IL-6 to chronicity remains uncertain [42,43], perhaps because it can be produced not only by macrophages but also by Th2 cells [8,39].
TNF-α mRNA can be clearly documented in the regional LN and spleen (Figs 4c and 5c), in temporal coincidence with the severity of the joint swelling (Fig. 1), though with high variability from animal to animal, resulting in a lack of statistical significance. In AIA, therefore, the role of systemic TNF-α appears marginal, at least in relation to other cytokines, as has also been shown by the fact that successful anti-CD4 therapy reduced IL-6, but not TNF-α levels in local and systemic compartments [5]. This is at odds with more systemic models of arthritis, such as collagen-induced and adjuvant arthritis [19,44], in which there is highly significant TNF-α elevation in LN and/or spleen. Spleen TNF-α, in particular, is significantly correlated with the severity of the wasting syndrome in adjuvant arthritis [45]. The modesty of spleen TNF-α changes in AIA (Fig. 5c), therefore, is consistent with the lack of a wasting syndrome in this model [17].
Both inguinal LN and spleen showed an increase of IL-2 mRNA in the acute phase of AIA (Figs 4d and 5d), similar to a significant rise of IFN-γ in the spleen (Fig. 5e). This is clearly in line with the concept of Th1 dominance in AIA, as is also emphasized by the reduced levels of these Th1 cytokines in spleen and LN upon successful anti-CD4 treatment of AIA in mice [4]. The second elevation of IL-2 and IFN-γ in the spleen just before the transition to chronicity is similar to the situation in rat adjuvant arthritis [19], and suggests that recruitment of fresh, possibly disease-controlling, regulatory T cells [17,32] may have a systemic component even in the predominantly local AIA.
Although IL-4 mRNA underwent no changes in the spleen, it rose significantly in the inguinal LN on day 3 (Fig. 4f). Furthermore, there was a significant elevation of IL-5 mRNA in spleen and inguinal LN on day 3. Both findings are consistent with the transition to a regulatory phase of T-cell function in anticipation of chronicity. This time point may therefore represent the turning point of the disease, when LN-generated Th2-like responses may gradually replace Th1-like processes, or, according to the present data, when the Th1/Th2-like balance shifts in favor of Th2-like patterns [46]. The lack of significant elevation of Th2-like cytokines in lymphoid organs at the protein level may be due to the relatively weak clinical arthritis in this experimental series.
Overlap of Th1-like and Th2-like responses
Besides somewhat obvious Th2-like elevations in advance of chronicity (see above), in line with the expected regulatory properties of this group of cytokines [8,47,48], clear Th2-like peaks markedly overlap with the initial Th1-like surge. In the LN, the early Th2-like rise may include not only IL-5 and IL-10, but also IL-6, inasmuch as this cytokine can be produced by Th2-like cells [8] and not only by macrophages, which are modestly activated in this organ (Fig. 4a–c). These findings document that a sharp Th1/Th2 division, valid for some other systems [8], does not automatically apply to in vivo models of autoimmunity [10], including other models of arthritis [19,36]. The biological relevance of the early Th2-like rise in the SM and at systemic sites remains however unclear, that is, whether it contributes to inflammation or rather represents an attempt to limit the acute inflammatory insult [8].
The Th1-like/Th2-like responses overlap with some degree of anatomical segregation. While in the SM and spleen the expression of mRNA for IL-2 and IFN-γ overlaps with that of IL-5 and IL-10, the inguinal LN shows an overlap of mRNA for IL-2 with IL-4 and IL-5. A clear anatomical segregation of Th1-like/Th2-like responses, although in different patterns, is seen also in rat adjuvant arthritis [19]. The anatomic location of these responses varies, however, in the two models; the knee monoarticular arthritis in AIA seems to require a more regionally confined crosstalk with the inguinal LN, whereas the systemic adjuvant-induced polysynovitis requires a much stronger involvement of the spleen, the organ in which potentially regulatory T-cell cytokine responses appear generated.
Conclusion
The present study documents that the course of AIA is characterized by organ-specific overlaps of Th1-like and Th2-like responses. Activation of synovial macrophages and T cells is prominent in this prevalently local model of arthritis, although regional and systemic factors may also contribute to the disease processes. The cytokine patterns share some features with systemic adjuvant arthritis, although there are several clear differences, conceivably imputable to pathogenetic differences between the two models.
Abbreviations
AIA = antigen-induced arthritis; ELISA = enzyme-linked immunosorbent assay; IFN = interferon; IL = interleukin; LN = lymph node; mBSA = methylated bovine serum albumin; PBS = phosphate-buffered saline; RA = rheumatoid arthritis; RT-PCR = reverse transcriptase polymerase chain reaction; SM = synovial membrane; Th = T-helper; TNF = tumor necrosis factor.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
DP performed the assessment and analysis of the protein data and participated in writing the manuscript. AS, EB, and CBSW assessed and analyzed the mRNA data. EPK critically read and edited the manuscript. FE and RB participated in the design and coordination of the study, including the animal experiments. RWK contributed to the design of the study, including the animal experiments, and participated in the layout, writing, and finalization of the manuscript. All authors read and approved the final manuscript.
Acknowledgements
We thank Birthe Müller, Renate Stöckigt, and Freya Rost for excellent technical assistance, Dr H Schädlich for valuable advice, and Prof K von der Mark and Prof JR Kalden for generous support. Dr S Kunkel is gratefully aknowledged for critical reading of the manuscript. Prof F Emmrich, Prof R Bräuer, Prof RW Kinne, and Dr D Pohlers were supported by the Bundesministerium für Bildung und Forschung (BMBF; FKZ 01VM 8702, 01VM 9311, 01ZZ9602, and 01ZZ0105) and the Deutsche Forschungsgemeinschaft (DFG; Br 1372/5, Ki 439/6). Dr CB Schmidt-Weber and Dr E Buchner were supported by the Graduiertenkolleg Erlangen, Germany.
Figures and Tables
Figure 1 Time course of knee-joint swelling in rats with AIA used to evaluate cytokine mRNA. Values are means (n = 4 to 6); vertical bars indicate the standard error of the mean. The disease is characterized by rapid onset of acute inflammation, a peak on day 1, and a transition to chronicity on day 6. **P ≤ 0.01 in comparison with day 0; ++P ≤ 0.01 in comparison with the preceding date. AIA, antigen-induced arthritis.
Figure 2 Expression of mRNA of various cytokines in the synovial membrane of rats with AIA. Expression of mRNA was assessed before the induction of antigen-induced arthritis (AIA) (day 0) or afterwards (at 6 hours and on days 1, 3, and 6). Data were originally expressed as arbitrary units (1 unit = highest detectable dilution of the competitor fragment) and then related to the value of day 0 (fold change). Values are means;vertical bars indicate the standard error of the mean (n = 4 to 6 rats for each time point). Each determination was performed in duplicate. *P ≤ 0.05, **P ≤ 0.01 in comparison with day 0; +P ≤ 0.05, ++P ≤ 0.01 in comparison with the preceding date.
Figure 3 Expression of proteins for various cytokines in the synovial membrane of rats with AIA. Expression of protein was assessed before the induction of antigen-induced arthritis (AIA) (day 0) or afterwards (at 6 hours and on days 1, 3, and 6). Data were originally expressed as ng/mg total protein and then related to values of day 0 (fold change; peak values in ng/mg total protein are indicated for 6 hours). Values are means; vertical bars indicate the standard error of the mean (n = 5 to 6 rats for each time point). *P ≤ 0.05 in comparison with day 0; +P ≤ in comparison with the preceding date.
Figure 4 Expression of mRNA for various cytokines in the inguinal lymph nodes of rats with AIA. Time course and other details as in Fig. 2. AIA, antigen-induced arthritis.
Figure 5 Expression of mRNA for various cytokines in the spleens of rats with AIA. Time course and details as in Fig. 2. AIA, antigen-induced arthritis.
Table 1 Cytokine protein per total protein (ng/mg) on day 0 of antigen-induced arthritis in rats
Cytokine Synovial membrane Inguinal lymph node Spleen
IL-1β 338.60 (90.78) 9.81 (2.61) 0.983 (0.117)
IL-6 20.26 (5.77) 0.10 (0.03) 0.015 (0.006)
TNF-α 80.46 (15.48) 0.71 (0.19) 0.041 (0.004)
IL-2 n.d. 1.73 (0.54) 0.054 (0.012)
IFN-γ 43.11 (9.83) 0.74 (0.21) 0.028 (0.004)
IL-4 2.47 (0.62) 0.02 (0.01) 0.007 (0.001)
IL-10 11.51 (3.49) 0.56 (0.11) 0.011 (0.001)
Values are means (standard error of the mean). n.d., not determined.
Table 2 Correlation between mRNA expression and severity of knee joint swelling in rats with AIA
Cytokine Day(s) Correlationa Pa ρa n
Synovial membrane
IL-1β 0–6 + 0.05 0.527 25
IL-6 0–6 + 0.03 0.405 28
Inguinal lymph node
IL-6 3 + 0.02 0.886 6
Spleen
IL-6 0–6 + 0.01 0.511 25
IL-5 6 - <0.001 1.000 4
IL-10 6 - 0.04 0.829 6
aSpearman rank correlation. +, positive; -, negative; AIA, antigen-induced arthritis.
Table 3 Changes (fold change relative to day 0) in cytokine protein concentrations during rat AIA
Time point IL-1β IL-6 TNF-α IL-2 IFN-γ IL-4 IL-10
Inguinal lymph node
Day 0 1.00 (0.27) 1.00 (0.26) 1.00 (0.26) 1.00 (0.31) 1.00 (0.28) 1.00 (0.25) 1.00 (0.20)
6 hours 1.67 (0.60) 0.98 (0.38) 1.93 (0.78) 1.06 (0.31) 0.98 (0.25) 1.38 (0.55) 1.60 (0.32)
Day 1 1.74 (0.58) 0.59 (0.42) 2.53 (0.78) 1.31 (0.45) 1.14 (0.34) 1.86 (0.72) 2.18* (0.49)
Day 3 1.24 (0.29) 0.40 (0.21) 1.38 (0.28) 0.66 (0.12) 0.55 (0.11) 0.53 (0.28) 1.41 (0.24)
Day 6 1.86* (0.41 0.10* (0.10) 1.87 (0.38) 0.69 (0.14) 0.81 (0.12) 0.91 (0.27) 0.89 (0.17)
Spleen
Day 0 1.00 (0.12) 1.00 (0.41) 1.00 (0.10) 1.00 (0.23) 1.00 (0.13) 1.00 (0.12) 1.00 (0.05)
6 hours 0.75 (0.09) 0.32 (0.14) 0.61* (0.04) 0.70 (0.07) 0.78 (0.04) 0.77 (0.04) 0.94 (0.06)
Day 1 0.72 (0.11) 0.41 (0.17) 0.95 (0.13) 0.84 (0.16) 0.81 (0.07) 0.80 (0.08) 0.87 (0.08)
Day 3 0.81 (0.06) 0.24 (0.14) 1.07 (0.08) 0.70 (0.13) 0.71 (0.06) 0.85 (0.06) 0.90 (0.06)
Day 6 0.67 (0.06) 0.82*+ (0.07) 1.36 (0.21) 1.10*+ (0.13) 0.91+ (0.06) 0.96 (0.10) 0.99 (0.07)
Values (fold changes) are means (standard error of the mean) (n = 5 to 6 rats for each time point). AIA, antigen-induced arthritis. *P ≤ 0.05 in comparison with day 0; +P ≤ 0.05 in comparison with the preceding date.
==== Refs
Burmester GR Stuhlmüller B Keyszer G Kinne RW Mononuclear phagocytes and rheumatoid synovitis. Mastermind or workhorse in arthritis? Arthritis Rheum 1997 40 5 18 9008595
Kinne RW Palombo-Kinne E Emmrich F T-cells in the pathogenesis of rheumatoid arthritis villains or accomplices? Biochim Biophys Acta 1997 1360 109 141 9128178
Fox DA The role of T cells in the immunopathogenesis of rheumatoid arthritis: new perspectives Arthritis Rheum 1997 40 598 609 9125240
Pohlers D Nissler K Frey O Simon J Petrow PK Kinne RW Bräuer R Anti-CD4 monoclonal antibody treatment in acute and early chronic antigen-induced arthritis: Influence on T helper cell activation Clin Exp Immunol 2004 135 409 415 15008972 10.1111/j.1365-2249.2003.02381.x
Nissler K Pohlers D Hückel M Simon J Bräuer R Kinne RW Anti-CD4 monoclonal antibody treatment in acute and early chronic antigen-induced arthritis: influence on macrophage activation Ann Rheum Dis 2004 63 1470 1477 15479897 10.1136/ard.2003.013060
Miossec P An update on the cytokine network in rheumatoid arthritis Curr Opin Rheumatol 2004 16 218 222 15103248 10.1097/00002281-200405000-00009
Abramson SB Amin A Blocking the effects of IL-1 in rheumatoid arthritis protects bone and cartilage Rheumatology (Oxford) 2002 41 972 980 12209029
Abbas AK Murphy KM Sher A Functional diversity of helper T lymphocytes Nature 1996 383 787 793 8893001 10.1038/383787a0
Mosmann TR Sad S The expanding universe of T-cell subsets: Th1, Th2 and more Immunol Today 1996 17 138 146 8820272 10.1016/0167-5699(96)80606-2
Gor DO Rose NR Greenspan NS TH1-TH2: a procrustean paradigm Nat Immunol 2003 4 503 505 12774069 10.1038/ni0603-503
O'Garra A Steinman L Gijbels K CD4+ T-cell subsets in autoimmunity Curr Opin Immunol 1997 9 872 883 9492992 10.1016/S0952-7915(97)80192-6
Simon AK Seipelt E Sieper J Divergent T-cell cytokine patterns in inflammatory arthritis Proc Natl Acad Sci USA 1994 91 8562 8566 8078923
Schulze-Koops H Kalden JR The balance of Th1/Th2 cytokines in rheumatoid arthritis Best Pract Res Clin Rheumatol 2001 15 677 691 11812015 10.1053/berh.2001.0187
Cohen SB Katsikis PD Chu CQ Thomssen H Webb LM Maini RN Londei M Feldmann M High level of interleukin-10 production by the activated T cell population within the rheumatoid synovial membrane Arthritis Rheum 1995 38 946 952 7612044
Griffiths RJ Characterisation and pharmacological sensitivity of antigen arthritis induced by methylated bovine serum albumin in the rat Agents Actions 1992 35 88 95 1509982
Buchner E Bräuer R Schmidt C Emmrich F Kinne RW Induction of flare-up reactions in rat antigen-induced arthritis J Autoimmun 1995 8 61 74 7734037
Billingham ME Models of arthritis and the search for anti-arthritic drugs Pharmacol Ther 1983 21 389 428 6138788 10.1016/0163-7258(83)90062-1
Kinne RW Bräuer R Stuhlmüller B Palombo-Kinne E Burmester GR Macrophages in rheumatoid arthritis Arthritis Res 2000 2 189 202 11094428 10.1186/ar86
Schmidt-Weber CB Pohlers D Siegling A Schädlich H Buchner E Volk HD Palombo Kinne E Emmrich F Kinne RW Cytokine gene activation in synovial membrane, regional lymph nodes, and spleen during the course of rat adjuvant arthritis Cell Immunol 1999 195 53 65 10433797 10.1006/cimm.1999.1509
Siegling A Lehmann M Platzer C Emmrich F Volk HD A novel multispecific competitor fragment for quantitative PCR analysis of cytokine gene expression in rats JYImmunol Methods 1994 177 23 28 10.1016/0022-1759(94)90139-2
Hersmann GHW Kriegsmann J Simon J Hüttich C Bräuer R Expression of cell adhesion molecules and cytokines in murine antigen-induced arthritis Cell Adhes Commun 1998 6 69 82 9759522
Henderson B Rowe FM Bird CR Gearing AJ Production of interleukin 1 in the joint during the development of antigen-induced arthritis in the rabbit Clin Exp Immunol 1988 74 371 376 3266117
Kleinau S Dencker L Klareskog L Oil-induced arthritis in DA rats: tissue distribution of arthritogenic 14C-labelled hexadecane Int J Immunopharmacol 1995 17 393 401 7591363 10.1016/0192-0561(95)00020-3
Dijkstra CD Dopp EA Vogels IM Van Noorden CJ Macrophages and dendritic cells in antigen-induced arthritis. An immunohistochemical study using cryostat sections of the whole knee joint of rat Scand J Immunol 1987 26 513 523 3479836
Verschure PJ Van Noorden CJ Dijkstra CD Macrophages and dendritic cells during the early stages of antigen-induced arthritis in rats: immunohistochemical analysis of cryostat sections of the whole knee joint Scand J Immunol 1989 29 371 381 2717881
Dijkstra CD Dopp EA Joling P Kraal G The heterogeneity of mononuclear phagocytes in lymphoid organs: distinct macrophage subpopulations in the rat recognized by monoclonal antibodies ED1, ED2 and ED3 Immunology 1985 54 589 599 3882559
Graabaek PM Ultrastructural evidence for two distinct types of synoviocytes in rat synovial membrane J Ultrastruct Res 1982 78 321 339 7086939 10.1016/S0022-5320(82)80006-3
Koch AE Kunkel SL Harlow LA Mazarakis DD Haines GK Burdick MD Pope RM Strieter RM Macrophage inflammatory protein-1 alpha. A novel chemotactic cytokine for macrophages in rheumatoid arthritis J Clin Invest 1994 93 921 928 8132778
Rioja I Bush KA Buckton JB Dickson MC Life PF Joint cytokine quantification in two rodent arthritis models: kinetics of expression, correlation of mRNA and protein levels and response to prednisolone treatment Clin Exp Immunol 2004 137 65 73 15196245 10.1111/j.1365-2249.2004.02499.x
Billingham MEJ Hicks CA Carney SL Monoclonal antibodies and arthritis Agents Actions 1990 29 77 87 2327319
Boissier MC Chiocchia G Bessis N Hajnal J Garotta G Nicoletti F Fournier C Biphasic effect of interferon-gamma in murine collagen-induced arthritis Eur J Immunol 1995 25 1184 1190 7774621
Malek TR Bayer AL Tolerance, not immunity, crucially depends on IL-2 Nat Rev Immunol 2004 4 665 674 15343366 10.1038/nri1435
Walmsley M Katsikis PD Abney E Parry S Williams RO Maini RN Feldmann M Interleukin-10 inhibition of the progression of established collagen-induced arthritis Arthritis Rheum 1996 39 495 503 8607899
Finnegan A Kaplan CD Cao Y Eibel H Glant TT Zhang J Collagen-induced arthritis is exacerbated in IL-10-deficient mice Arthritis Res Ther 2003 5 R18 R24 12716449 10.1186/ar601
Horsfall AC Butler DM Marinova L Warden PJ Williams RO Maini RN Feldmann M Suppression of collagen-induced arthritis by continuous administration of IL-4 JYImmunol 1997 159 5687 5696
Svensson L Nandakumar KS Johansson A Jansson L Holmdahl R IL-4-deficient mice develop less acute but more chronic relapsing collagen-induced arthritis Eur J Immunol 2002 32 2944 2953 12355448 10.1002/1521-4141(2002010)32:10<2944::AID-IMMU2944>3.0.CO;2-4
De Franco M Gille-Perramant MF Mevel JC Couderc J T helper subset involvement in two high antibody responder lines of mice (Biozzi mice): HI (susceptible) and HII (resistant) to collagen-induced arthritis Eur J Immunol 1995 25 132 136 7843223
Bräuer R Kette H Henzgen S Thoss K Influence of cyclosporin A on cytokine levels in synovial fluid and serum of rats with antigen-induced arthritis Agents Actions 1994 41 96 98 8079831
Theisen-Popp P Pape H Müller-Peddinghaus R Interleukin-6 (IL-6) in adjuvant arthritis of rats and its pharmacological modulation Int J Immunopharmacol 1992 14 565 571 1521924 10.1016/0192-0561(92)90117-4
Sack U Kinne RW Marx T Heppt P Bender S Emmrich F Interleukin-6 in synovial fluid is closely associated with chronic synovitis in rheumatoid arthritis Rheumatol Int 1993 13 45 51 8356390 10.1007/BF00307733
Hunneyball IM Spowage M Crossley MJ Rowe IF Baltz ML Acute phase protein changes in antigen-induced mono-articular arthritis in rabbits and mice Clin Exp Immunol 1986 65 311 318 2431818
van de Loo FA Kuiper S van Enckevort FH Arntz OJ van den Berg WB Interleukin-6 reduces cartilage destruction during experimental arthritis. A study in interleukin-6-deficient mice Am J Pathol 1997 151 177 191 9212744
de Hooge AS van de Loo FA Arntz OJ van den Berg WB Involvement of IL-6, apart from its role in immunity, in mediating a chronic response during experimental arthritis Am J Pathol 2000 157 2081 2091 11106580
Müssener A Litton MJ Lindroos E Klareskog L Cytokine production in synovial tissue of mice with collagen- induced arthritis (CIA) Clin Exp Immunol 1997 107 485 493 9067522 10.1046/j.1365-2249.1997.3181214.x
Roubenoff R Freeman LM Smith DE Abad LW Dinarello CA Kehayias JJ Adjuvant arthritis as a model of inflammatory cachexia Arthritis Rheum 1997 40 534 539 9082942
Charlton B Lafferty KJ The Th1/Th2 balance in autoimmunity Curr Opin Immunol 1995 7 793 798 8679122 10.1016/0952-7915(95)80050-6
Katsikis PD Chu CQ Brennan FM Maini RN Feldmann M Immunoregulatory role of interleukin 10 in rheumatoid arthritis J Exp Med 1994 179 1517 1527 8163935 10.1084/jem.179.5.1517
Bessis N Honiger J Damotte D Minty A Fournier C Fradelizi D Boissier M Encapsulation in hollow fibres of xenogeneic cells engineered to secrete IL-4 or IL-13 ameliorates murine collagen-induced arthritis (CIA) Clin Exp Immunol 1999 117 376 382 10444273 10.1046/j.1365-2249.1999.00959.x
| 15899031 | PMC1174936 | CC BY | 2021-01-04 16:02:36 | no | Arthritis Res Ther. 2005 Feb 17; 7(3):R445-R457 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1689 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar16921589902910.1186/ar1692Research ArticleThe utility of pathway selective estrogen receptor ligands that inhibit nuclear factor-κB transcriptional activity in models of rheumatoid arthritis Keith James C [email protected] Leo M [email protected] Yelena [email protected] Max [email protected] Lili [email protected] Lisa [email protected] Christopher C [email protected] Robert J [email protected] Douglas C [email protected] Cardiovascular and Metabolic Disease Research, Wyeth Research, Cambridge, MA, USA2 Department Biological Technologies, Cambridge, MA, USA3 Cardiovascular and Metabolic Disease Research, Collegeville, PA, USA4 Women's Health Research Institute, Collegeville, PA, USA5 Chemical and Screening Sciences, Collegeville, PA, USA2005 21 2 2005 7 3 R427 R438 3 6 2004 29 6 2004 12 1 2005 17 1 2005 Copyright © 2005 Keith et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Rheumatoid arthritis (RA) is a chronic inflammatory disease that produces synovial proliferation and joint erosions. The pathologic lesions of RA are driven through the production of inflammatory mediators in the synovium mediated, in part, by the transcription factor NF-κB. We have identified a non-steroidal estrogen receptor ligand, WAY-169916, that selectively inhibits NF-κB transcriptional activity but is devoid of conventional estrogenic activity. The activity of WAY-169916 was monitored in two models of arthritis, the HLA-B27 transgenic rat and the Lewis rat adjuvant-induced model, after daily oral administration. In both models, a near complete reversal in hindpaw scores was observed as well as marked improvements in the histological scores. In the Lewis rat adjuvant model, WAY-169916 markedly suppresses the adjuvant induction of three serum acute phase proteins: haptoglobin, α1-acid glycoprotein (α1-AGP), and C-reactive protein (CRP). Gene expression experiments also demonstrate a global suppression of adjuvant-induced gene expression in the spleen, liver, and popliteal lymph nodes. Finally, WAY-169916 was effective in suppressing tumor necrosis factor-α-mediated inflammatory gene expression in fibroblast-like synoviocytes isolated from patients with RA. Together, these data suggest the utility of WAY-169916, and other compounds in its class, in treating RA through global suppression of inflammation via selective blockade of NF-κB transcriptional activity.
==== Body
Introduction
Rheumatoid arthritis (RA) is a chronic, debilitating condition affecting 0.5 to 1% of the world's population. The major goals of treatment of RA are to reduce pain and discomfort, to prevent deformities and loss of joint function, and to maintain a productive and active lifestyle. RA is characterized by chronic joint inflammation mediated by inflammatory cell infiltration into synovial tissues as well as joint destruction through the overexpression of matrix metalloproteinase (MMP) in articular synoviocytes and chondrocytes. The pathologic lesions of RA are driven, in part, by the production of inflammatory mediators in synoviocytes and macrophages, probably involving the transcription factor NF-κB. Because NF-κB is localized in the nuclei of synovial cells in patients with RA [1,2] and the inducers and targets of NF-κB almost perfectly match the list of pivotal mediators increased in RA [3], an important role for activated NF-κB in human RA is likely.
NF-κB is a dimeric transcription factor composed of homodimeric and heterodimeric complexes of the Rel family of proteins, p65 (Rel A), p50/105, c-Rel, p52/100, and Rel B. Binding of cytoplasmic inhibitory protein-κB (IκB) to NF-κB masks the NF-κB nuclear localization signal and sequesters NF-κB in a non-activated form in the cytoplasm. Cell activation by a variety of extracellular signals such as oxidative stress, cytokines, and lipopolysaccharide induces a cascade of events that leads to the degradation of IκB; activated NF-κB then translocates to the nucleus, where it binds to DNA elements in the promoters of several proinflammatory gene families [4].
Activation of NF-κB has been observed in synovial cells from patients with RA [5] and results in the induction of proinflammatory genes such as tumor necrosis factor-α (TNF-α), IL-1β, IL-6, MMP-1, and MMP-3 in ex vivo synovial membrane cultures [6]. Moreover, NF-κB activation might also be a pivotal factor protecting cells from apoptosis, thus contributing to synovial hyperplasia (reviewed in [7]). Inactivation of NF-κB in transgenic mice expressing a 'super-repressor' IκBα or in rel- /- and nfkb1- /- knockout mice rendered the animals refractory to development of collagen-induced arthritis [8,9]. In another study performed in the rat adjuvant-induced arthritis model, intra-articular injection of an adenoviral construct encoding a dominant-negative from of IκB kinase-2 significantly ameliorated the severity of the adjuvant arthritis and was correlated with a decrease in NF-κB DNA binding in the nucleus of synovial cells [10]. Because NF-κB is involved in normal immune and homeostatic processes, its prolonged inhibition might be harmful. Therefore, more indirect methods of targeting NF-κB might provide a safer pharmacological profile.
In tissues that express estrogen receptor (ER), 17β-estradiol inhibits NF-κB-driven transcription through multiple mechanisms that might include direct protein–protein interactions [11,12], inhibition of NF-κB binding to DNA [13,14], induction of IκB expression [15], or coactivator sharing [16,17]. Two nuclear estrogen receptors have been identified (ERα and ERβ). Both receptors are widely distributed throughout numerous organs [18] and are present in T cells, monocytes, dendritic cells, synovial macrophages, articular chondrocytes, and proliferating fibroblasts present in the RA joint [19-22]. These two receptors have a nearly identical DNA-binding domain, both activate transcription through binding to identical ER response elements [23,24], and both can antagonize NF-κB transcriptional activity [25,26]. Taken together, these findings identify RA as a disease amenable to treatment with ER-selective NF-κB inhibitors.
The selective inflammatory modulator WAY-169916 is a non-steroidal ER-dependent inhibitor of NF-κB transcriptional activity. Although it inhibits the expression of a range of inflammatory proteins, including cytokines, chemokines, and cell adhesion molecules that are expressed after activation of NF-κB, WAY-169916 lacks estrogenic activity such as the stimulation of uterine proliferation [27]. Here we demonstrate that WAY-169916 is efficacious in two models of arthritis, the HLA-B27 transgenic rat and a Lewis rat model of adjuvant-induced arthritis. The activity of WAY-169916 is related to its ability to suppress inflammatory processes globally, as demonstrated by the decrease in serum acute-phase protein levels of haptoglobin, α1-acid glycoprotein (α1-AGP), and C-reactive protein (CRP) as well as the inhibition of adjuvant-induced gene expression in the spleen, liver, and popliteal lymph nodes in the rat adjuvant arthritis model. Moreover, WAY-169916 was also active in suppressing cytokine and adhesion molecule expression in fibroblast-like synoviocytes (FLS) isolated from patients with RA. Taken together, these data suggest the potential utility of the pathway-selective ER ligand WAY-169916 and other compounds in its class in the treatment of RA.
Materials and Methods
Animals
Male HLA-B27 transgenic rats were obtained from Taconic; the Lewis rats were purchased from Charles River Laboratories. The rats were housed in accordance with standard operating procedures and were provided with food and water ad libitum. All experiments were approved and performed in accordance with the Wyeth Animal Care and Use Committee standards.
HLA-B27 transgenic rat model arthritis
HLA-B27 transgenic rats, 26 to 28 weeks old, experiencing maximal clinical signs of arthritis with a score of 12, using a scale of 0 to 3 for swelling and for erythema of the hindpaws (0, normal paw; 1, mild; 2, moderate; 3, severe) were treated with vehicle (2% Tween 80, 0.5% methylcellulose), prednisolone (0.6 mg/kg), or WAY-169916 (10 mg/kg) given orally once daily for 29 days with four rats per group. At necropsy, the tarsal joints were removed and prepared for histological examination. After decalcification, histological sections were stained with hematoxylin and eosin or Safranin O/Fast Green stain. Synovial tissue from tarsal joints was evaluated on the basis of synovial hyperplasia, fibroplasia, inflammatory cell infiltration, and pannus formation [28].
Articular cartilage was evaluated with Mankin's histological grading system [29]. The scoring system evaluates the structure of the articular cartilage, ranging from 0 (normal), 1 (surface irregularity), 2 (pannus and surface irregularity), 3 (clefts to transititional zone), 4 (clefts to radial zone), 5 (clefts to the calcified zone), to 6 (complete disorganization; cartilage cells, ranging from 0 (normal), 1 (diffuse hypercellularity), 2 (cloning), to 3 (hypocellularity); Safranin-O staining to assess proteoglycan content, ranging from 0 (normal), 1 (slight reduction), 2 (modest reduction), 3 (severe reduction), to 4 (no staining); and tidemark integrity, ranging from 0 (intact) to 1 (crossed by blood vessels). The scores for eachtarsal joint were tabulated and summed, and amean score was derived for each animal, ranging from 0 to 14. Statistical analysis was performed with Abacus Concepts Super ANOVA (Abacus Concepts, Inc., Berkeley, CA). All parameters of interest were subjected to analysis of variance (ANOVA) with Duncan's new multiple-range post hoc testing between groups. Data are expressed throughout as means ± standard deviation (SD), and differences were deemed significant if P < 0.05.
Rat adjuvant-induced arthritis model
Arthritis was induced in the Lewis rats with complete Freund's adjuvant (CFA) by intradermal injection of 0.1 mg of heat-killed and dried Mycobacterium tuberculosis in 0.1 ml of mineral oil, at the base of the tail. Eight days after adjuvant injection, when the rats were experiencing maximal clinical signs of arthritis with a score of 12 using the same hindpaw scoring system described above, treatment began. Male Lewis rats (n = 6) received orally delivered vehicle (2.0% Tween 80, 0.5% methylcellulose, 1 ml/kg) or WAY-169916 (5.0, 0.3, or 0.1 mg/kg) once daily for 10 to 14 days, with six rats in each group. The clinical signs of arthritis were monitored daily. At the end of the experiment, terminal blood samples were obtained and the tarsal joints were prepared for histological examination and graded as described above. Statistical analysis was performed with Abacus Concepts Super ANOVA. All parameters of interest were subjected to ANOVA with Duncan's new multiple-range post hoc testing between groups. The serum samples were used to determine the levels of haptoglobin, α1-AGP, and CRP by radial immunodiffusion test kits in accordance with manufacturer's protocol (Life Diagnostics Inc.). The data were analyzed by one-way ANOVA and are expressed as means ± SD, and differences were deemed significant if P < 0.05.
Gene expression profiling experiments were conducted with RNA isolated from the spleen, liver, and popliteal lymph nodes, using Affymetrix REA230A oligonucleotide arrays (Affymetrix) in accordance with the manufacturer's recommendations. The arrays were washed and stained with Streptavidin R–phycoerythrin (Molecular Probes) with the use of the GeneChip® Fluidics Station 400, and scanned with a Hewlett Packard GeneArray Scanner in accordance with the manufacturer's instructions. Fluorescent data were collected and converted to gene-specific difference averages with MicroArray Suite 4.0 software. A representative set of genes regulated by WAY-169916 was confirmed by real-time RT–PCR analysis. All mRNA levels were normalized for glyceraldehyde-3-phosphate dehydrogenase expression. The data were analyzed by one-way ANOVA and expressed as means ± SD, and differences were deemed significant if P < 0.05.
NF-κB DNA binding experiments
Mouse splenocytes were prepared by creating single-cell suspensions, with the subsequent removal of red blood cells with Tris-NH4Cl solution. After lysis of red blood cells, the cells were cultured in 24-well plates at a concentration of 106 cells/ml in RPMI-10 (RPMI medium containing 10% heat-inactivated fetal bovine serum, 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM glutamine, and 50 μM 2-mercaptoethanol; Invitrogen, Carlsbad, CA). The cells were stimulated with concanavalin A (ConA) and co-treated with either WAY-169916 (1 μM) or pyrrolidine dithiocarbamate (PDTC; 100 μM) for 18 hours. Nuclear extract preparation and NF-κB DNA binding experiments were conducted with kits purchased from Active Motif.
Experiments with FLS
Human FLS isolated from patients with RA were purchased from Cell Applications, Inc. The cells were cultured in synoviocyte growth medium (Cell Applications, Inc.) and seeded at 3 × 105 cells per well in a 12-well dish. After overnight culture, the cells were pretreated for 1 hour with vehicle, WAY-169916 (1 μM), or PDTC (100 μM), followed by stimulation for 1 hour with TNF-α (100 U/ml). Synoviocyte RNA was isolated after the 1 hour of TNF-α treatment, and gene expression analysis was performed using real-time RT–PCR with an ABI PRISM 7900 Sequence Detection System, in accordance with the manufacturer's protocol (Applied Biosystems). The data were analyzed with Sequence Detector v2.1 software (Applied Biosytems) and normalized to glyceraldehyde-3-phosphate dehydrogenase with the Applied Biosystems primer set. Values are reported as means ± SEM for each group from two experiments, with n = 3. The data were analyzed by one-way ANOVA and differences were deemed significant if P < 0.05.
Results
Activity in the hla-b27 transgenic rat
The HLA-B27 transgenic rat expresses two human proteins (HLA-B27 and β2-microglobulin) that, over time, provoke a misdirected immune response. This model represents a chronic intestinal inflammation with associated arthritis induced by the human class I major histocompatibility allele HLA-B27, which is strongly associated with human disease. Treatment of male HLA-B27 transgenic rats with concentrations of WAY-169916 as low as 0.05 mg/kg rapidly converts the chronic diarrhea that these rats experience to a normal stool [27]. If the disease is allowed to progress, they begin to show symptoms of arthritis. In these settings, treatment of WAY-169916 at a single oral dosage of 10 mg/kg per day restored the clinical joint scores to baseline after 10 days, while a sub-optimal dose of prednisolone (0.6 mg/kg) resulted in a 50% improvement in the joint scores (Fig. 1). Histological scoring of synovitis and cartilage damage in the tarsal joints after 29 days of treatment was also conducted. Treatment with WAY-169916 significantly decreased the synovitis parameters of synovial structure, fibroplasia, inflammatory cell infiltrates, and total synovitis score, and also significantly improved all cartilage parameters monitored (Table 1).
Activity of WAY-169916 in the Lewis rat adjuvant-induced arthritis model
WAY-169916 was then given a more thorough evaluation in the male Lewis rat adjuvant-induced arthritis model. The disease in this model is a migratory polyarthritis affecting primarily the tarsal, metatarsal, and interphalangeal joints. The hallmarks of the model include polyarticular inflammation, marked bone resorption, and periosteal bone proliferation. When immunized with CFA, the joints of Lewis rats swell markedly over a period of 8 days. After maximal swelling had occurred, rats received an oral daily dose of WAY-169916, making this a therapeutic dosing regimen. Joint swelling was rapidly and markedly reduced in rats treated with WAY-169916. Full efficacy was seen with oral doses of 0.3 mg/kg or higher (Fig. 2) but efficacy was decreased at a dose of 0.1 mg/kg. However, both doses were effective at reversing tarsal joint destruction as assessed by synovitis and cartilage (Mankin) scores (Table 2). Incremental improvements in the histology scores were observed with higher doses of WAY-169916 (data not shown), suggesting that continued improvements in joint lesions might occur with a longer duration of treatment or with higher dosages.
Because both the HLA-B27 transgenic rat and Lewis rat studies used males, the efficacy of WAY-169916 (5 mg/kg) was compared in intact male and female Lewis rats with the same experimental design as described above. The joint (Fig. 2b) and histology scores (not shown) for the two sexes were equivalent; it therefore does not seem that the utility of WAY-169916 is restricted by gender.
Mechanism of action of WAY-169916
Because WAY-169916 has been shown to antagonize NF-κB transcriptional activity selectively [27], we wished to begin to address how WAY-169916 might be functioning to improve disease symptoms in the rat adjuvant model. Previous studies have shown changes in concentrations of rat serum proteins induced by adjuvant administration (reviewed in [30]). We decided to look at three acute-phase proteins, haptoglobin, α1-AGP, and CRP, that are induced by the adjuvant and have been correlated with RA progression in humans [31]. Serum was analyzed from male Lewis rats treated with 5 mg/kg WAY-169916 for 10 days. As shown in Fig. 3, both haptoglobin and α1-AGP serum levels were induced about 300 to 400% by adjuvant treatment, whereas CRP inductions were more modest (40%); this was consistent with previous reports [30]. WAY-169916 inhibited the adjuvant induction of all three acute-phase proteins but had no effect on their basal levels.
We also performed gene expression profile analysis from the spleen, liver, and popliteal lymph nodes from these rats. In the spleen, 36 genes were identified that were induced twofold by adjuvant treatment (average fold change; Table 3). Of those 36 genes, WAY-169916 decreased the expression of 29 of them by at least 50%. Several genes that have been implicated in the pathogenesis of RA that were regulated by WAY-169916 include LBS binding protein (LBP), CD14, MMP-9, IL1R2, S100A8, and S100A9. As a control, the regulation of LBP, haptoglobin, and S100A9 was confirmed by real-time RT–PCR (Fig. 4a). A similar global inhibition of adjuvant-induced genes by WAY-169916 was also observed in liver and popliteal lymph node gene-profiling studies. In the liver, 47 genes were induced and WAY-169916 inhibited 43 of those by 50%; in the lymph node, 143 genes were induced and 61 of those were repressed by 50% by WAY-169916 (data not shown).
In addition, we attempted to determine whether treatment with WAY-169916 resulted in direct interference of NF-κB DNA binding in primary spleen cell cultures. The cells were stimulated with ConA (5 μg/ml) for 24 hours and co-treated with either WAY-169916 (1 μM) or PDTC (100 μM), a general inhibitor of NF-κB. As shown in Fig. 4b, activation by ConA resulted in an 80% increase in NF-κB DNA binding. Although PDTC treatment could completely block NF-κB activation, WAY-169916 was without effect. Control experiments demonstrated that the binding of NF-κB was specific, because competition experiments with wild-type oligonucleotide interfered with binding activity whereas a mutated oligonucleotide was without effect (data not shown). These results are consistent with our previous observations [16,27] demonstrating that liganded ER inhibits NF-κB at the transcriptional level downstream from NF-κB DNA binding. Overall, these data indicate a marked anti-inflammatory effect for WAY-169916 that seems to cross multiple signaling pathways and tissues consistent with NF-κB's ubiquitous role in inflammation.
WAY-169916 anti-inflammatory activity in synoviocytes isolated from patients with RA
Finally, we wished to test whether WAY-169916 is active in FLS, a human cell type that is thought to have a pathologic function in joint destruction through its production of inflammatory cytokines and MMPs [32]. Activation of NF-κB in FLS is necessary for the production of these inflammatory mediators [5,6]. FLS obtained from male patients with RA were stimulated with TNF-α and treated with vehicle, WAY-169916 (1 μM), or PDTC (100 μM). RNA was analyzed for gene expression changes of intercellular cell-adhesion molecule-1 (ICAM-1), IL-6, and TNF-α by real-time RT–PCR. As shown in Fig. 5, TNF-α-stimulated expression of all three inflammatory genes was significantly blocked by both WAY-169916 and PDTC, which was consistent with previous observations [5]. The cells were confirmed to express ERα mRNA [33] but no ERβ mRNA was detected (data not shown). In total, these data suggest the potential utility of non-steroidal selective NF-κB modulators such as WAY-169916 in treating patients with RA.
Discussion
RA might occur as a result of an autoimmune response, and recent studies suggest that hypersensitivity to microbial antigens contributes to the development of the arthritis. Microbial or self-antigen presentation to T lymphocytes results in chronic activation of the immune system. Multiple proinflammatory mediators, including IL-1, TNF-α, interferon-γ, and MMPs mediate the inflammation of the joints. Biochemical and histological changes in synovial tissue, cartilage, and bone have been documented in various animal models of arthritis [34,35]. In many respects the synovial and cartilage lesions that develop in these models closely resemble those seen in rheumatoid arthritis. We have investigated the role of WAY-169916 in two such models.
The HLA-B27 transgenic rats spontaneously develop arthritis similar to the human spondyloarthropathies associated with the HLA-B27 and β2-microglobulin genes through a T cell-mediated process [34]. In this model, WAY-169916 restored the clinical joint scores to baseline after 10 days. Histological scoring of synovitis and cartilage damage in the tarsal joints after 29 days of treatment was also significantly improved with WAY-169916 treatment.
In the rat adjuvant-induced arthritis model [36], 3 to 6 days after the injection of adjuvant, induction of an αβ T cell response occurs and leads to clinical lesion development in the tarsal joints within 5 to 8 days. Because activated NF-κB was detected in the synovial lining layer and around blood vessels in the inflamed synovium as early as day 3 after adjuvant injection in the Lewis rats and is thought to be correlated with disease development [37], this model was used to test the therapeutic treatment with WAY-169916. We demonstrated that WAY-169916 was effective in improving both joint and histology scores at doses as low as 0.3 mg/kg given orally once daily. Improvement in the synovitis and Mankin scores did occur with higher doses of WAY-169916 even though the joint score reduction was already maximal at 0.3 mg/kg. When the rats were dosed at 5 mg/kg the total synovitis score decreased to 4.14 (data not shown). The beneficial effects of WAY-169916 on joint histology might therefore continue with increasing dose or longer exposure.
A benefit on arthritis progression with non-selective estrogens such a 17β-estradiol has also been demonstrated in both the rat adjuvant-induced arthritis model [38] and the collagen-induced mouse model [39,40]. Indeed, 17β-estradiol has been shown to affect several processes involved in the pathogenesis of RA, including immunoregulation, regulation of adhesion molecules, and modulation of cytokine signaling. However, the role of estrogen has not been well defined in patients with RA. There is evidence that gender might affect the occurrence and progression of RA. Women have a higher risk of developing RA than men. During pregnancy, the disease activity is ameliorated in 75% of women, whereas after delivery, flares occur in up to 90% of patients [41]. The highest incidence of developing RA coincides with menopause, indicating that a decrease in estrogen production might increase the risk of joint inflammation. In a recent randomized clinical trial, post-menopausal women taking hormone therapy had suppressed signs of inflammation and significantly improved disease severity scores (DS28) after 12 months of treatment, which was consistent with previous trials [42].
With the identification of selective NF-κB transcriptional inhibitors such as WAY-169916, the expectation is to accentuate the anti-inflammatory, anti-rheumatic activity observed with the non-selective estrogens. Whereas non-selective estrogens have been documented to contain an anti-inflammatory activity through the suppression of NF-κB transcriptional activity [11,12], hormone therapy can simultaneously elicit both proinflammatory and anti-inflammatory activities as exemplified by the decrease in haptoglobin and α1-AGP levels in women taking hormone therapy [43] while also inducing MMP9 and CRP levels [44,45]. In the rat adjuvant model, WAY-169916 inhibited the adjuvant induction of CRP levels and those of haptoglobin and α1-AGP. Moreover, WAY-169916 had no effect on the basal levels of CRP whereas treatment with estradiol has been shown to increase rat CRP serum levels [46], suggesting a potential differential effect of WAY-169916 in comparison with estradiol. This differential activity has been demonstrated on several classic estrogenic effects. WAY-169916 neither stimulates creatine kinase gene expression driven via an estrogen receptor response element in vitro nor promotes uterine proliferation in vivo [27] while retaining the anti-inflammatory activity as demonstrated here.
The anti-inflammatory activity of WAY-169916 was further demonstrated in a series of gene-profiling experiments. In the spleens from the adjuvant-treated rats, 36 genes were identified that were induced greater than twofold by the adjuvant treatment. WAY-169916, when dosed at 5 mg/kg, repressed 29 of those genes by at least 50%, and 17 of them by more than 75%. An attractive hypothesis for WAY-169916-mediated activity in the spleen involves the downregulation of LBP and CD14 expression on monocytes and macrophages, resulting in a diminished immune response and ultimately resulting in the observed decreases in MMP9, IL1R2, chemokine-like factor 1, S100A8, and S100A9 through the repression of NF-κB activity [47,48]. In spleen cell cultures, WAY-169916 treatment did not interfere with ConA-stimulated NF-κB DNA binding; however, the downregulation of S100A9 mRNA was confirmed (data not shown), which was consistent with our hypothesis that ER regulates NF-κB at the transcriptional level [16,27]. A similar suppression of adjuvant-induced inflammatory gene expression was also observed in liver and lymph node studies. These data demonstrate that WAY-169916 can have an effect on a global level, both in terms of the tissues targeted and the different inflammatory signaling pathways, to suppress adjuvant-induced gene expression.
Infiltration of inflammatory cells into the synovial tissue and lining layer results in the formation of pannus, a highly vascularized tissue comprising FLS, macrophages, and lymphocytes. FLS are known for their role in joint destruction through the production of cytokines and MMP, which contribute to cartilage degradation (reviewed in [49]). Expression of ER has been detected in synovial tissues from patients with RA [50] and localized to synoviocytes in the synovial lining [22], providing another potential cell type by which WAY-169916 functions. Synoviocytes isolated from a male RA patient were confirmed to express ERα mRNA [33], but no ERβ mRNA was detected (data not shown). The ERα was functional in these cells, because WAY-169916 could effectively block the TNF-α-mediated inflammatory gene expression of IL-6, TNF-α, and ICAM-1. The potential involvement of NF-κB in mediating TNF-α gene induction was demonstrated with the use of a general NF-κB inhibitor, as shown previously [5]. Given the importance of TNF-α signaling in RA disease progression, the ability of WAY-169916 to interfere with this signaling pathway in human synoviocytes suggests a potential clinical benefit for WAY-169916 in patients with RA.
Conclusions
We detailed the activity of the first pathway-selective ER ligand, WAY-169916, in two models of RA. This compound selectively inhibits NF-κB activity via the ER and imparts significant efficacy in the HLA-B27 and Lewis rat adjuvant-induced models of arthritis. More importantly, no evidence for classic estrogenic activity has been observed with this compound [27]. These data provide evidence that the non-steroidal, pathway-selective ER ligand, WAY-169916, and other compounds in its class might be therapeutically useful in the treatment of RA.
Abbreviations
α1-AGP = α1-acid glycoprotein; ANOVA = analysis of variance; CFA = complete Freund's adjuvant; ConA = concanavalin A; CRP = C-reactive protein; ER = estrogen receptor; FLS = fibroblast-like synoviocytes; ICAM-1 = intercellular cell-adhesion molecule-1; IκB = inhibitory protein-κB; IL = interleukin; LBP = LBS binding protein; MMP = matrix metalloproteinase; NF-κB = nuclear factor-κB; PDTC = pyrrolidine dithiocarbamate; RA = rheumatoid arthritis; RT–PCR = reverse transcriptase polymerase chain reaction; TNF-α = tumor necrosis factor-α.
Competing interests
The authors are employees of Wyeth.
Authors' contributions
JCK, LMA and YL performed the in vivo experiments. MF and LW performed the gene-profiling experiments. LBM performed the cell-based assays and serum analysis. CCC, RJS and DCH were involved in the conception and identification of the molecule, and DCH wrote the manuscript. All authors contributed intellectually to the work and read and approved the final manuscript.
Acknowledgements
We thank all the members of the Discovery and Development Teams that contributed to this program.
Figures and Tables
Figure 1 WAY-169916 improves joint scores in HLA-B27 transgenic rat model of arthritis. HLA-B27 transgenic rats, 26 to 28 weeks old, presenting signs of arthritis were treated orally daily with vehicle, prednisolone (0.6 mg/kg), or WAY-169916 (10 mg/kg) for 29 days. Joint scores were assessed by evaluating hindpaws for erythema and swelling (0 to 3 each; maximal score of 12).
Figure 2 WAY-169916 improves joint scores in a dose-dependent fashion in rat adjuvant-induced arthritis model. (a) Male Lewis rats were injected with complete Freund's adjuvant on day 1 and maximal inflammation was allowed to develop. Beginning on day 8 and continuing until day 22, rats were treated daily with oral vehicle or WAY-169916 (0.3 and 0.1 mg/kg). Joint scores were assessed by evaluating hindpaws for erythema and swelling (0 to 3 each; maximal score of 12). (b) WAY-169916 improves joint scores in rat adjuvant-induced arthritis model in both males and females. Experiments were performed as in (a) except that WAY-169916 was dosed daily at 5 mg/kg orally in both intact male and intact female rats.
Figure 3 WAY-169916 inhibits the adjuvant-induced expression of serum acute-phase protein. Serum from control male rats or adjuvant-induced rats treated with either vehicle or WAY-169916 (5 mg/kg) was analyzed for the expression of (a) haptoglobin, (b) α1-acid glycoprotein (α1-AGP) or (c) C-reactive protein (CRP) by radial immunodiffusion assay. Results are expressed as means ± SEM from six rats per group. *P < 0.05 compared with vehicle control. CFA, complete Freund's adjuvant.
Figure 4 The effect of WAY-169916 in spleen cells. (a) The regulation of LBS binding protein (LBP), haptoglobin and S100A9 gene expression by WAY-169916 from spleens from the rat adjuvant model were confirmed by real-time RT–PCR (grey bars) compared with the regulation observed in the gene-profiling experiments (black bars). Results are expressed as means ± SEM from six rats per group. *P < 0.05 compared with vehicle control. (b) Treatment with WAY-169916 does not interfere with the binding of NF-κB to DNA. Nuclear extracts from primary mouse spleen cell cultures were co-treated for 18 hours with vehicle or concanavalin A (ConA; 5 μg/ml) and either WAY-169916 (1 μM) or pyrrolidine dithiocarbamate (PDTC) (100 μM) as indicated. CFA, complete Freund's adjuvant.
Figure 5 Inhibitory effect of treatment with WAY-169916 on inflammatory gene expression in FLS induced by TNF-α. Fibroblast-like synoviocytes (FLS) were pretreated for 1 hour with WAY-169916 (1 μM) or pyrrolidine dithiocarbamate (PDTC) (100 μM) before treatment with tumor necrosis factor-α (TNF-α) for 1 hour. The mRNA levels for TNF-α, IL-6 and intercellular cell-adhesion molecule-1 (ICAM-1) were determined by real-time RT–PCR and normalized to glyceraldehyde-3-phosphate dehydrogenase. Results are reported as means ± SEM for each group, with the mean level of the stimulated cells treated with vehicle defined as 1. *P < 0.05 compared with vehicle control.
Table 1 Histological scoring of synovitis and cartilage damage in the tarsal joints from HLA-B27 transgenic rats
Group Synovial structure (0–3) Fibroplasia (0–3) Inflammatory cells (0–3) Pannus (0–2) Total synovitis score (0–11)
T/MC vehicle 3.00 ± 0.00 2.80 ± 0.45 3.00 ± 0.00 1.60 ± 0.89 10.40 ± 1.34
WAY-169916, 10 mg/kg 1.80 ± 0.45* 1.20 ± 0.84* 1.40 ± 0.55*† 0.40 ± 0.89 5.00 ± 2.24*†
Prednisolone, 0.6 mg/kg 2.40 ± 0.55 2.00 ± 0.71* 2.00 ± 0.00* 1.60 ± 0.89 8.00 ± 1.87
Group Cartilage structure (0–6) Cartilage cells Safranin-O/Fast Green staining (0–4) Tidemark integrity (0–2) Total Mankin score
T/MC vehicle 3.80 ± 0.84 2.80 ± 0.45 3.20 ± 0.45 0 9.80 ± 1.48
WAY-169916, 10 mg/kg 2.20 ± 0.45*† 2.00 ± 0.45*† 1.60 ± 0.55*† 0 5.80 ± 0.84*†
Prednisolone, 0.6 mg/kg 3.40 ± 0.55 2.20 ± 0.45* 3.00 ± 0.00 0 8.60 ± 0.89
Results are means ± SD.
*Significantly less than vehicle (P < 0.005). †Significantly less than prednisolone (P < 0.005).
Table 2 Histological scoring of synovitis and cartilage changes in the tarsal joints from rats with adjuvant-induced arthritis
Group Synovial structure (0–3) Fibroplasia (0–3) Inflammatory cells (0–3) Pannus (0–2) Total synovitis score (0–11)
T/MC vehicle 2.92 ± 0.21 2.67 ± 0.41 2.92 ± 0.21 2.00 ± 0.00 10.5 ± 0.63
WAY-169916, 0.3 mg/kg 2.33 ± 0.41* 2.33 ± 0.52 1.58 ± 0.38* 1.17 ± 0.75 7.42 ± 1.88*
WAY-169916, 0.1 mg/kg 2.17 ± 0.68* 1.92 ± 0.49* 1.50 ± 0.45* 0.83 ± 0.98* 6.42 ± 2.90*
Group Cartilage structure (0–6) Cartilage cells (0–3) Safranin-O/Fast Green staining (0–4) Tidemark integrity (0–1) Total Mankin score (0–14)
T/MC vehicle 3.52 ± 0.42 2.33 ± 0.41 3.00 ± 0.00 0 8.58 ± 0.74
WAY-169916, 0.3 mg/kg 1.75 ± 0.69* 1.58 ± 0.38* 1.83 ± 0.41* 0 5.17 ± 1.77*
WAY-169916, 0.1 mg/kg 2.25 ± 0.42* 1.42 ± 0.49* 1.67 ± 0.41* 0 5.33 ± 1.21*
Results are means ± SD.
*Significantly less than vehicle (P < 0.005).
Table 3 WAY-169916 gene-profiling experiment with spleen from rat adjuvant arthritis model
Av. control Av. CFA Av. CFA + WAY-169916 AFC CFA AFC WAY-169916 Inhibition by WAY-169916 (%)
Transcription factors
CCAAT/enhancer binding protein (C/EBP), beta 16.8 39.8 19.6 2.37 0.49 87.6
NF-E2-related factor 2 14.2 30.3 19.6 2.13 0.65 66.2
Immune mediators
Mast cell protease 2 5.3 11.8 5.9 2.20 0.50 91.4
Arachidonate 5-lipoxygenase-activating protein 18.9 46.5 23.9 2.46 0.51 82.0
Chemokine-like factor 1 10.5 28.9 18.9 2.75 0.65 54.3
Phospholipase A2, group IIA 33.8 71.7 60.3 2.12 0.84 30.0
Proteoglycan 2, bone marrow 27.6 147.4 59.1 5.34 0.40 73.7
Immune related
CD14 antigen 15.9 32.3 19.1 2.04 0.59 80.2
Peptidoglycan recognition protein 10.2 39.6 14.5 3.88 0.37 85.2
Lipopolysaccharide-binding protein 3.3 10.2 4.4 3.10 0.43 83.6
IL-1 receptor, type II 2.2 8.1 3.5 3.71 0.44 76.8
Defensin RatNP-3 precursor 31.7 105.3 63.5 3.32 0.60 56.8
Suppressor of cytokine signalling 3 15.2 35.3 15.2 2.33 0.43 99.7
Complement component 3 13.3 28.4 24.2 2.14 0.85 27.6
Defensin NP-2 precursor 58.9 166.6 102.8 2.83 0.62 59.3
Defensin NP-4 precursor 54.1 164.9 104.7 3.05 0.64 54.3
Paired immunoglobulin-like receptor-B 11.0 22.7 18.3 2.07 0.81 37.6
25 oligoadenylate synthetase 24.6 51.4 24.8 2.09 0.48 99.2
Tumor necrosis factor receptor II 7.3 15.8 8.9 2.18 0.56 81.0
S100 calcium-binding protein A8 (calgranulin A) 52.2 184.5 91.5 3.54 0.50 70.3
S100 calcium-binding protein A9 (calgranulin B) 144.9 382.7 242.1 2.64 0.63 59.1
Ficolin B 42.3 121.1 74.7 2.86 0.62 58.9
Myelin and lymphocyte protein 11.8 27.6 19.9 2.34 0.72 49.1
Protease
Matrix metalloproteinase 9 2.0 8.4 3.5 4.24 0.42 75.8
Chymase 1 2.9 9.2 5.1 3.16 0.56 64.8
Transport
Monocarboxylate transporter 10.3 24.4 10.0 2.38 0.41 102.3
Lipocalin 2 16.3 55.3 33.0 3.39 0.60 57.1
Acute phase
Haptoglobin 10.5 37.5 12.9 3.58 0.34 90.9
Metabolism
Uridine phosphorylase I 8.9 32.8 9.1 3.68 0.28 99.3
Guanine deaminase 17.8 39.5 20.7 2.21 0.52 86.9
Microsomal glutathione S-transferase 1 40.0 87.6 51.2 2.19 0.58 76.6
GTP cyclohydrolase 1 11.6 25.5 23.6 2.20 0.93 13.8
Hepatic steroid hydroxylase II A2 9.0 18.7 21.5 2.07 1.15 -29.0
Adhesion
C-CAM4 protein 11.5 24.5 16.0 2.12 0.65 65.6
Fibronectin 1 57.2 116.6 91.3 2.04 0.78 42.6
Unknown function
Expressed sequence tag 20.0 41.5 25.1 2.08 0.61 76.1
AFC, average fold change; Av., average; CFA, complete Freund's adjuvant.
==== Refs
Handel ML McMorrow LB Gravallese EM Nuclear factor-kappa B in rheumatoid synovium. Localization of p50 and p65 Arthritis Rheum 1995 38 1762 1770 8849348
Asahara H Asanuma M Ogawa N Nishibayashi S Inoue H High DNA-binding activity of transcription factor NF-kappa B in synovial membranes of patients with rheumatoid arthritis Biochem Mol Biol Int 1995 37 827 832 8624487
Makarov SS NF-kappa B in rheumatoid arthritis: a pivotal regulator of inflammation, hyperplasia, and tissue destruction Arthritis Res 2001 3 200 206 11438035 10.1186/ar300
Pahl HL Activators and target genes of Rel/NF-kappaB transcription factors Oncogene 1999 18 6853 6866 10602461 10.1038/sj.onc.1203239
Fujisawa K Aono H Hasunuma T Yamamoto K Mita S Nishioka K Activation of transcription factor NF-kappa B in human synovial cells in response to tumor necrosis factor alpha Arthritis Rheum 1996 39 197 203 8849369
Bondeson J Foxwell B Brennan F Feldmann M Defining therapeutic targets by using adenovirus: blocking NF-kappaB inhibits both inflammatory and destructive mechanisms in rheumatoid synovium but spares anti-inflammatory mediators Proc Natl Acad Sci USA 1999 96 5668 5673 10318942 10.1073/pnas.96.10.5668
Pope RM Apoptosis as a therapeutic tool in rheumatoid arthritis Nature Rev Immunol 2002 2 527 535 12094227 10.1038/nri846
Seetharaman R Mora AL Nabozny G Boothby M Chen J Essential role of T cell NF-kappa B activation in collagen-induced arthritis J Immunol 1999 163 1577 1583 10415062
Campbell IK Gerondakis S O'Donnell K Wicks IP Distinct roles for the NF-kappaB1 (p50) and c-Rel transcription factors in inflammatory arthritis J Clin Invest 2000 105 1799 1806 10862795
Tak PP Gerlag DM Aupperle KR van de Geest DA Overbeek M Bennett BL Boyle DL Manning AM Firestein GS Inhibitor of nuclear factor kappaB kinase beta is a key regulator of synovial inflammation Arthritis Rheum 2001 44 1897 1907 11508443 10.1002/1529-0131(200108)44:8<1897::AID-ART328>3.0.CO;2-4
Stein B Yang MX Repression of the interleukin-6 promoter by estrogen receptor is mediated by NF-kappa B and C/EBP beta Mol Cell Biol 1995 15 4971 4979 7651415
Ray A Prefontaine KE Ray P Down-modulation of interleukin-6 gene expression by 17 beta-estradiol in the absence of high affinity DNA binding by the estrogen receptor J Biol Chem 1994 269 12940 12946 8175711
Ray P Ghosh SK Zhang DH Ray A Repression of interleukin-6 gene expression by 17 beta-estradiol: inhibition of the DNA-binding activity of the transcription factors NF-IL6 and NF-kappa B by the estrogen receptor FEBS Letters 1997 409 79 85 9199508 10.1016/S0014-5793(97)00487-0
Deshpande R Khalili H Pergolizzi RG Michael SD Chang MD Estradiol down-regulates LPS-induced cytokine production and NFκB activation in murine macrophages Am J Reprod Immunol 1997 38 46 54 9266010
Sun WH Keller ET Stebler BS Ershler WB Estrogen inhibits phorbol ester-induced I kappa B alpha transcription and protein degradation Biochem Biophys Res Commun 1998 244 691 695 9535726 10.1006/bbrc.1998.8324
Harnish DC Scicchitano MS Adelman SJ Lyttle CR Karathanasis SK The role of CBP in estrogen receptor cross-talk with nuclear factor-kappaB in HepG2 cells Endocrinology 2000 141 3403 3411 10965913 10.1210/en.141.9.3403
Speir E Yu ZX Takeda K Ferrans VJ Cannon RO Competition for p300 regulates transcription by estrogen receptors and nuclear factor-kappa B in human coronary smooth muscle cells Circ Res 2000 87 1006 1011 11090545
Kuiper GGJM Carlsson B Grandian K Enmark E Haggblad J Nilsson S Gustafsson JA Comparison of the ligand binding specificity and transcript tissue distribution of estrogen receptor alpha and beta Endocrinology 1997 138 863 870 9048584 10.1210/en.138.3.863
Tornwall J Carey AB Fox RI Fox HS Estrogen in autoimmunity: expression of estrogen receptors in thymic and autoimmune T cells J Gend Specif Med 1999 2 33 40 11252833
Ushiyama T Ueyama H Inoue K Ohkubo I Hukuda S Expression of genes for estrogen receptors alpha and beta in human articular chondrocytes Osteoartheritis Cartilage 1999 7 560 566 10.1053/joca.1999.0260
Cutolo M Accardo S Villaggio B Barone A Sulli A Coviello DA Carabbio C Felli L Miceli D Farruggio R Androgen and estrogen receptors are present in primary cultures of human synovial macrophages J Clin Endocrinol Metab 1996 81 820 827 8636310 10.1210/jc.81.2.820
Liu SH al-Shaikh R Panossian V Yang RS Nelson SD Soleiman N Finerman GA Lane JM Primary immunolocalization of estrogen and progesterone target cells in the human anterior cruciate ligament J Orthop Res 1996 14 526 533 8764860 10.1002/jor.1100140405
Cowley SM Hoare S Mosselman S Parker MG Estrogen receptors alpha and beta form heterodimers on DNA J Biol Chem 1997 272 19858 19862 9242648 10.1074/jbc.272.32.19858
Pace P Taylor J Suntharalingam S Coombes RC Ali S Human estrogen receptor beta binds DNA in a manner similar to and dimerizes with estrogen receptor alpha J Biol Chem 1997 272 25832 25838 9325313 10.1074/jbc.272.41.25832
Tyree CM Zou A Allegretto EA 17beta-Estradiol inhibits cytokine induction of the human E-selectin promoter J Steroid Biochem Mol Biol 2002 80 291 297 11948013 10.1016/S0960-0760(02)00022-5
Valentine JE Kalkhoven E White R Hoare S Parker MG Mutations in the estrogen receptor ligand binding domain discriminate between hormone-dependent transactivation and transrepression J Biol Chem 2000 275 25322 25329 10840033 10.1074/jbc.M002497200
Chadwick CC Chippari S Matelan E Borges-Marcucci L Eckert AM Keith JC JrAlbert LM Leathurby Y Harris HA Bhat RA Identification of pathway selective estrogen receptor ligands that inhibit NF-κB transcriptional activity Proc Natl Acad Sci USA
Poole AR Coombs RR Rheumatoid-like joint lesions in rabbits injected intravenously with bovine serum Int Arch Allergy Appl Immunol 1977 54 97 113 863527
Mankin HJ Dorfman H Lippiello L Zarins A Biochemical and metabolic abnormalities in articular cartilage from osteo-arthritic human hips. II. Correlation of morphology with biochemical and metabolic data J Bone Joint Surg 1971 53 523 537 5580011
Gianazza E Eberini I Villa P Fratelli M Pinna C Wait R Gemeiner M Miller I Monitoring the effects of drug treatment in rat models of disease by serum protein analysis J Chromatogr B Analyt Technol Biomed Life Sci 2002 771 107 130 12015995 10.1016/S0378-4347(01)00562-X
Otterness IG The value of C-reactive protein measurement in rheumatoid arthritis Semin Arthritis Rheum 1994 24 91 104 7839158
Muller-Ladner U Gay RE Gay S Activation of synoviocytes Curr Opin Rheumatol 2000 12 186 194 10803747 10.1097/00002281-200005000-00005
Khalkhali-Ellis Z Seftor EA Nieva DR Handa RJ Price RH JrKirschmann DA Baragi VM Sharma RV Bhalla RC Moore TL Estrogen and progesterone regulation of human fibroblast-like synoviocyte function in vitro: implications in rheumatoid arthritis J Rheum 2000 27 1622 1631 10914842
Taurog JD Arthritis in HLA-B27 transgenic animals Am J Med Sci 1998 316 250 256 9766486 10.1097/00000441-199810000-00005
Waksman BH Immune regulation in adjuvant disease and other arthritis models: relevance to pathogenesis of chronic arthritis Scan J Immunol 2002 56 12 34 10.1046/j.1365-3083.2002.01106.x
Van Eden W Waksman BH Immune regulation in adjuvant-induced arthritis: possible implications for innovative therapeutic strategies in arthritis Arthritis Rheum 2003 48 1788 1796 12847671 10.1002/art.11077
Tsao PW Suzuki T Totsuka R Murata T Takagi T Ohmachi Y Fujimura H Takata I The effect of dexamethasone on the expression of activated NF-kappa B in adjuvant arthritis Clin Immunol Immunopathol 1997 83 173 178 9143378 10.1006/clin.1997.4333
Badger AM Blake SM Dodds RA Griswold DE Swift BA Rieman DJ Stroup GB Hoffman SJ Gowen M Idoxifene, a novel selective estrogen receptor modulator, is effective in a rat model of adjuvant-induced arthritis J Pharmacol Exp Ther 1999 291 1380 1386 10565864
Jansson L Olsson T Holmdahl R Estrogen induces a potent suppression of experimental autoimmune encephalomyelitis and collagen-induced arthritis in mice J Neuroimmunol 1994 53 203 207 8071434 10.1016/0165-5728(94)90030-2
Holmdahl R Jansson L Estrogen-induced suppression of collagen arthritis. III. Adult thymectomy does not affect the course of arthritis or the estrogen-mediated suppression of T-cell immunity Brain Behav Immun 1988 2 123 132 3233356 10.1016/0889-1591(88)90013-X
Kanik KS Wilder RL Hormonal alterations in rheumatoid arthritis, including the effects of pregnancy Rheum Dis Clin North Am 2000 26 805 823 11084945
D'Elia HF Larsen A Mattsson LA Waltbrand E Kvist G Mellstrom D Saxne T Ohlsson C Nordborg E Carlsten H Influence of hormone replacement therapy on disease progression and bone mineral density in rheumatoid arthritis J Rheum 2003 30 1456 1463 12858441
Tuck CH Holleran S Berglund L Hormonal regulation of lipoprotein(a) levels: effects of estrogen replacement therapy on lipoprotein(a) and acute phase reactants in postmenopausal women Arterioscler Thromb Vasc Biol 1997 17 1822 1829 9327783
Zanger D Yang BK Ardans J Waclawiw MA Csako G Wahl LM Cannon RO 3rd Divergent effects of hormone therapy on serum markers of inflammation in postmenopausal women with coronary artery disease on appropriate medical management J Am Coll Cardiol 2000 36 1797 1802 11092646 10.1016/S0735-1097(00)00952-9
Pradhan AD Manson JE Rossouw JE Siscovick DS Mouton CP Rifai N Wallace RB Jackson RD Pettinger MB Ridker PM Inflammatory biomarkers, hormone replacement therapy, and incident coronary heart disease: prospective analysis from the Women's Health Initiative observational study JAMA 2002 288 980 987 12190368 10.1001/jama.288.8.980
El-Swefy SE Asker ME Ali SI Mohammed HE A novel concept to preserve the beneficial effects of hormone replacement therapy in bilaterally female ovariectomized rats: role of lovastatin therapy Pharmacol Res 2002 45 167 173 11884211 10.1006/phrs.2001.0876
Kido J Kido R Suryono Kataoka M Fagerhol MK Nagata T Calprotectin release from human neutrophils is induced by Porphyromonas gingivalis lipopolysaccharide via the CD-14–Toll-like receptor–nuclear factor kappaB pathway J Periodontal Res 2003 38 557 563 14632917 10.1034/j.1600-0765.2003.00691.x
Han W Lou Y Tang J Zhang Y Chen Y Li Y Gu W Huang J Gui L Tang Y Molecular cloning and characterization of chemokine-like factor 1 (CKLF1), a novel human cytokine with unique structure and potential chemotactic activity Biochem J 2001 357 127 135 11415443 10.1042/0264-6021:3570127
McInnes IB Leung BP Liew FY Cell–cell interactions in synovitis. Interactions between T lymphocytes and synovial cells Arthritis Res 2000 2 374 378 11094451 10.1186/ar115
Ushiyama T Inoue K Nishioka J Expression of estrogen receptor related protein (p29) and estradiol binding in human arthritic synovium J Rheum 1995 22 421 426 7540206
| 15899029 | PMC1174937 | CC BY | 2021-01-04 16:02:37 | no | Arthritis Res Ther. 2005 Feb 21; 7(3):R427-R438 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1692 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar16931589903010.1186/ar1693Research ArticleClinical response to discontinuation of anti-TNF therapy in patients with ankylosing spondylitis after 3 years of continuous treatment with infliximab Baraliakos Xenofon [email protected] Joachim 2Brandt Jan 1Rudwaleit Martin 3Sieper Joachim 3Braun Juergen [email protected] Rheumazentrum Ruhrgebiet, Herne, Germany2 German Rheumatism Research Center, Berlin, Germany3 Charité, Medical University of Berlin, Campus Benjamin Franklin, Department of Rheumatology, Germany2005 21 2 2005 7 3 R439 R444 30 11 2004 22 12 2004 7 1 2005 17 1 2005 Copyright © 2005 Baraliakos et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is cited.
We analyzed the clinical response and the time to relapse after discontinuation of continuous long-term infliximab therapy in patients with ankylosing spondylitis (AS). After 3 years of infliximab therapy, all AS patients (n = 42) discontinued treatment (time point (TP)1) and were visited regularly for 1 year in order to assess the time to relapse (TP2). Relapse was defined as an increase to a value ≥ 4 on the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) and a physician's global assessment ≥ 4 according to the recommendations of the Assessments in Ankylosing Spondylitis (ASAS) working group. After 52 weeks, 41 of the 42 patients (97.6%) had to be reinfused because of relapse. The mean change in the BASDAI between TP1 and TP2 was 3.6 ± 1.7 and that in the physician's global assessment was 4.4 ± 1.8 (both P < 0.001). The mean time to relapse was 17.5 weeks (± 7.9 weeks, range 7 to 45). Ten patients (24%) showed a relapse within 12 weeks and 38 patients (90.5%), within 36 weeks. After 52 weeks, only one patient had remained in ongoing remission without further treatment with anti-tumor-necrosis factor. Patients who were in partial remission according to the ASAS criteria and those with normal C-reactive protein levels at the time point of withdrawal had longer times to relapse after discontinuation of the treatment. Retreatment with infliximab was safe and resulted in clinical improvement in all patients to a state similar to that before the treatment was stopped. Discontinuation of long-term therapy with infliximab eventually led to relapse of disease activity in all patients but one.
==== Body
Introduction
Ankylosing spondylitis (AS) is a chronic, immune-mediated inflammatory disease that is associated with inflammation in the sacroiliac joints, the axial skeleton, entheses, peripheral joints, the uvea, and other structures [1-3]. In randomized clinical trials, agents targeting the proinflammatory cytokine tumor necrosis factor (TNF)-α, such as the monoclonal antibody infliximab, have produced significant improvement of signs and symptoms in AS patients [4]. Persistence of clinical response was reported in long-term follow-up studies over 2 [5] and 3 years [6]. These results have been substantiated in studies using magnetic resonance imaging of the spine [7].
We reasoned that it was unclear whether after 3 years of successful therapy with infliximab our patients still needed treatment. Similarly, it was unknown whether discontinuation of the infliximab would be tolerated and whether a restart would be efficacious and safe. Furthermore, nothing was known about the clinical parameters predictive of flare after discontinuation of infliximab therapy. Therefore, we decided to study these questions in our cohort, who had been treated with infliximab for the preceding 3 years [6].
Materials and methods
Patients and study protocol
The AS patients included in this study had all been receiving infliximab for the preceding 3 years, having participated in the first published randomized clinical trial on this therapy in active AS [4,5,8,9]. After the initial, placebo-controlled phase of that trial, the patients entered open extension phases, in which they were treated continuously with 5 mg/kg infliximab every 6 weeks. At the end of the third year of the study (defined as time point (TP)1), all the patients (n = 43) had the opportunity to continue for another extension phase. Only one patient discontinued, because of a side effect. All the others (n = 42) were included in the present extension. In accordance with the study protocol, they gave their informed consent and agreed to discontinuation of the infliximab treatment.
The study was approved by the local ethics committee of each site that participated in this multicenter trial.
Thereafter they were visited regularly at 6-week intervals for assessment of their clinical disease state and the time to relapse (TtR). Relapse was defined as a Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) value ≥ 4 [10]and a physician's global assessment score ≥ 4 according to the recommendations of the Assessments in Ankylosing Spondylitis (ASAS) working group [11]. Patients were invited to present to the centers between the 6-week intervals at any time if symptoms suggestive of relapse or other problems occurred, and if they did, their clinical symptoms were documented accordingly. In cases of relapse, the patients were reinfused with infliximab at 5 mg/kg (TP2) and were then followed up for 12 weeks after the first reinfusion. All the patients were offered an opportunity to enter the next phase of the trial, for another 2 years.
Assessment of the individual disease course after discontinuation
Clinical data were assessed at TP1 and TP2 by use of the standard indicators: disease activity as measured by the BASDAI, C-reactive protein (CRP), and erythrocyte sedimentation rate (ESR). Function was assessed according to the Bath Ankylosing Spondylitis Functional Index (BASFI) [12], and mobility was assessed according to the Bath AS Metrology Index (BASMI) [13]. The patient's global assessment score, the physician's global assessment score, and the numerical rating scale for pain (NRS-P) were each assessed on a numerical rating scale ranging from 0 to 10.
Statistical analysis
The correlation of the data at the two time points was calculated using Pearson's correlation coefficient. The clinical and laboratory data for the patients who experienced a relapse (that is, at TP2) were compared with the data found at TP1.
A Kaplan-Meier survival analysis was used to calculate the probability of a relapse, with duration of response as survival time and relapse as a binomial covariate for the end point. A Cox proportional hazards regression analysis was used to identify possible predictors of flare.
In addition, patients were stratified both according to their BASDAI values at the time of discontinuation, using a cutoff value of 3 at TP1, and also according to the ASAS working group criteria for partial remission at TP1 [14]. Partial remission was defined as a score ≤ 2 (on a scale of 0 to 10) in each of the four ASAS 20% domains, according to the ASAS criteria. The TtR in these groups was compared using a log-rank test. All statistical tests were two-tailed.
Results
Baseline findings, at discontinuation of anti-TNF therapy
Table 1 summarizes the mean ages of the patients, their scores on the various measures of AS, the mean ESR, and the mean CRP concentration at TP1, when anti-TNF treatment was discontinued.
The BASDAI values at TP1 were >3 for 13 (31%) of the 42 patients and >4 for 8 (19%) of the 42. The latter were still receiving treatment, because they had experienced a significant decrease of their BASDAI values, of about 30% compared with their baseline value at the start of the study (mean BASDAI 7.4 at baseline versus 5.3 at TP1) and reported definite subjective improvement. At TP 1, 14 (33%) of the 42 patients were in partial remission [14].
Duration of response after discontinuation
By 3 weeks after the last patient reached TP2, 41 of the 42 patients were being reinfused because of relapse (Fig. 1). Although the first patient reached TP2 after 7 weeks, it took the last patient more than 52 weeks. However, most patients (64%) experienced a flare between week 12 (10/42 patients; 23.8%) and week 24 (37/42 patients; 88.1%). The mean time between TP 1 and TP 2 was 17.5 weeks (± 7.9 weeks, range 7 to 45) and the median time was 15 weeks.
Means and changes of the assessed parameters after discontinuation of treatment
Between TP1 and TP2, the mean increase of the BASDAI was 3.6 (± 1.7), the mean increase of CRP was 17.6 mg/l (± 23.4 mg/l), and the mean increase of the ESR was 21.0 mm/hour (± 29.7 mm/hour), (all P < 0.001 in comparison of TP1 with TP2). All changes between the two time points were statistically significant (Table 1).
Correlations between the individual parameters
The changes in the BASDAI correlated well with the changes in the BASMI (r = 0.35, P = 0.03) and the BASFI (r = 0.79, P < 0.001). The changes in these three indexes correlated well with the changes in the patient's global assessment score (r = 0.81, r = 0.32, and r = 0.74, respectively; all P < 0.05) and in the physician's global assessment score (r = 0.49, r = 0.39, and r = 0.46, respectively; all P < 0.05). The change in the NRS-P correlated well with the change in all clinical findings but not with the laboratory values (data not shown). The TtR was not correlated with any clinical parameter.
Correlations between clinical remission and disease activity and response to discontinuation of treatment
Patients in partial remission at TP1 (n = 15) had a longer duration of response than patients who did not fulfill remission criteria (P = 0.059). The mean TtR was 21.3 weeks (95% confidence interval (CI), 15.5 to 27.2 weeks) for patients in remission but only 15.4 weeks (12.7 to 18.1) for the other group (Fig. 2a).
Similarly, in the analysis of the disease status at TP1, there was also a difference between the patients with low (BASDAI <3) and high (BASDAI ≥ 3) disease activity (Fig. 2b; P = 0.039); the mean TtR of the patients with high disease activity was 14.8 weeks (CI 10.0 to 19.6) and the mean TtR of the patients with low disease activity was 18.9 (CI 15.4 to 22.4). This result was confirmed by a Cox regression analysis. A higher BASDAI, an elevated CRP, older age, and a longer disease duration were associated with a shorter TtR. Three of seven patients with a CRP >6 mg/l at the end of year 3 (TP1) had already experienced a relapse by 12 weeks, and the remaining four patients, by 16 weeks (Fig. 2c; P = 0.009). The cumulative probability of relapse was less in patients with low CRP levels (20% by week 12 and 60% by week 16, respectively) than in patients with elevated CRP levels (43% by week 12 and 100% by week 16, respectively).
Response to retreatment
All 41 patients who were reinfused responded well to the restart of therapy with infliximab. They showed a clear improvement of signs and symptoms and reached a disease state similar to that before the treatment was discontinued. The main inclusion and outcome parameter, the BASDAI, had improved from 6.1 ± 1.4 at TP2 to 3.2 ± 2.6 by 6 weeks after reinfusion and to 2.9 ± 2.1 by 12 weeks after reinfusion, respectively (both P < 0.001). All other parameters improved similarly well in comparison with TP2 (not shown).
There was no adverse event and no other safety concern after resumption of infliximab therapy.
Discussion
Infliximab has proven clinical efficacy in patients with active AS, which is associated with definite improvement of disease activity in both the short and the long term, for up to 3 years [5,6]. Our study is the first to examine the clinical response to discontinuation of long-term infliximab therapy in patients with AS. Several important observations were made.
First, we found that discontinuation of long-term therapy with infliximab in patients with AS leads to a clinical relapse of the disease, with deterioration of signs and symptoms, after several weeks to months. This indicates that the majority of patients may, rather, need continuous anti-TNF therapy.
Another finding is that even though there were relapses eventually, in many patients the low disease activity at discontinuation of therapy persisted for some weeks after discontinuation, although only one patient was in ongoing remission for more than 1 year. The mean duration of ongoing response was almost 4 months. Since the time of persistent clinical efficacy of infliximab after discontinuation varied widely between patients, the optimal dose and the optimal infusion interval for infliximab is also likely to be different from patient to patient. The best dosage probably needs to be defined individually.
We also found that there seem to be predictive factors for the duration of clinical improvement after discontinuation of infliximab therapy in AS patients. The data suggest that clinical improvement persists longer when a state of partial remission, low disease activity, and low CRP levels are present at the time of discontinuation. Thus, the outcome after discontinuation can be partly predicted.
These conclusions are complementary to those predictive of major response that have been reported recently [15]. Overall, it seems that patients who may be candidates for discontinuation or a possible extension of infusion intervals of infliximab therapy have a better outcome if this decision is made while the patients are in a state of low disease activity. Such patients are more likely to have ongoing benefit from previous therapy for several more months.
The favorable response after retreatment argues against an important role of formation of antibodies to infliximab (ATI) in these patients. This response is probably due to the preselection of the patients by the previous 3 years of persistent high-dose therapy with infliximab, which clearly differs from other approaches [16].
Discontinuation of infliximab may become necessary in various patients: those who are in remission for long periods and simply want to test the remission; those who want to become pregnant and wish to exclude the risk of medication toxicity (although there is no indication that infliximab may be harmful); those with more severe or repetitive infection(s); and those who have to undergo surgery (although there is no reason to think that ongoing infliximab therapy may be harmful, good data are lacking).
Another finding of our study is that discontinuation of infliximab therapy seems justified, since we found that retreatment with infliximab was safe, resulting in a good clinical response, similar to that before discontinuation. There was no loss of efficacy and no need for an increased dose after the new start of infliximab therapy. Thus, if for any reason discontinuation of anti-TNF therapy is considered necessary, that seems possible with no major problems regarding efficacy and safety. This may have definite implications for daily practice, since discontinuation of therapy at certain intervals, such as after 1 or 2 years of therapy, may become a standard approach. Payers and patients may want to make sure that further anti-TNF therapy is needed. An intermittent cessation of anti-TNF therapy may be considered in the case of patients who respond well to infliximab therapy for longer periods of time. Since it is unknown how long the patients should receive anti-TNF therapy, it is unclear how to deal with this uncertainty in clinical practice. One possible approach would be to check from time to time whether the disease is still active or has become active again after initial improvement due to infliximab therapy. Another possibility would be to slowly extend the intervals between infusions. This approach would obviously have important economic implications.
However, we think that no clear recommendation for such an approach can be given in the light of present knowledge. More work is needed to confirm our findings and further studies are required to better clarify these issues.
The decision to use a BASDAI cutoff score of 4 is based on the ASAS recommendations. The decision to use a cutoff score of 3 to indicate low disease activity is, at the moment, arbitrary but may serve as a basis for further discussion. It will be especially interesting to learn from the patients whether a score of 3 comes closer to indicating an acceptable state.
Conclusion
Therapy with infliximab has definite long-term clinical efficacy and safety in patients with AS. Patients who discontinue therapy are likely to have a clinical relapse within several weeks to months. Therefore, continuous therapy seems to be necessary for most patients with AS. Importantly, however, we found that retreatment is safe and the clinical efficacy is as good as that before discontinuation. Patients in partial remission or with low disease activity have a longer duration of response after discontinuation than patients with higher disease activity. Overall, anti-TNF therapy is a major step forward in the treatment of patients with AS.
Competing Interests
Dr Braun and Dr Sieper have received reimbursements and fees from the Centocor Amgen, and Wyeth and Abbott.
Abbreviations
AS = ankylosing spondylitis; ASAS = Assessments in Ankylosing Spondylitis [working group]; BASDAI = Bath Ankylosing Spondylitis Disease Activity Index; BASFI = Bath Ankylosing Spondylitis Function Index; BASMI = Bath AS Metrology Index; CI = confidence interval; CRP = C-reactive protein; ESR = erythrocyte sedimentation rate; NRS-P = numerical rating scale for pain; TNF = tumor necrosis factor; TP = time point; TtR = time to relapse.
Authors' contributions
XB: Preparation of data analysis, preparation of the manuscript, study coordination. JL: Data analysis, statistical evaluation. JB: Monitoring and investigation of the patients, study coordination. MR: Monitoring and investigation of the patients, study coordination. JS: Investigator, writing of the manuscript. JB: Idea, writing of the manuscript, principal investigator, responsible for the study. All authors read and approved the final manuscript.
Figures and Tables
Figure 1 Cumulative percentages (confidence intervals) of retreatment after discontinuation of infliximab in patients treated for ankylosing spondylitis. Retreatment depended on the duration of response to the initial treatment. Of the 42 patients, 10 had to be retreated within 12 weeks after discontinuation of infliximab infusions, 37 within 24 weeks, and 38 within 36 weeks. By week 48, 1 of the 42 patients had not needed retreatment and 41 were again receiving infliximab.
Figure 2 Kaplan-Meier analysis of time to relapse in AS patients after discontinuation of infliximab treatment. (a) Cumulative probability of relapse analyzed according to state of remission as measured by ASAS partial remission criteria at TP1. Patients were (bold line) or were not (thin line) in partial remission at TP1. (b) Cumulative probability of relapse according to state of disease activity at TP1 as indicated by a BASDAI ≥ 3 (high disease activity) (thin line) or <3 (low disease activity) (bold line). (c) Cumulative probability of relapse according to state of disease activity at TP1 as indicated by a CRP ≤ 6 mg/l (bold line; low disease activity) or >6 mg/l (thin line; increased disease activity). AS, ankylosing spondylitis; ASAS, Assessments in Ankylosing Spondylitis [working group]; BASDAI, Bath Ankylosing Spondylitis Disease Activity Index; CRP, C-reactive protein; pts., patients; TP1, time point 1 (when infliximab treatment was discontinued).
Table 1 Clinical and laboratory findings for 42 patients with ankylosing spondylitis treated with infliximab
Finding BASDAI BASMI BASFI PatGA PhysGA NRS-P ESR (mm/h) CRP (mg/l)
At time point 1a
Mean ± SD 2.5 ± 1.8** 2.7 ± 2.0* 2.9 ± 2.4** 2.6 ± 1.5** 2.6 ± 2.1** 2.6 ± 2.1** 10.5 ± 7.3** 3.1 ± 4.2**
Median 2.4 2.0 2.5 4.0 2.0 2.0 8.0 1.1
Range 0.0 - 6.8 0.0 - 7.0 0.0 - 8.3 0.0 - 8.0 0.0 - 4.0 0.0 - 7.0 2.0 - 32.0 0.0 - 19.0
At time point 2a
Mean ± SD 6.1 ± 1.4** 3.2 ± 2.2* 5.8 ± 1.8** 6.9 ± 2.1** 7.0 ± 1.5** 7.1 ± 1.7** 31.5 ± 29.7** 20.7 ± 23.7**
Median 6.2 3.0 5.7 7.0 7.0 7.0 23.0 14.0
Range 4.0 - 9.2 0.0 - 9.0 1.2 - 9.1 4.0 - 10.0 4.0 - 10.0 0.0 - 10.0 4.0 - 150.0 0.3 - 126.0
Change between time points 1 and 2
Mean ± SD 3.6 ± 1.7 0.5 ± 1.5 2.9 ± 2.0 4.3 ± 1.9 4.4 ± 1.8 4.5 ± 2.2 21.0 ± 29.7 17.6 ± 23.4
Median 3.6 0.5 2.5 4.0 4.0 4.0 12.0 11.5
Range -1.2 - 6.7 -4.0 - 3.0 -0.5 - 7.8 -2.0 - 8.0 -2.0 - 8.0 -1.0 - 8.0 -6.0 - 146.0 -6.3 - 123.0
aTime point 1 is the time point at which infliximab treatment was discontinued; time point 2 is that when retreatment began. *P < 0.05, **P < 0.001, when means at time points 1 and 2 are compared. BASDAI, Bath Ankylosing Spondylitis Disease Activity Index; BASFI, Bath Ankylosing Spondylitis Function Index; BASMI, Bath Ankylosing Spondylitis Metrology Index; CRP, C-reactive protein; ESR, erythrocyte sedimentation rate; NRS-P, numerical rating scale for pain; PatGA, patient's global assessment; PhysGA, physician's global assessment; SD, standard deviation.
==== Refs
Braun J Sieper J The sacroiliac joint in the spondyloarthropathies Curr Opin Rheumatol 1996 8 275 287 8864578
Braun J Bollow M Eggens U Konig H Distler A Sieper J Use of dynamic magnetic resonance imaging with fast imaging in the detection of early and advanced sacroiliitis in spondylarthropathy patients Arthritis Rheum 1994 37 1039 1045 8024613
McGonagle D Gibbon W O'Connor P Green M Pease C Emery P Characteristic magnetic resonance imaging entheseal changes of knee synovitis in spondylarthropathy Arthritis Rheum 1998 41 694 700 9550479 10.1002/1529-0131(199804)41:4<694::AID-ART17>3.0.CO;2-#
Braun J Brandt J Listing J Zink A Alten R Golder W Gromnica-Ihle E Kellner H Krause A Schneider M Treatment of active ankylosing spondylitis with infliximab: a randomised controlled multicentre trial Lancet 2002 359 1187 1193 11955536 10.1016/S0140-6736(02)08215-6
Braun J Brandt J Listing J Zink A Alten R Burmester G Golder W Gromnica-Ihle E Kellner H Schneider M Two-year maintenance of efficacy and safety of infliximab in the treatment of ankylosing spondylitis Ann Rheum Dis 2005 64 229 234 15388511 10.1136/ard.2004.025130
Braun J Baraliakos X Brandt J Listing J Zink A Alten R Burmester G Golder W Gromnica-Ihle E Kellner H Persistent clinical response to the anti-TNFa antibody infliximab in patients with ankylosing spondylitis over 3 years J Rheumatol 2005
Braun J Baraliakos X Golder W Brandt J Rudwaleit M Listing J Bollow M Sieper J Van Der Heijde D Magnetic resonance imaging examinations of the spine in patients with ankylosing spondylitis, before and after successful therapy with infliximab: evaluation of a new scoring system Arthritis Rheum 2003 48 1126 1136 12687557 10.1002/art.10883
Brandt J Haibel H Sieper J Reddig J Braun J Infliximab treatment of severe ankylosing spondylitis: one-year followup Arthritis Rheum 2001 44 2936 2937 11762956 10.1002/1529-0131(200112)44:12<2936::AID-ART483>3.0.CO;2-M
Braun J Brandt J Listing J Zink A Alten R Burmester G Golder W Gromnica-Ihle E Kellner H Schneider M Long-term efficacy and safety of infliximab in the treatment of ankylosing spondylitis: an open, observational, extension study of a three-month, randomized, placebo-controlled trial Arthritis Rheum 2003 48 2224 2233 12905476 10.1002/art.11104
Garrett S Jenkinson T Kennedy LG Whitelock H Gaisford P Calin A A new approach to defining disease status in ankylosing spondylitis: the Bath Ankylosing Spondylitis Disease Activity Index J Rheumatol 1994 21 2286 2291 7699630
Braun J Pham T Sieper J Davis J van der Linden S Dougados M van der Heijde D International ASAS consensus statement for the use of anti-tumour necrosis factor agents in patients with ankylosing spondylitis Ann Rheum Dis 2003 62 817 824 12922952 10.1136/ard.62.9.817
Calin A Garrett S Whitelock H Kennedy LG O'Hea J Mallorie P Jenkinson T A new approach to defining functional ability in ankylosing spondylitis: the development of the Bath Ankylosing Spondylitis Functional Index J Rheumatol 1994 21 2281 2285 7699629
Jenkinson TR Mallorie PA Whitelock HC Kennedy LG Garrett SL Calin A Defining spinal mobility in ankylosing spondylitis (AS). The Bath AS Metrology Index J Rheumatol 1994 21 1694 1698 7799351
Anderson JJ Baron G van der Heijde D Felson DT Dougados M Ankylosing spondylitis assessment group preliminary definition of short-term improvement in ankylosing spondylitis Arthritis Rheum 2001 44 1876 1886 11508441 10.1002/1529-0131(200108)44:8<1876::AID-ART326>3.0.CO;2-F
Rudwaleit M Listing J Brandt J Braun J Sieper J Prediction of a major clinical response (BASDAI 50) to tumor necrosis factor alpha blockers in ankylosing spondylitis Ann Rheum Dis 2004 63 665 670 15037444 10.1136/ard.2003.016386
Baert F Noman M Vermeire S Van Assche G G DH Carbonez A Rutgeerts P Influence of immunogenicity on the long-term efficacy of infliximab in Crohn's disease N Engl J Med 2003 348 601 608 12584368 10.1056/NEJMoa020888
| 15899030 | PMC1174938 | CC BY | 2021-01-04 16:02:36 | no | Arthritis Res Ther. 2005 Feb 21; 7(3):R439-R444 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1693 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar16941589903410.1186/ar1694Research ArticleSmall GTP-binding protein Rho-mediated signaling promotes proliferation of rheumatoid synovial fibroblasts Nakayamada Shingo [email protected] Hitoshi [email protected] Kazuyoshi [email protected] Akira [email protected] Yoshiya [email protected] First Department of Internal Medicine, University of Occupational and Environmental Health, School of Medicine, Fukuoka, Japan2 Department of Pharmacology and Toxicology, Graduate School of Pharmaceutical Sciences, Kyushu University, Fukuoka, Japan3 Pharmaceuticals Research Unit, Research & Development Division, Mitsubishi Pharma Corporation, Yokohama, Japan2005 18 2 2005 7 3 R476 R484 11 11 2004 24 11 2004 10 1 2005 18 1 2005 Copyright © 2005 Nakayamada et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Rho is a major small GTP-binding protein that is involved in the regulation of various cell functions, including proliferation and cell migration, through activation of multiple signaling molecules in various types of cells. We studied its roles in synovial fibroblasts (SFs) in patients with rheumatoid arthritis (RA) and clarified its relevance to RA synovitis, with the following results. 1)We found that the thrombin receptor was overexpressed on RA synovial fibroblasts (RA SFs) and that thrombin induced a marked proliferation and progression of the cell cycle to the S phase in these cells. 2)We also found that thrombin efficiently activated Rho. 3)Rho activation and proliferation and the progression of the cell cycle to the S phase were completely blocked by p115RGS (an N-terminal regulator of the G-protein signaling domain of p115RhoGEF) and by the C-terminal fragments of Gα13 (an inhibitor of the interaction of receptors with G13). 4)Thrombin induced the secretion of IL-6 by RA SFs, but this action was blocked by p115RGS or Gα13. Our findings show that the actions of thrombin on the proliferation of RA SFs, cell-cycle progression to the S phase, and IL-6 secretion were mainly mediated by the G13 and RhoGEF pathways. These results suggest that p115RGS and Gα13 could be potent inhibitors of such functions. A rational design of future therapeutic strategies for RA synovitis could perhaps include the exploitation of the Rho pathway to directly reduce the growth of synovial cells.
==== Body
Introduction
Rheumatoid arthritis (RA) is characterized by synovial proliferation, neovascularization, and accumulation of inflammatory cells in inflamed joints. Synovial cells are markedly activated by cytokines, adhesion molecules, and coagulation factors, resulting in hyperplasia of the synovial tissue, and the activated synovial cells produce inflammatory cytokines and degradative enzymes. These pathological processes in RA synoviocytes are tightly regulated by intracellular signaling. The small GTPase Rho is a pivotal regulator of several signaling pathways, including the remodeling of actin cytoskeleton, transcriptional regulation, and cell-cycle progression [1-4]. Like other regulatory GTPases, Rho requires GDP/GTP exchange dependent on guanine nucleotide exchange factors (GEFs) for its activation [5]. GEFs are critical regulators of Rho activation and thereby control a variety of cellular responses such as cell proliferation and cytokine production. However, the relevance of Rho-mediated signaling to inflammatory processes in RA is largely unknown.
Among the various stimuli that activate the GEF–Rho pathway, thrombin is the best-known activator through the following sequence of events: binding thrombin to protease-activated receptor-1 including a thrombin receptor; activation of heterotrimeric G proteins Gq, Gi, and G12/13 [6-8]; activation of p115RhoGEF by the α subunit of G12/13; binding of a Rho-specific GEF containing a Dbl homology domain to Rho; and activation of Rho by GDP/GTP exchange [9-12].
Recent studies have indicated that Rho regulates cellular functions in inflammatory cells [13-16]. Rho GTPases have been implicated in the regulation of cell proliferation and IL-2 production in T cells [13,17,18]. RA is a representative inflammatory disease and is characterized by accumulation of T cells and proliferation of synovial fibroblasts [19,20]. Although many molecules, including inflammatory cytokines such as IL-1, tumor necrosis factor (TNF), and IL-6 and growth factors, have been implicated as pathogenic factors in RA, the coagulation system is also involved in the inflammatory processes in RA synovitis. High levels of various clotting and fibrinolytic factors such as thrombin are found in the synovial fluid of patients with RA [21-24], and high concentrations of thrombin are detected in RA synovial tissue [25,26]. Moreover, thrombin promotes chemotaxis and adhesion of inflammatory cells such as lymphocytes and the production of various proinflammatory molecules [25,27,28]. Thrombin may therefore play an important pathological role in RA synovitis.
The aim of the present study was to determine the role of Rho-mediated signaling in the regulation of synovial proliferation and cytokine production in RA SFs. The results indicate that thrombin stimulation induces proliferation and IL-6 secretion by RA SFs through G13 and Rho pathways and suggest that the G13–GEF–Rho pathway plays an important role in the RA inflammatory process.
Materials and methods
The study protocol was approved by the Human Ethics Review Committee of the University of Occupational and Environmental Health, Japan, and we obtained a signed consent form from each subject before taking tissue samples used in the present study.
Synovial tissues and culture of synovial fibroblasts
Synovial tissues were obtained from five women (aged 45 to 66 years) with active RA or osteoarthritis (OA) whose disease had been diagnosed according to the criteria of the American College of Rheumatology [29-32] and who were treated by joint replacement surgery. All the enrolled patients with RA had more than six swollen joints, more than three tender joints, and an erythrocyte sedimentation rate (Westergren) of >28 mm/hour.
Samples were dissected under sterile conditions in PBS and were immediately prepared for culture of fibroblast-like synovial cells. Briefly, the tissue samples were minced into small pieces and digested with collagenase (Sigma Aldrich, Tokyo, Japan) in serum-free DMEM (Gibco BRL, Grand Island, NY, USA). The cells were filtered through a nylon mesh and then were washed extensively and suspended in DMEM supplemented with 10% FCS (Bio-Pro, Karlsruhe, Germany) and streptomycin/penicillin (10 units/ml; Sigma Aldrich). Finally, isolated cells were seeded in 25-cm2 culture flasks (Falcon, Lincoln Park, NJ, USA) and cultured in a humidified 5% CO2 atmosphere. After overnight culture, nonadherent cells were removed and incubation of adherent cells was continued in fresh medium. At confluence, the cells were trypsinized, passaged at a 1:3 split ratio, and recultured. The medium was changed twice each week, and the cells were used after 2 to 5 passages. We characterized cultured synovial cells derived from the synovium of RA patients. The cells were spindle-shaped and grew in a cobblestone pattern. Flow cytometric analysis of these cells indicated that they lacked macrophage markers such as major-histocompatibility-complex class II antigens CD14 and CD11b (data not shown). Thus, RA synovial cells are type B synovial-fibroblast-like cells.
Materials
Human thrombin was purchased from Sigma Aldrich. The following mAbs were used: fluorescein-isothiocyanate-conjugated control mAb anti-Thy 1.2 (Becton Dickinson, San Jose, CA, USA) and antithrombin receptor mAb ATAP2 (Santa Cruz Biotechnology, Santa Cruz, CA, USA). A Rho activation kit containing glutathione S-transferase (GST)-Rhotekin-Rho-binding domain (RBD) (GST-RBD) beads was purchased from Cytoskeleton (Denver, CO, USA).
Adenoviral infection
Recombinant adenoviruses encoding green-fluorescent protein (GFP), the C-terminal regions of Gα12 (Gα12-ct), the C-terminal regions of Gα13 (Gα13-ct), and the regulator of the G-protein signaling domain of p115RhoGEF (p115RGS) were produced as described previously [33]. RA SFs were plated onto a six-well culture dish and cultured in DMEM containing 10% FCS. After 24 hours, the cells were infected with recombinant adenoviruses at a multiplicity of infection of 30 for 1 hour at 37°C. Cells infected or not infected with adenovirus were then starved in DMEM with 1% FCS and cultured for an additional 48 hours before treatment. Under these conditions, infection with adenoviruses coding for GFP made almost 100% of cells GFP-positive. None of these vectors produced cytotoxic effects on RA SFs until 96 hours after infection, as confirmed by trypan blue staining (data not shown).
Flow microfluorometry
Staining and flow-cytometric analysis of RA SFs were conducted by standard procedures, as described previously [34], using a FACScan (Becton Dickinson, Mountain View, CA, USA). Briefly, cells (2 × 105) were incubated with fluorescein-isothiocyanate-conjugated negative control mAb anti-Thy-1.2 or antithrombin receptor mAb at saturating concentrations in fluorescence-activated cell sorter (FACS) medium consisting of Hanks' balanced salt solution (Nissui, Tokyo, Japan), 0.5% human serum albumin (Mitsubishi Pharma, Osaka, Japan), and 0.2% NaN3 (Sigma Aldrich) for 30 min at 4°C. After three washes in FACS medium, the cells were analyzed with the FACScan. Cell-surface antigens on single cells were quantified using standard beads, QIFKIT (Dako Japan, Kyoto, Japan), as described previously [35,36]. The data were used to construct the calibration curve of mean fluorescence intensity versus antibody-binding capacity. The cell specimen was analyzed on the FACScan and antibody-binding capacity calculated by interpolation on the calibration curve. When green-fluorescence laser detection was set at 450 nm in the FACScan used, antibody-binding capacity = 414.45 × exp (0.0092 × Mean fluorescence intensity) (R2 = 0.9999). Subsequently, specific antibody-binding capacity was obtained after corrections for background, apparent antibody-binding capacity of the negative control mAb anti-Thy-1.2. The specific antibody-binding capacity is the mean number of accessible antigenic sites per cell, referred to as antigen density and expressed in sites/cell.
Proliferation assay
RA SFs (1 × 104) infected with or free of adenoviruses were seeded and incubated in 96-well flat-bottomed microfilter plates (Costar, Cambridge, MA, USA) in DMEM containing 1% FCS for 48 hours at 37°C and were then stimulated with the indicated amount of thrombin. At 24 hours after the thrombin stimulation, cells were stained with TetraColor One (Seikagaku, Tokyo, Japan) including tetrazolium and an electron-carrier mixture for detecting cell proliferation. After the cells had been stained in this way for 1 hour at 37°C, the optical density value of each well was measured using an ELISA plate reader at 450 nm.
Cell-cycle analysis
RA SFs infected with or free of adenoviruses were cultured for 48 hours in DMEM containing 1% FCS and then stimulated with the indicated amount of thrombin. At 24 hours after thrombin stimulation, the cells were collected, washed with PBS, and fixed in 70% ethanol for 2 hours at 4°C. After treatment of cells with 10 μg/ml ribonuclease (Wako, Osaka, Japan) for 15 min at 37°C, the cells were stained with 50 μg/ml propidium iodide (Sigma Aldrich) for 2 minutes. The DNA content was subsequently measured by FACScan.
Rho activation assay
Rho activation was determined by a pull-down assay using GST-RBD beads [37,38]. Forty-eight hours after adenovirus infection, RA SFs were stimulated with 10 units/ml thrombin, quickly washed with ice-cold Tris-buffered saline, and lysed in 500 μl of lysis buffer (50 mM Tris, pH 7.5, 10 mM MgCl2, 0.5 M NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, 500 μg/ml tosyl arginine methyl ester, and 10 μg/ml each of leupeptin and aprotinin). Cell lysates were immediately centrifuged at 8,000 rpm at 4°C for 5 min and equal volumes of lysates were incubated with 30 μg GST-RBD beads for 1 hour at 4°C. The beads were washed twice with wash buffer (in mmoles: 25 Tris, pH 7.5; 30 MgCl2; 40 NaCl), and bound Rho was eluted by boiling each sample in Laemmli sample buffer. Eluted samples from the beads and total cell lysate were then electrophoresed on 12% SDS–PAGE gels, transferred to nitrocellulose, blocked with 5% nonfat milk, and analyzed by western blotting using a polyclonal anti-Rho antibody.
Statistical analysis
Data are expressed as mean ± standard deviation (SD) of the number of indicated patients. Differences from the control were examined for statistical significance by the Mann–Whitney U test. A P value less than 0.05 denoted the presence of a statistically significant difference.
Results
High expression of thrombin receptor on RA SFs
First, we assessed the expression of thrombin receptor on synovial fibroblasts using FACScan. Fig. 1a shows the histogram of thrombin receptor expressed on RA and OA synovial fibroblasts. Although thrombin receptor was expressed on both types of synovial fibroblasts, its level was significantly higher in RA than OA fibroblasts (Fig. 1b). These results were identified in synovial fibroblasts from patients with RA and OA (n = 5 each).
Thrombin induces synovial proliferation and S-phase progression of the cell cycle in RA SFs
To assess the effect of thrombin on the proliferation of RA SFs, we performed proliferation assay and cell-cycle analysis. After cells were starved for 48 hours in DMEM containing 1% FCS, cells were stimulated with the indicated amount of thrombin for 24 hours. Thrombin significantly induced cell proliferation in a dose-dependent manner (Fig. 2a). As shown in Fig. 2b, the vast majority of the starved cells existed at the propidium-iodide-low G0/G1 phase and showed little progression to S phase. However, thrombin significantly increased the S/G2/M phase of the cell in a dose-dependent manner (Fig. 2b,c). The maximum effects of thrombin on cell proliferation and progression to S phase were noted at 10 units/ml. These results indicate that thrombin acts as an important stimulator of RA synovial proliferation.
We next compared cell growth and the cell cycle of RA SFs with those of OA SFs. As shown in Fig. 3a, thrombin induced cell growth and cell-cycle progress in both RA and OA SFs, whereas unstimulated cells did not proliferate well. However, thrombin-induced proliferation of RA SFs was significantly higher than that of OA SFs after incubation for 24 or 48 hours. As shown in Fig. 3b, thrombin induced cell-cycle progress to S phase in both RA and OA SFs, but the responses of RA SFs were significantly higher than those of OA SFs after 24 hours of incubation.
Involvement of small GTP-binding protein Rho activation in thrombin-induced signaling in RA SFs
Thrombin is known to bind to protease-activated receptor-1 such as thrombin receptor, and binding of thrombin to receptors leads to activation of the G protein Gα13 and induces activation of p115RhoGEF, a Rho-specific GEF, and thereby activates Rho [9-12]. We measured Rho activation and its inhibition in RA SFs using the GST-Rhotekin fusion protein. We used the C-terminal regions of Gα12 and Gα13, which inhibit Gα12 and Gα13 from coupling with each receptor, or the regulator of the G-protein signaling domain of p115RhoGEF, which inhibits endogenous p115RhoGEF function by blocking the interaction of p115RhoGEF with Gα12/13 and by its GTPase-activating-protein activity on Gα12/13 [33,39,40]. After adenoviral infection, cells were stimulated with 10 units/ml thrombin, and the lysates from cells were incubated with Rhotekin bound to GST beads to determine Rho activation. As shown in Fig. 4, thrombin stimulation increased the amount of activated Rho in RA SFs infected or not infected with adenoviruses encoding control vector, suggesting that thrombin activates Rho in RA SFs. However, the expression of Gα13-ct or p115RGS completely prevented thrombin-induced Rho activation (Fig. 4). These results suggest that Rho is involved in signaling via thrombin-stimulation, which leads to synovial proliferation.
Involvement of G protein Gα13 and Rho-GEF in thrombin-induced proliferation of RA SFs
Thrombin (10 unit/ml) significantly induced synovial proliferation and S-phase progression in RA SFs infected or not infected with adenoviruses encoding control vector and Gα12-ct (Fig. 5a,b). In contrast, thrombin failed to induce both cell proliferation and S-phase progression in RA SFs expressing Gα13-ct and p115RGS (Fig. 5a,b). These data suggest that Gα13 (but not Gα12) and RhoGEF are involved in signaling via thrombin-stimulation, which leads to synovial proliferation.
Induction of IL-6 secretion via Rho-mediated signaling in RA SFs
Finally, we assessed IL-6 secretion by RA SFs, using ELISA. Fig. 6 shows the concentrations of IL-6 in supernatants of thrombin-stimulated RA SFs. Stimulation with 10 units/ml thrombin significantly increased IL-6 secretion at 6 hours (Fig. 6b), and this effect was dose-dependent (Fig. 6a). The same dose of thrombin produced a significant induction of IL-6 secretion by RA SFs infected or not infected with adenoviruses encoding control vector (Fig. 6c). However, the thrombin-induced IL-6 secretion in RA SFs transfected with Gα12-ct and Gα13-ct was partially reduced and that in RA SFs expressing p115RGS was markedly inhibited (Fig. 6c). These results suggest that thrombin-induced IL-6 secretion by RA SFs is mainly mediated through RhoGEF.
Discussion
The multiple functions of synovial cells including proliferation, apoptosis, adhesion, and cytokine production are induced by intracellular signaling, which plays a pivotal role in the pathological processes of RA, a representative inflammatory disease. Among the signaling molecules, we document here the relevance of Rho to the pathogenesis of RA synovitis, based on the following results: high expression of thrombin receptor on RA SFs and thrombin markedly increased the proliferation of these cells and progression of the cell cycle to S phase; thrombin induced the activation of Rho; Rho activation as well as proliferation and S-phase progression were completely blocked by either p115RGS or Gα13-ct; and thrombin-induced IL-6 secretion was also reduced by p115RGS and Gα13-ct.
Rho is activated by the G12/13 family of heterotrimeric GTP-binding proteins through the stimulation of GEF activity of p115RhoGEF [12,39]. However, p115RhoGEF can activate as well as inhibit Rho signaling after stimulation of protease-activated receptor-1 [9,41]. p115RhoGEF also contains the regulator of G-protein signaling (RGS) domain at its N terminus, through which it interacts with Gα12/13 and functions as a GTPase-activating protein for G12/13 [39,42]. Furthermore, several groups have reported that C-terminal fragments of Gα12 and Gα13 can inhibit the interaction of receptors with G12 and G13, respectively [33,40,43]. Therefore, we used p115RGS domains and C-terminal fragments of Gα subunits as inhibitors to analyze Gα12- and Gα13-mediated signaling pathways. Using the N-terminal RGS domain of p115RhoGEF, we observed that thrombin activated Rho-dependent signaling and induced synovial proliferation via the G13 pathway.
Previous studies postulated that local fibrin deposition promotes inflammation and tissue destruction, based on the findings of activation of the coagulation system and local generation of fibrin in inflamed arthritic joints [25,44]. Furthermore, recent studies indicate that the coagulation system is closely associated with the inflammation of RA. The level of thrombin, a ligand for G13, is markedly increased in the synovial fluid and tissue of RA patients, compared with those of OA patients, and significantly correlates with RA activity [24]. Our results also indicated that the expression of the thrombin receptor was significantly higher in RA than OA synovial fibroblasts. Since the expression of thrombin receptor is up-regulated by thrombin itself, it is conceivable that up-regulation of thrombin receptor in RA SFs is a natural consequence of exposure to extravasated plasma thrombin and tissue remodeling during the inflammatory response [22,44,45]. Since thrombin induced cell growth and cell-cycle progress of both RA and OA SFs, thrombin-mediated activation of fibroblasts may not be specific for RA SFs. However, responses of RA SFs to thrombin were significantly higher than those of OA SFs, suggesting that the thrombin–Rho pathway could be activated in RA SF.
In the present study, we observed failure of thrombin to induce both cell proliferation and S-phase progression in RA SFs that expressed Gα13-ct and p115RGS, but not Gα12-ct. These data suggest that G13 and p115RhoGEF, which is directly stimulated by G13, are involved in signaling via thrombin-stimulation, subsequent Rho activation, and synovial proliferation.
IL-6 plays an important role in the pathogenesis of RA [46-51], since it is induced by a variety of stimuli such as IL-1 and TNF, is produced abundantly in RA synovium, and is detected at high concentrations in the synovial fluid and serum of RA. Among the various inflammatory cytokines, production of IL-1α, IL-1β, and TNF-α from RA SFs did not change after thrombin stimulation, as reported previously [25]. The obtained results showed that thrombin markedly induced IL-6 secretion from RA SFs. The thrombin-induced IL-6 secretion was completely inhibited in RA SFs expressing p115RGS, whereas it was partially suppressed in the cells that expressed Gα12-ct and Gα13-ct, suggesting the possible existence of another pathway for the activation of RhoGEF during IL-6 secretion in RA SFs.
Considering these findings all together, we conclude that Rho plays a key role in synovial proliferation, S-phase cell-cycle progression, and IL-6 secretion by thrombin-stimulated RA SFs during the pathological process of synovial inflammation. Furthermore, because p115RGS and Gα13 appear to be potent inhibitors of these cellular functions by targeting the thrombin–G13–GEF–Rho pathway, a rational design of future therapeutic strategies for RA synovitis could perhaps include the exploitation of the Rho pathway to directly reduce synovial cell growth in vivo.
Conclusion
Our results indicate that stimulation with thrombin induced proliferation and IL-6 secretion by RA SFs through G13 and Rho pathways and suggest that the G13–GEF–Rho pathway plays an important role in the RA inflammatory process. A rational design of future therapeutic strategies for RA synovitis could perhaps include the exploitation of the Rho pathway to directly reduce synovial cell growth.
Abbreviations
DMEM = Dulbecco's modified Eagle's medium; ELISA = enzyme-linked immunosorbent assay; FACS = fluorescence-activated cell sorter; FCS = fetal calf serum; Gα12-ct = the carboxy-terminal regions of Gα12; Gα13-ct = the carboxy-terminal regions of Gα13; GAP = GTPase-activating protein; GEF = guanine nucleotide exchange factor; GFP = green-fluorescent protein; GST = glutathione S-transferase; IL = interleukin; mAb = monoclonal antibody; OA = osteoarthritis; p115RGS = regulator of G-protein signaling domain of p115RhoGEF; PBS = phosphate-buffered saline; RA = rheumatoid arthritis; RBD = Rhotekin-Rho-binding domain; RGS = regulator of G-protein signaling; SD = standard deviation; SF = synovial fibroblast; TNF = tumor necrosis factor.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SN designed the experimental design of the study, carried out the experiments, and drafted the manuscript. HK provided adenovirus and participated in the preparation of the manuscript. KS participated in the experimental design of the study. AM performed statistical analyses and participated in the design of the study. YT conceived of the study, participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.
Acknowledgements
We thank Ms T Adachi for her excellent technical assistance. This work was supported in part by a Research Grant-In-Aid for Scientific Research by the Ministry of Health, Labor and Welfare of Japan; the Ministry of Education, Culture, Sports, Science and Technology of Japan; and the University of Occupational and Environmental Health, Japan.
Figures and Tables
Figure 1 Overexpression of thrombin receptor on SFs from patients with RA. (a) Histograms representing thrombin receptor expression on synovial fibroblasts (SFs) from rheumatoid arthritis (RA) and osteoarthritis (OA) patients. Cells were stained with the antithrombin receptor mAb ATAP2. Flow-cytometric analyses were performed using FACScan. Open histograms represent the number of cells stained with ATAP2 in each logarithmic scale on a fluorescence amplifier. Shaded histograms represent profiles of anti-Thy1.2 mAb as a negative control. (b) Comparison of thrombin receptor expression in RA and OA SFs. The expression of thrombin receptor was analyzed by FACScan. Each value represents the number of molecules expressed per cell, calculated using standard QIFKIT beads from five similar experiments, as described in Materials and methods. Data are expressed as mean ± standard deviation for five independent donors. FACS, fluorescence-activated cell sorter. **P < 0.01.
Figure 2 Thrombin induces synovial proliferation and progression to S phase in RA SFs. Cells were cultured for 48 hours in DMEM containing 1% FCS and then stimulated with the indicated amount of thrombin. For the proliferation assay, at 24 hours after thrombin stimulation, cells were stained with TetraColor One including tetrazolium and electron-carrier mixture for detecting cell proliferation. The optical density (OD) was measured by ELISA plate reader at 450 nm. For analysis of the cell cycle, at 24 hours after thrombin stimulation, cells were collected, washed with PBS, and fixed in 70% ethanol for 2 hours at 4°C. After treatment of cells with 10 μg/ml ribonuclease for 15 min at 37°C, fixed cells were stained with 50 μg/ml propidium iodide for 2 min. The DNA content was subsequently measured by FACScan fluorescence-activated cell sorter. (a) Dose-dependent proliferation of synovial fibroblasts (SFs) from rheumatoid arthritis (RA) patients. The OD was measured by ELISA plate reader at 450 nm. (b) Histogram representing the cell cycle in RA SFs, as detected by FACScan. (c) Dose-dependent S-phase progression of the cell cycle in RA SFs. Numbers represent the percentage of cells exhibiting mean channel fluorescence (FL2-H) in the S/G2/M phase of the cell. Data are expressed as mean ± standard deviation for five experiments, using five independent donors. **P < 0.01 in comparison with the value found without thrombin stimulation.
Figure 3 Effects of thrombin on proliferation and cell cycle in SFs from RA and OA. Cells were cultured for 48 hours in DMEM containing 1% FCS and then stimulated with 10 units/ml thrombin. (a) Time-course of proliferation of synovial fibroblasts (SFs) from patients with rheumatoid arthritis (RA) (dark symbols) and patients with osteoarthritis (OA) (light symbols). Continuous lines, thrombin-stimulated cells; dotted lines, unstimulated cells. The optical density (OD) was measured by ELISA plate reader at 450 nm. (b) Comparison of cell cycle in RA SFs and OA SFs. Numbers represent the percentage of cells exhibiting mean channel fluorescence (FL2-H) in the S/G2/M phase of the cell-division cycle. Data are expressed as mean ± standard deviation of five experiments, using five independent donors. **P < 0.01.
Figure 4 Inhibition of thrombin-induced Rho activation by expression of Gα13-ct and p115RGS in RA synovial fibroblasts. Rheumatoid arthritis (RA) synovial fibroblasts (SFs) were or were not infected with adenoviruses encoding green-fluorescent protein (GFP) (control vector), the C-terminal regions of Gα13 (Gα13-ct), or P115RGS. Cells that were not infected with adenovirus were incubated in the medium alone. Cells were then cultured for 48 hours in DMEM containing 1% FCS, then stimulated with 10 units/ml thrombin for 1 minute or were loaded with GTPγS (positive control), after which they were lysed to measure Rho activity. Rho activity is indicated by the amount of Rho bound by the Rhotekin-Rho-binding domain (RBD) (top). The percentage of activated Rho (graph) is expressed as a ratio relative to 4% of total Rho (4% of total protein used in the RBD bead pull-down experiments). Results are representative of three experiments. Western blot analysis confirmed that equal amounts of total Rho were used for the pull-down assay under each condition (data not shown). (-), cells without infection; p115RGS, regulator of G-protein signaling domain of p115Rho guanine nucleotide exchange factor.
Figure 5 G13 and Rho signaling in thrombin-mediated synovial proliferation and S-phase progression in RA SFs. Rheumatoid arthritis (RA) synovial fibroblasts (SFs) infected or not infected with adenoviruses encoding GFP (control vector), the C-terminal regions of Gα12 (Gα12-ct), Gα13-ct, or p115RGS were cultured for 48 hours in DMEM containing 1% FCS and then stimulated with 10 units/ml thrombin. At 24 hours after the thrombin stimulation, proliferation assay and cell-cycle analysis of RA SFs were performed. (a) Effect of Rho signaling inhibition on thrombin-induced cell proliferation. Numbers represent the optical density (OD) as measured by ELISA plate reader at 450 nm. (b) Effect of Rho signaling inhibition on thrombin-induced cell-cycle progression. Numbers represent the percentage of cells that exhibited mean channel fluorescence (FL2-H) in the S/G2/M phase. Data are expressed as mean ± standard deviation of five experiments, using five independent donors. (-), cells without infection; p115RGS, regulator of G-protein signaling domain of p115Rho guanine nucleotide exchange factor. *P < 0.05, **P < 0.01, in comparison with thrombin stimulation.
Figure 6 Thrombin induces IL-6 secretion via Rho-mediated signaling in RA synovial fibroblasts. Rheumatoid arthritis (RA) synovial fibroblasts (SFs) were or were not infected with adenoviruses encoding green-fluorescent protein (control vector), the C-terminal regions of Gα12 (Gα12-ct), Gα13-ct, or p115RGS. Cells were cultured for 48 hours in DMEM containing 1% FCS, then stimulated with the indicated amount of thrombin. At 3 to 24 hours after thrombin stimulation, the supernatants of cultured cells were collected and assayed for IL-6 using commercial ELISA kits. (a) Dose-dependent IL-6 production by RA SFs at 12 hours after thrombin stimulation. (b) Time course of IL-6 secretion by RA SFs stimulated with 10 units/ml thrombin. (c) Effect of Rho signaling inhibition on thrombin-induced IL-6 secretion at 12 hours after thrombin stimulation. Data are expressed as mean ± standard deviation of five experiments, using five independent donors. (-), cells without infection; p115RGS, regulator of the G-protein signaling domain of p115Rho guanine nucleotide exchange factor. *P < 0.05, **P < 0.01, in comparison with (a) time 0, (b) no thrombin stimulation, and (c) the indicated data.
==== Refs
Etienne-Manneville S Hall A Rho GTPases in cell biology Nature 2002 420 629 635 12478284 10.1038/nature01148
Bar-Sagi D Hall A Ras and Rho GTPases: a family reunion Cell 2000 13 227 238 10.1016/S0092-8674(00)00115-X
Yamamoto M Marui N Sakai T Morii N Kozaki S Ikai K Imamura S Narumiya S ADP-ribosylation of the rhoA gene product by botulinum C3 exoenzyme causes Swiss 3T3 cells to accumulate in the G1 phase of the cell cycle Oncogene 1993 8 1449 1455 8502473
Rodriguez-Fernandez JL Sanchez-Martin L Rey M Vicente-Manzanares M Narumiya S Teixido J Sanchez-Madrid F Cabanas C Rho and Rho-associated kinase modulate the tyrosine kinase PYK2 in T-cells through regulation of the activity of the integrin LFA-1 J Biol Chem 2001 276 40518 40527 11489881 10.1074/jbc.M102896200
Van Aelst L D'Souza-Schorey C Rho GTPases and signaling networks Genes Dev 1997 11 2295 2322 9308960
Coughlin SR How the protease thrombin talks to cells Proc Natl Acad Sci USA 1999 96 11023 11027 10500117 10.1073/pnas.96.20.11023
Barr AJ Brass LF Manning DR Reconstitution of receptors and GTP-binding regulatory proteins (G proteins) in Sf9 cells. A direct evaluation of selectivity in receptor. G protein coupling J Biol Chem 1997 272 2223 2229 8999927 10.1074/jbc.272.52.32979
Vogt S Grosse R Schultz G Offermanns S Receptor-dependent RhoA activation in G12/G13-deficient cells: genetic evidence for an involvement of Gq/G11 J Biol Chem 2003 278 28743 28749 12771155 10.1074/jbc.M304570200
Holinstat M Mehta D Kozasa T Minshall RD Malik AB Protein kinase Calpha-induced p115RhoGEF phosphorylation signals endothelial cytoskeletal rearrangement J Biol Chem 2003 278 28793 28798 12754211 10.1074/jbc.M303900200
Hart MJ Jang X Kozasa T Roscoe W Singer WD Gilman AG Sternweis PC Bollag G Direct stimulation of the guanine nucleotide exchange activity of p115 RhoGEF by Galpha13 Science 1998 280 2112 2114 9641916 10.1126/science.280.5372.2112
Bhattacharyya R Wedegaertner PB Galpha 13 requires palmitoylation for plasma membrane localization, Rho-dependent signaling, and promotion of p115-RhoGEF membrane binding J Biol Chem 2000 275 14992 14999 10747909 10.1074/jbc.M000415200
Wells CD Liu MY Jackson M Gutowski S Sternweis PM Rothstein JD Kozasa T Sternweis PC Mechanisms for reversible regulation between G13 and Rho exchange factors J Biol Chem 2002 277 1174 1181 11698392 10.1074/jbc.M105274200
Tharaux PL Bukoski RC Rocha PN Crowley SD Ruiz P Nataraj C Howell DN Kaibuchi K Spurney RF Coffman TM Rho kinase promotes alloimmune responses by regulating the proliferation and structure of T cells J Immunol 2003 171 96 105 12816987
Salazar-Fontana LI Barr V Samelson LE Bierer BE CD28 engagement promotes actin polymerization through the activation of the small Rho GTPase Cdc42 in human T cells J Immunol 2003 171 2225 2232 12928366
Vicente-Manzanares M Rey M Perez-Martinez M Yanez-Mo M Sancho D Cabrero JR Barreiro O de la Fuente H Itoh K Sanchez-Madrid F The RhoA effector mDia is induced during T cell activation and regulates actin polymerization and cell migration in T lymphocytes J Immunol 2003 171 1023 1034 12847276
Lee JR Ha YJ Kim HJ Cutting edge: induced expression of a RhoA-specific guanine nucleotide exchange factor, p190RhoGEF, following CD40 stimulation and WEHI 231 B cell activation J Immunol 2003 170 19 23 12496377
Woodside DG Wooten DK McIntyre BW Adenosine diphosphate (ADP)-ribosylation of the guanosine triphosphatase (GTPase) rho in resting peripheral blood human T lymphocytes results in pseudopodial extension and the inhibition of T cell activation J Exp Med 1998 188 1211 1221 9763600 10.1084/jem.188.7.1211
Costello PS Walters AE Mee PJ Turner M Reynolds LF Prisco A Sarner N Zamoyska R Tybulewicz VL The Rho-family GTP exchange factor Vav is a critical transducer of T cell receptor signals to the calcium, ERK, and NF-kappaB pathways Proc Natl Acad Sci USA 1999 96 3035 3040 10077632 10.1073/pnas.96.6.3035
Feldmann M Pathogenesis of arthritis: recent research progress Nat Immunol 2001 2 771 773 11526382 10.1038/ni0901-771
Mojcik CF Shevach EM Adhesion molecules: a rheumatologic perspective Arthritis Rheum 1997 40 991 1004 9182908
Weinberg JB Pippen AM Greenberg CS Extravascular fibrin formation and dissolution in synovial tissue of patients with osteoarthritis and rheumatoid arthritis Arthritis Rheum 1991 34 996 1005 1677574
Shin H Nakajima T Kitajima I Shigeta K Abeyama K Imamura T Okano T Kawahara K Nakamura T Maruyama I Thrombin receptor-mediated synovial proliferation in patients with rheumatoid arthritis Clin Immunol Immunopathol 1995 76 225 233 7554443 10.1006/clin.1995.1120
Gabazza EC Osamu T Yamakami T Ibata H Sato T Sato Y Shima T Correlation between clotting and collagen metabolism markers in rheumatoid arthritis Thromb Haemost 1994 71 199 202 7514817
Ohba T Takase Y Ohhara M Kasukawa R Thrombin in the synovial fluid of patients with rheumatoid arthritis mediates proliferation of synovial fibroblast-like cells by induction of platelet derived growth factor J Rheumatol 1996 23 1505 1511 8877916
Shin H Kitajima I Nakajima T Shao Q Tokioka T Takasaki I Hanyu N Kubo T Maruyama I Thrombin receptor mediated signals induce expressions of interleukin 6 and granulocyte colony stimulating factor via NF-kappa B activation in synovial fibroblasts Ann Rheum Dis 1999 58 55 60 10343541
Maruyama N Hirano F Yoshikawa N Migita K Eguchi K Tanaka H Thrombin stimulates cell proliferation in human fibroblast-like synoviocytes in nuclear factor-kappaB activation and protein kinase C mediated pathway J Rheumatol 2000 27 2777 2785 11128663
Malik AB Fenton JWII Thrombin-mediated increase in vascular endothelial permeability Semin Thromb Hemost 1992 18 193 199 1631567
Varisco PA Peclat V van Ness K Bischof-Delaloye A So A Busso N Effect of thrombin inhibition on synovial inflammation in antigen induced arthritis Ann Rheum Dis 2000 59 781 787 11005778 10.1136/ard.59.10.781
Arnett FC Edworthy SM Bloch DA McShane DJ Fries JF Cooper NS Healey LA Kaplan SR Liang MH Luthra HS The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis Arthritis Rheum 1988 31 315 324 3358796
Altman R Alarcon G Appelrouth D Bloch D Borenstein D Brandt K Brown C Cooke TD Daniel W Gray R The American College of Rheumatology criteria for the classification and reporting of osteoarthritis of the hand Arthritis Rheum 1990 33 1601 1610 2242058
Altman R Alarcon G Appelrouth D Bloch D Borenstein D Brandt K Brown C Cooke TD Daniel W Feldman D The American College of Rheumatology criteria for the classification and reporting of osteoarthritis of the hip Arthritis Rheum 1991 34 505 514 2025304
Altman R Asch E Bloch D Bole G Borenstein D Brandt K Christy W Cooke TD Greenwald R Hochberg M Development of criteria for the classification and reporting of osteoarthritis. Classification of osteoarthritis of the knee. Diagnostic and Therapeutic Criteria Committee of the American Rheumatism Association Arthritis Rheum 1986 29 1039 1049 3741515
Maruyama Y Nishida M Sugimoto Y Tanabe S Turner JH Kozasa T Wada T Nagao T Kurose H Galpha(12/13) mediates alpha(1)-adrenergic receptor-induced cardiac hypertrophy Circ Res 2002 91 961 969 12433842 10.1161/01.RES.0000043282.39776.7C
Tanaka Y Albelda SM Horgan KJ van Seventer GA Shimizu Y Newman W Hallam J Newman PJ Buck CA Shaw S CD31 expressed on distinctive T cell subsets is a preferential amplifier of beta 1 integrin-mediated adhesion J Exp Med 1992 176 245 253 1377224 10.1084/jem.176.1.245
Tanaka Y Minami Y Mine S Hirano H Fujimoto H Fujii K Saito K Tsukada J van Kooyk Y Figdor CG H-Ras signals to cytoskeletal machinery in induction of integrin-mediated adhesion of T cells J Immunol 1999 163 6209 6216 10570313
Tanaka Y Mine S Figdor CG Wake A Hirano H Tsukada J Aso M Fujii K Saito K van Kooyk Y Constitutive chemokine production results in activation of leukocyte function-associated antigen-1 on adult T-cell leukemia cells Blood 1998 91 3909 3919 9573029
van Nieuw Amerongen GP van Delft S Vermeer MA Collard JG van Hinsbergh VW Activation of RhoA by thrombin in endothelial hyperpermeability: role of Rho kinase and protein tyrosine kinases Circ Res 2000 87 335 340 10948069
Ren XD Kiosses WB Schwartz MA Regulation of the small GTP-binding protein Rho by cell adhesion and the cytoskeleton EMBO J 1999 18 578 585 9927417 10.1093/emboj/18.3.578
Wells C Jiang X Gutowski S Sternweis PC Functional characterization of p115 RhoGEF Methods Enzymol 2002 345 371 382 11665621
Yuan J Slice LW Rozengurt E Activation of protein kinase D by signaling through Rho and the alpha subunit of the heterotrimeric G protein G13 J Biol Chem 2001 276 38619 38627 11507098 10.1074/jbc.M105530200
Kozasa T Jiang X Hart MJ Sternweis PM Singer WD Gilman AG Bollag G Sternweis PC p115 RhoGEF, a GTPase activating protein for Galpha12 and Galpha13 Science 1998 280 2109 2111 9641915 10.1126/science.280.5372.2109
Kozasa T Regulation of G protein-mediated signal transduction by RGS proteins Life Sci 2001 68 2309 2317 11358341 10.1016/S0024-3205(01)01020-7
Gilchrist A Vanhauwe JF Li A Thomas TO Voyno-Yasenetskaya T Hamm HE G alpha minigenes expressing C-terminal peptides serve as specific inhibitors of thrombin-mediated endothelial activation J Biol Chem 2001 276 25672 25679 11274183 10.1074/jbc.M100914200
Morris R Winyard PG Brass LF Blake DR Morris CJ Thrombin receptor expression in rheumatoid and osteoarthritic synovial tissue Ann Rheum Dis 1996 55 841 843 8976643
Marra F DeFranco R Grappone C Milani S Pinzani M Pellegrini G Laffi G Gentilini P Expression of the thrombin receptor in human liver: up-regulation during acute and chronic injury Hepatology 1998 27 462 471 9462645 10.1002/hep.510270221
Ceuppens JL Baroja ML Lorre K van Damme J Billiau A Human T cell activation with phytohemagglutinin. The function of IL-6 as an accessory signal J Immunol 1988 141 3868 3874 3263438
Gauldie J Richards C Harnish D Lansdorp P Baumann H Interferon beta 2/B-cell stimulatory factor type 2 shares identity with monocyte-derived hepatocyte-stimulating factor and regulates the major acute phase protein response in liver cells Proc Natl Acad Sci USA 1987 84 7251 7255 2444978
Houssiau FA Devogelaer JP van Damme J de Deuxchaisnes CN van Snick J Interleukin-6 in synovial fluid and serum of patients with rheumatoid arthritis and other inflammatory arthritides Arthritis Rheum 1988 31 784 788 3260102
Fujii K Tanaka Y Hubscher S Saito K Ota T Eto S Crosslinking of CD44 on rheumatoid synovial cells augment interleukin 6 production Lab Invest 1999 79 1439 1446 10616194
Mihara M Moriya Y Kishimoto T Ohsugi Y Interleukin-6 (IL-6) induces the proliferation of synovial fibroblastic cells in the presence of soluble IL-6 receptor Br J Rheumatol 1995 34 321 325 7788145
Kotake S Sato K Kim KJ Takahashi N Udagawa N Nakamura I Yamaguchi A Kishimoto T Suda T Kashiwazaki S Interleukin-6 and soluble interleukin-6 receptors in the synovial fluids from rheumatoid arthritis patients are responsible for osteoclast-like cell formation J Bone Miner Res 1996 11 88 95 8770701
| 15899034 | PMC1174939 | CC BY | 2021-01-04 16:02:37 | no | Arthritis Res Ther. 2005 Feb 18; 7(3):R476-R484 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1694 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar16951589903610.1186/ar1695Research ArticlePeripheral blood but not synovial fluid natural killer T cells are biased towards a Th1-like phenotype in rheumatoid arthritis Linsen Loes [email protected] Marielle [email protected] Kurt [email protected] Veerle [email protected] Piet [email protected] Jef [email protected] Piet [email protected] Biomedisch Onderzoeksinstituut, Limburgs Universitair Centrum and School of Life Sciences, Transnationale Universiteit Limburg, Universitaire Campus, Diepenbeek, Belgium2005 18 2 2005 7 3 R493 R502 13 10 2004 17 11 2004 14 1 2005 19 1 2005 Copyright © 2005 Linsen et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Natural killer T (NKT) cells have been implicated in the regulatory immune mechanisms that control autoimmunity. However, their precise role in the pathogenesis of rheumatoid arthritis (RA) remains unclear. The frequency, cytokine profile and heterogeneity of NKT cells were studied in peripheral blood mononuclear cells (PBMCs) from 23 RA patients and 22 healthy control individuals, including paired PBMC–synovial fluid samples from seven and paired PBMC–synovial tissue samples from four RA patients. Flow cytometry revealed a decreased frequency of NKT cells in PBMCs from RA patients. NKT cells were present in paired synovial fluid and synovial tissue samples. Based on the reactivity of PBMC-derived NKT cells toward α-galactosylceramide, RA patients could be divided into responders (53.8%) and nonresponders (46.2%). However, NKT cells isolated from synovial fluid from both responders and nonresponders expanded upon stimulation with α-galactosylceramide. Analysis of the cytokine profile of CD4+ and CD4- PBMC derived NKT cell lines from RA patients revealed a significantly reduced number of IL-4 producing cells. In contrast, synovial fluid derived NKT cell lines exhibited a Th0-like phenotype, which was comparable to that in healthy control individuals. This suggests that synovial fluid NKT cells are functional, even in patients with nonresponding NKT cells in their blood. We conclude that, because the number of Vα24+Vβ11+CD3+ NKT cells is decreased and the cytokine profile of blood-derived NKT cells is biased toward a Th1-like phenotype in RA patients, NKT cells might be functionally related to resistance or progression of RA. Providing a local boost to the regulatory potential of NKT cells might represent a useful candidate therapy for RA.
==== Body
Introduction
Natural killer T (NKT) cells are a distinct subset of lymphocytes that share the characteristics of both T cells and natural killer cells. They express a semi-invariant TCR (TCR Vα24Jα18 and Vβ11 in human; Vα14Jα281 and Vβ8, Vβ7 or Vβ2 in mouse) and recognize glycolipid antigens presented by the major histocompatibility complex class I-like molecule CD1d [1]. Two subsets can be distinguished [2,3]: CD4+ NKT cells that produce T-helper (Th)1-type and Th2-type cytokines, and CD4-CD8- (double negative) NKT cells that primarily produce Th1-type cytokines. The ability to secrete cytokines and chemokines rapidly is thought to underlie their regulatory function in a variety of diseases, including cancer and autoimmunity [4]. Although the natural ligand of NKT cells remains to be elucidated, it has been reported that the sponge derived glycolipid α-galactosylceramide (α-GalCer) is a potent activator of mouse and human NKT cells, both in vitro and in vivo [5,6]. When α-GalCer is administered to mice it polarizes the adaptive immune response toward production of Th2 cytokines [7,8], which therefore raises the possibility that α-GalCer can temper or even prevent Th1-mediated autoimmune diseases.
Several studies have shown that NKT cells are decreased or dysfunctional in autoimmune conditions such as insulin-dependent diabetes mellitus, systemic sclerosis, systemic lupus erythematosus, rheumatoid arthritis (RA) and multiple sclerosis [9-12]. Significant therapeutic effects of α-GalCer have been demonstrated in animal models of autoimmunity, such as experimental allergic encephalomyelitis [13-15] and nonobese diabetic mice [16,17].
Because the NKT/CD1d system is phylogenetically conserved among mammals, findings in mice are expected to have a direct parallel in humans. The NKT cell frequency in peripheral blood mononuclear cells (PBMCs) is lower in humans than in mice [1], which may be an obstacle in translating results from animal studies to the clinic. However, results from a phase I study conducted in advanced cancer patients revealed that treating patients with α-GalCer can increase NKT cell numbers above pretreatment levels. This again indicates that α-GalCer could be applied to the treatment of patients with autoimmune disease [18].
RA is an autoimmune disease that is characterized by a chronic inflammation of the joints, followed by progressive destruction of cartilage and underlying bone [19]. Autoreactive Th1 T cells are believed to play a major role in the disease process [20-22]. In RA patients, the frequency of NKT cells is decreased, but the functional characteristics of NKT cells have not yet been fully elucidated. Chiba and coworkers [23] demonstrated that administration of a truncated form of α-GalCer to mice suffering from collagen-induced arthritis – a frequently used animal model of RA – resulted in protection from disease, indicating that this might represent a therapy that can enhance NKT cell numbers in RA patients.
In the present study we analyzed the frequency, functional characteristics and heterogeneity of NKT cells in peripheral blood, synovial fluid and synovial tissue from RA patients. In parallel, we assessed these parameters in α-GalCer-stimulated short-term cell lines of both peripheral blood and synovial fluid NKT cells. We found that NKT cells were decreased and had altered functional properties in peripheral blood, but they were not impaired in synovial fluid from RA patients. Our data indicate that NKT cells may be involved in the disease process of RA and that a strategy to boost the regulatory potential of NKT cells might be useful in the treatment of RA.
Materials and methods
Patients and healthy control individuals
NKT cell characteristics were examined in 23 RA patients (mean age 52.1 ± 2.0 years, 11 males and 12 females, mean disease duration 8.0 ± 1.6 years), who were diagnosed in accordance with the criteria of the American College of Rheumatology [24], and in 22 healthy individuals (mean age 48.6 ± 2.0 years, 10 males and 12 females). When RA patients presented with a swollen knee, paired peripheral blood and synovial fluid samples were obtained. Synovial tissue samples were obtained from four RA patients after total knee/hip arthroplasty. Patients were informed about the purpose of the study and gave written consent. Approval for the study was granted by our ethics committee. Patient characteristics are summarized in Table 1.
Flow cytometric analysis of natural killer T cells
Expression of cell surface markers was analyzed by flow cytometry. Fluorescein isothiocyanate (FITC)-labelled anti-TCR Vα24 and phycoerythrin (PE)-labelled TCR Vβ11 were purchased from Serotec Ltd (Oxford, UK). Anti-CD3-PE, anti-CD3-PerCP, anti-CD4-FITC, anti-CD8-PE, anti-CD25-FITC, anti-IFN-γ-FITC and anti-IL-4-PE were obtained from Becton Dickinson (Erembodegem, Belgium). The frequency of invariant NKT cells was estimated using three-colour anti-Vα24/anti-Vβ11/anti-CD3 staining. For intracellular cytokine detection, α-GalCer expanded Vα24+Vβ11+ NKT cells or Vα24+ isolated NKT cell lines were stimulated with 25 ng/ml phorbol-12-myristate-13-acetate and 1 μg/ml ionomycine in the presence of 10 μg/ml brefeldin A for 4 hours. Intracellular staining was performed as previously described [25]. Cells were analyzed on a FACSCalibur flow cytometer using Cellquest software (Becton Dickinson).
Direct ex vivo analysis of the cytokine profile of natural killer T cells by ELISPOT
ELISPOT procedure was performed as previously described [25]. Briefly, 2 × 105 PBMCs were stimulated with 100 ng/ml α-GalCer in anti-IFN-γ or anti-IL-4 (Mabtech, Nacka, Sweden) coated nitrocellulose bottomed plates (Millipore Corp, Bedford, MA, USA). After 20 hours of culture, biotinylated anti-IFN-γ or anti-IL-4 antibody (Mabtech) was added for 2 hours followed by incubation with streptavidin-alkaline phosphatase (Mabtech) and NBT/BCIP (Nitro Blue Tetrazolium/5-Bromo-4 Chloro-3-Indolyphosphate; Pierce, Rockford, IL, USA) as substrate. The number of cytokine-secreting cells was calculated by subtracting the number of spots in control wells (without antigen) from the number of spots obtained in the presence of α-GalCer.
Expansion and culture of Vα24+Vβ11+ natural killer T cells
PBMCs and synovial fluid mononuclear cells (SFMCs) were isolated using Ficoll-Hypaque (Sigma Diagnostics, St Louis, MO, USA) density gradient centrifugation. PBMCs and SFMCs were cultured in the presence of 100 ng/ml α-GalCer (Kirin Brewery Ltd, Gunma, Japan) at a density of 7.5 × 105 cells/ml RPMI supplemented with 10% heat-inactivated foetal bovine serum, 1 mmol/l sodiumpyruvate and 1% nonessential amino acids (Invitrogen, Merelbeke, Belgium). After 7 days, cells were re-stimulated with irradiated autologous, α-GalCer pulsed PBMCs and supplemented with 2 U/ml recombinant human IL-2 (Roche Diagnostics, Brussels, Belgium). On day 7 after re-stimulation, NKT cells were isolated using Vα24+ magnetic isolation (EasySep; Stemcell Technologies, Meylan, France), in accordance with the manufacturer's instructions. Reactivity of the isolated NKT cells toward α-GalCer was tested in a standard [3H]thymidine incorporation assay. During the last 16 hours of culture, cells were pulsed with 1 μCi [3H]thymidine (Amersham, Buckinghamshire, UK) and subsequently harvested using an automated cell harvester (Pharmacia, Uppsala, Sweden). Incorporated radioactivity was measured using a β-plate liquid scintillation counter (Wallac, Turku, Finland). A NKT cell line was considered to be antigen reactive when the mean counts per minute in the presence of α-GalCer exceeded 1000 and the stimulation index (mean counts with α-GalCer/mean counts without α-GalCer) was greater than 3.
Analysis of clonal heterogeneity by T-cell receptor CDR3 region fragment length analysis
RNA was isolated from snap frozen synovial tissue samples using the Absolutely RNA RT-PCR Miniprep Kit (Stratagene, Amsterdam, The Netherlands). For isolation of total RNA from PBMCs, SFMCs and isolated NKT cells, the High Pure total RNA Isolation kit (Roche Diagnostics, Brussels, Belgium) was used, in accordance with the manufacturer's instructions. RNA was reverse transcribed into cDNA using AMV reverse transcriptase and an oligo-dT primer (Promega, Madison, WI, USA).
CDR3 spectratyping analysis was performed as described previously [26]. Briefly, 2 μl cDNA was used for first-round PCR analysis performed in 1 × PCR buffer, 0.9 U Taq polymerase, 0.02 mmol/l dNTP mix (all from Roche Diagnostics), 1 μmol/l forward primer specific for TCR Vα24 (5'-GAA CGG AAG ATA TAC AGC AAC TC-3') or TCR Vβ11 (5'-TCC ACA GAG AAG GGA GAT CTT TCC TCT GAG-3') region, and 1 μmol/l reverse primer specific for TCR constant α (5'-ATC ATA AAT TCG GGT AGG ATC C-3') or constant β (5'-CTC TTG ACC ATG GCC ATC-3') region. PCR was performed for 40 cycles (95°C for 20 s, 55°C for 20 s, and 72°C for 40 s) on a GeneAmp PCR system 9600 thermal cycler (Perkin Elmer, Zaventem, Belgium). PCR amplicons were used in a second amplification procedure of 25 cycles using the TCR Vα24 or TCR Vβ11 specific primer as forward primer and a FAM labelled TCR constant α (5'-FAM-CTG TTG CTC TTG AAG TCC ATA G-3') or TCR constant β (5'-FAM-GTG GCA AGG CAC ACC AGT GTG GGC C-3') as reverse primer (Eurogentec, Liege, Belgium) under the same PCR conditions as described above.
PCR amplicon lengths were analyzed on the 310 ABI DNA sequencer (Applied Biosystems, Warrington, UK). Fragment sizes of gene products were calculated using an internal Genescan-500 ROX labelled standard and analysis was performed with 672 Genescan Software (both from Applied Biosystems). The heterogeneity of the CDR3 spectratype profiles provides an indication of the clonality of T-cell populations (Fig. 1): monoclonal with one peak, oligoclonal with two to four peaks, and polyclonal with more than four peaks. Identical peak lengths strongly indicate the presence of identical T cell clones in different samples. A 350 base pair fragment was obtained for the invariant TCR.
Sequence analysis of the invariant T-cell receptor
Purified TCR Vα24 PCR amplicons obtained from first round PCR (as described above) were sequenced with a TCR constant α primer (5'-CTG TTG CTC TTG AAG TCC ATA G-3') using the Big DyeTM Terminator Cycle Sequence Ready Reaction Kit II (Applied Biosystems). Sequences were analyzed on a ABI Prism 310 Genetic Analyser (Applied Biosystems).
Statistical analysis
Differences in the percentage of NKT cells between healthy control individuals and RA patients and between peripheral blood and synovial fluid from RA patients were analyzed using the Mann–Whitney U-test. For comparisons between matched peripheral blood and synovial fluid samples, the Wilcoxon matched pairs signed rank test was used. P < 0.05 was considered statistically significant.
Results
Frequency of Vα24+Vβ11+CD3+ natural killer T cells in rheumatoid arthritis
The frequency of Vα24+Vβ11+CD3+ NKT cells in PBMCs from RA patients and healthy control individuals was analyzed by flow cytometry (Fig. 2). Significantly fewer Vα24+Vβ11+CD3+ NKT cells were found in PBMCs from RA patients (0.03 ± 0.01%) than in healthy control individuals (0.11 ± 0.03%; P < 0.01). We simultaneously determined the NKT cell frequency in paired blood–synovial fluid samples from seven RA patients. Although a tendency toward a higher frequency was observed in the synovial fluid (0.08 ± 0.03%) as compared with the concordant PBMC samples (0.05 ± 0.02%), this finding could not be demonstrated for all patients. These data indicate that the NKT cell frequency is decreased in the blood of RA patients but not increased in synovial fluid as compared with blood from these patients.
Cytokine profile of α-galactosylceramide stimulated peripheral blood mononuclear cells
To assess the cytokine profile of NKT cells directly ex vivo, we tested the reactivity of PBMCs to α-GalCer in 10 RA patients and eight healthy control individuals using an ELISPOT technique with IFN-γ and IL-4 readout. Similar to the frequency analysis by flow cytometry, a significantly decreased number of α-GalCer reactive cells was found for IFN-γ as well as for IL-4 in RA patients as compared with healthy control individuals (2.3 ± 0.6 spots versus 24.3 ± 10.1 spots for IFN-γ and 0.2 ± 0.1 spots versus 3.9 ± 1.1 spots for IL-4 per 2 × 105 cells for RA patients and healthy control individuals, respectively; P < 0.05). To determine whether this diminished frequency was also associated with an altered cytokine profile, the IL-4/IFN-γ ratio was calculated as the number of IL-4 producing cells to the number of IFN-γ producing cells (Fig. 3). The IL-4/IFN-γ ratio in RA patients was decreased as compared with that in healthy control individuals (0.07 ± 0.03 in RA patients versus 0.30 ± 0.10 in healthy control individuals; P = 0.06). This was mainly due to a reduced number of IL-4 producing cells, because the frequency of IL-4 producing cells in RA patients as compared with healthy control individuals was relatively more reduced than that of IFN-γ producing cells. These data indicate that NKT cells derived from RA patients are biased toward a Th1-like phenotype.
Analysis of the invariant T-cell receptor in synovial tissue
NKT cells express the invariant Vα24Jα18 TCR-α chain combined with a variable Vβ11 TCR-β chain. To compare the Vα24 expression profile in PBMCs from RA patients and healthy control individuals, PBMCs from five healthy control individuals and paired PBMCs–SFMCs and PBMCs–synovial tissue samples from four RA patients were subjected to TCR CDR3 size analysis using primers for Vα24 and TCR-α constant region. PBMCs from healthy control individuals exhibited a polyclonal peak profile or a Gaussian-like distribution for Vα24, containing a peak at 350 base pairs, which corresponds to the invariant TCR-α chain that is characteristic for NKT cells (not shown). Although PBMCs from RA patients exhibited an oligoclonal or monoclonal distribution, indicating a restricted usage for Vα24 (Table 2), the invariant TCR peak was present in all patients. We determined whether the invariant TCR could also be found in SFMCs and synovial tissue samples. As in PBMCs, the TCR Vα24 usage in SFMCs and synovial tissue tissue samples was skewed for some patients but polyclonal for others. Again, the invariant TCR peak was detected in SFMCs and synovial tissue samples for all RA patients. Sequence analysis of the PCR products obtained from the CDR3 fragment length analysis confirmed that the peak size of the synovial tissue samples corresponded with the invariant TCR sequence (not shown). These data show that NKT cells are present in rheumatoid synovial fluid as well as in synovial tissue.
Natural killer T-cell reactivity to α-galactosylceramide in rheumatoid arthritis patients
To assess whether the reduced NKT cell frequency in peripheral blood from RA patients was due to an inadequate response to the glycolipid antigen, we stimulated PBMCs from nine healthy control individuals and 13 RA patients and SFMCs from five RA patients with α-GalCer. At day 7, cells were re-stimulated with autologous α-GalCer pulsed, irradiated PBMCs. The NKT cell frequency was determined by flow cytometry at day 14 (Fig. 4). NKT cells from healthy control individuals expanded in response to α-GalCer to 15.8 ± 2.7%, whereas the number of peripheral blood and synovial fluid NKT cells from RA patients was significantly lower after α-GalCer stimulation (8.4 ± 2.9% and 4.4 ± 1.6%, respectively; P < 0.01). A more detailed analysis revealed that this decrease was due to the existence of two subpopulations of RA patients based on the NKT cell numbers reached after 14 days of α-GalCer stimulation. As shown in Fig. 5, NKT cells from six out 13 RA patients did not respond to α-GalCer stimulation (mean frequency after 14 days: 1.0 ± 0.2%, P < 0.01; nonresponders), whereas NKT cells from the remaining seven patients reached frequencies comparable with those in healthy control individuals (14.7 ± 4.0%; responders). Moreover, NKT cells of responder patients appeared to have increased ability to respond to α-GalCer because the expansion was greater than that in healthy control individuals (294-fold versus 149-fold, respectively). No relation between disease parameters (disease duration, disease status) or treatment and responsiveness/nonresponsiveness of NKT cells could be demonstrated. Remarkably, synovial fluid NKT cells, even from nonresponding RA patients, did expand after α-GalCer stimulation (4.94 ± 1.90%). These findings indicate that the reactivity of peripheral blood NKT cells to α-GalCer is impaired in some RA patients, whereas it is intact and even increased in others.
Cytokine profile of peripheral blood and synovial fluid natural killer T cell lines
Next, we analyzed the cytokine profile of peripheral blood derived NKT cells from five healthy control individual and five RA patients, and synovial fluid derived NKT cells from five RA patients by intracellular staining of 14-day-old, α-GalCer stimulated cultures gated on Vα24+ cells. Figure 6 shows that the Vα24+ NKT cell fraction of healthy control individuals contained 64.5 ± 13.1% IFN-γ producing cells, 15.7 ± 6.9% IL-4 producing cells, and 19.7 ± 6.4% cells producing both IFN-γ and IL-4. In contrast, peripheral blood NKT cells from RA patients consisted of significantly more IFN-γ producing cells and significantly fewer cells producing both IFN-γ and IL-4 (92.5 ± 2.7% and 6.1 ± 2.3%, respectively; P < 0.05). Remarkably, synovial fluid derived NKT cells exhibited a cytokine profile similar to that of healthy control individuals, although the number of IL-4 producing cells tended to be lower and the number of cells producing both IFN-γ and IL-4 was somewhat higher (5.3 ± 5.3% and 28.7 ± 6.7%, respectively; P > 0.05). No differences were found between the cytokine profiles of NKT cells of α-GalCer responding and nonresponding patients. Furthermore, no relation with treatment or any disease parameter was found. These observations show that, although NKT cells in PBMCs from RA patients are biased toward a Th1-like cytokine profile, NKT cells in the synovial fluid exhibit a Th0-like cytokine profile that is comparable with that in healthy control individuals.
Cytokine profile of CD4+ and CD4- natural killer T cell subsets in patients with rheumatoid arthritis and healthy control individuals
The observed Th1-like bias in NKT cells from RA patients might be due to an increased number of double-negative NKT cells or a decreased number of CD4+ NKT cells. To analyze the frequency of these NKT cell subtypes, we isolated the Vα24+ cells of α-GalCer stimulated, 14-day-old cultures derived from PBMCs from nine healthy control individuals and seven RA patients by immunomagnetic selection. Positively selected cells were tested for α-GalCer reactivity to ensure the NKT cell nature of the cells. The presence of CD4 was assessed by flow cytometry. NKT cells of healthy control individuals consisted of 33.3 ± 6.7% CD4+ NKT cells and 66.7 ± 6.7% CD4- (double-negative) NKT cells. The frequency of CD4+ and CD4- NKT cells in RA patients did not differ significantly from that in healthy control individuals (49.8 ± 6.3% and 50.2 ± 6.3%, respectively; data not shown).
Figure 7 shows the cytokine profile of each NKT cell subset, as determined by intracellular staining. Peripheral blood derived CD4- NKT cells from healthy control individuals predominantly consisted of IFN-γ producing cells (IFN-γ+ 57.6 ± 8.8%; IL-4+ 19.4 ± 6.6%; IFN-γ+IL-4+ 23.0 ± 6.0%), whereas CD4+ NKT cells contained almost as many IL-4 producing cells as IFN-γ producing cells (IFN-γ+ 40.1 ± 7.4%; IL-4+ 25.1 ± 7.5%; IFN-γ+IL-4+ 34.8 ± 6.4%). However, the CD4+ as well as the CD4- NKT cell fractions in RA patients contained significantly fewer IL-4 producing cells as compared with their counterparts in healthy control individuals (for CD4+ NKT cells: IFN-γ+ 57.2 ± 12.9%; IL-4+ 5.8 ± 1.5%; IFN-γ+IL-4+ 37.0 ± 13.2%; and for CD4- NKT cells: IFN-γ+ 72.1 ± 12.4%; IL-4+ 3.3 ± 1.9%; IFN-γ+IL-4+ 24.6 ± 11.9%), indicating that both CD4+ and CD4- NKT cells in the peripheral blood of RA patients are biased toward a Th1-like cytokine profile.
To exclude the possibility that the observations in NKT cell lines of RA patients were caused by the clonal expansion of one or a few NKT cells, we analyzed the heterogeneity of the Vα24 and Vβ11 TCR by means of CDR3 fragment length analysis. We found that the NKT cell lines of both RA patients and healthy control individuals exhibited a monoclonal Vα24 and polyclonal Vβ11 profile (data not shown), which shows that the differences between NKT cells from RA patients and healthy control individuals found in response to α-GalCer are not due to a skewed outgrowth of only one or a few NKT cells.
Discussion
Several studies have provided evidence that NKT cells are involved in autoimmune conditions [27]. Attempts to increase the number of NKT cells in animal models of autoimmunity by transgenic expression of the invariant TCR or by passive transfer of NKT cells resulted in a protective effect against disease induction [28,29]. Additionally, administration of α-GalCer resulted in prevention or suppression of disease. These studies indicate that NKT cells can play a role in the regulation of autoimmunity and that they are therefore an interesting subject for further investigation in human autoimmune diseases.
In the present study we demonstrated a decreased frequency of NKT cells in PBMCs from RA patients. Because we used anti-Vα24 and anti-Vβ11 monoclonal antibodies to identify invariant NKT cells, it is possible that conventional T cells were also stained by this combination. However, Araki and coworkers [12] showed that the frequency of Vα24+Vβ11+CD3+ T cells, even at low numbers, corresponded well with the NKT cell frequency determined by CD1d tetramers, which supports the specificity of anti-Vα24 and anti-Vβ11 staining for NKT cells.
Several mechanisms may account for NKT cell reduction in the peripheral blood of RA patients. First, NKT cells might preferentially migrate into the joint to fulfill their regulatory function. We therefore studied the frequency of NKT cells in synovial fluid and synovial tissue of RA patients. We found that the NKT cell frequency is not elevated in synovial fluid, but that the invariant TCR can be detected in both synovial tissue and synovial fluid samples from RA patients. Preferential migration of NKT cells into the synovium may have resulted in a monoclonal or oligoclonal Vα24 profile in synovial samples. However, we did not find such a profile in the synovial fluid or synovial tissue of all patients, indicating that the decrease cannot be accounted for by a selective migration of NKT cells toward the joint. A similar conclusion was reached by others for RA [30] and multiple sclerosis [31].
A second possibility might be that the reduced NKT cell frequency is caused by a selective loss of a limited number of NKT cell clones. It was shown in mice that NKT cells exhibit a highly diverse TCR-β repertoire and a small clone size [32], and hence a loss of NKT cells should result in a reduced diversity of TCR Vβ11. However, the Vβ11 profile of α-GalCer expanded peripheral blood NKT cells from RA patients was polyclonal, which suggests that RA patients do not suffer from a specific loss of NKT cells.
A third possible cause is a decreased reactivity toward the natural NKT cell ligand. To examine this possibility, we stimulated PBMCs of RA patients with α-GalCer and found that, in 53.8% of the patients ('responders'), NKT cells expanded upon α-GalCer stimulation and reached levels comparable to those in healthy control individuals. This suggests that an inadequate expression of CD1d [33] or an aberrant presentation of the natural NKT cell antigen, but not decreased reactivity, might account for the NKT cell reduction in these responder patients. In contrast, in 46.2% of the patients ('nonresponders') NKT cells did not react to α-GalCer. This impaired NKT cell function was also reported previously by Kojo and coworkers [11], who proposed that this decreased reactivity might result from an inherent NKT cell defect or a dysfunctional antigen presentation. However, those authors could exclude the possibility that antigen-presenting cells were dysfunctional in nonresponder patients. Remarkably, synovial fluid NKT cells of both responders and nonresponders expanded upon stimulation, indicating that the impaired NKT cell function in nonresponders is restricted to the blood compartment.
Additional mechanisms may account for the reduced frequency, including a decreased thymic output, as was described previously for conventional T cells in RA [34], and a chronic over-stimulation of NKT cells resulting in a decreased frequency due to TCR downregulation after activation [35]. Moreover, it is possible that a chronic activation might also lead to nonresponsiveness because it was shown that NKT cells in α-GalCer injected mice are anergic for an extended period of time [36].
When we analyzed the cytokine profiles of in vitro expanded NKT cells, we found that CD4- NKT cells from healthy control individuals mainly consisted of IFN-γ producing cells, whereas CD4+ NKT cells can produce both Th1-like and Th2-like cytokines. This reflects the direct ex vivo situation reported by others [2,3]. We observed that peripheral blood derived NKT cells from RA patients exhibited a Th1-like phenotype, which was due to a decreased number of IL-4 producing cells in both the CD4+ and CD4- NKT cell subsets compared with healthy control individuals. Although these data were obtained from in vitro cultured cells, our data obtained from direct ex vivo stimulation of PBMCs with α-GalCer confirm a Th1-like bias of NKT cells in RA patients. Strikingly, NKT cells in the synovial fluid do not show this Th1-like bias, but have a Th0-like profile that is similar to that of peripheral blood NKT cells from healthy control individuals. A Th1-like bias of peripheral blood derived NKT cells was also found in diabetes [9] and multiple sclerosis [12], indicating that NKT cell dysfunction is not specific for RA but might play a major role in the aetiology of autoimmune diseases.
Although no relation between reactivity to α-GalCer or NKT cell cytokine profiles and drug treatment was found, a possible effect of the medication cannot be excluded.
In summary, the presence, even in nonresponder patients, of functional NKT cells that exhibit a Th0-like cytokine profile in the synovial fluid may indicate that unimpaired NKT cells migrate from the peripheral blood toward the synovium in order to exert their regulatory function. NKT cells express a chemokine receptor profile similar to Th1-type inflammatory homing cells, which suggests that these cells perform their function mainly in the tissue [37]. However, their number and/or function are probably insufficient to resolve the ongoing autoimmune reaction. Hence, a strategy to enhance locally the number of NKT cells by α-GalCer represents a potential treatment for RA.
Conclusion
Because the number of Vα24+Vβ11+CD3+ NKT cells is decreased and the cytokine profile of blood derived NKT cells is biased toward a Th1-like phenotype in RA patients, NKT cells might be functionally related to resistance or progression of RA and are therefore an interesting target for the treatment of RA.
Abbreviations
α-GalCer = α-galactosylceramide; FITC = fluorescein isothiocyanate; IFN = interferon; IL = interleukin; NKT = natural killer T (cell); PBMC = peripheral blood mononuclear cell; PE = phycoerythrin; PCR = polymerase chain reaction; RA = rheumatoid arthritis; SFMC = synovial fluid mononuclear cell; TCR = T-cell receptor; Th = T-helper (cell).
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
LL carried out all experiments and drafted the manuscript. MT participated in frequency analysis of NKT cells. KB participated in reactivity assays. PG provided clinical material. VS and JR critically revised the manuscript. PS coordinated the study. All authors read and approved the final manuscript.
Acknowledgements
The authors wish to thank Kirin Brewery Ltd for kindly providing α-GalCer, Dr J Vanhoof and H Leroi for collecting patient material, and J Bleus for expert technical help. This study was supported by a grant of the 'Bijzonder onderzoeksfonds, LUC'.
Figures and Tables
Figure 1 Clonality of T-cell populations. (a) Monoclonal: one peak. (b) Oligoclonal: two to four peaks. (c) Polyclonal: more than four peaks.
Figure 2 Frequency of natural killer T (NKT) cells in rheumatoid arthritis (RA) patients and healthy control individuals. NKT cell frequency in freshly isolated peripheral blood (PB) mononuclear cells from 22 healthy control individuals and 23 RA patients, and in synovial fluid (SF) mononuclear cells from seven RA patients was determined by flow cytometry. Cells were stained with anti-Vα24, anti-Vβ11 and anti-CD3 monoclonal antibody. Error bars indicate the standard error of the mean. *P < 0.01.
Figure 3 IL-4/IFN-γ ratio in α-galactosylceramide (α-GalCer) stimulated peripheral blood mononuclear cells (PBMCs) evaluated by ELISPOT. PBMCs (2 × 105 cells/well) from 10 rheumatoid arthritis patients and eight healthy control individuals were stimulated with α-GalCer or no antigen for 20 hours. The number of cytokine secreting cells was calculated by subtracting the number of spots in control wells (without antigen) from the number of spots obtained in the presence of each stimulating agent. The IL-4/IFN-γ ratio is the number of IL-4 producing cells divided by the number of IFN-γ producing cells. Error bars indicate standard error of the mean.
Figure 4 Reactivity of peripheral blood (PB) and synovial fluid (SF) derived natural killer T (NKT) cells to α-galactosylceramide (α-GalCer). PB mononuclear cells (1.5 × 106 cells/well) of nine healthy control individuals and 13 rheumatoid arthritis (RA) patients as well as SF mononuclear cells of five RA patients were stimulated with α-GalCer and re-stimulated on day 7 with autologous, α-GalCer pulsed, irradiated PB mononuclear cells in the presence of 2 U/ml IL-2. NKT cell numbers were determined by flow cytometry at day 14. Error bars indicate standard error of the mean. *P < 0.01.
Figure 5 Rheumatoid arthritis (RA) patients can be divided into responder and nonresponder patients, based on peripheral blood derived natural killer T (NKT) cell reactivity to α-galactosylceramide (α-GalCer). Peripheral blood (PB) mononuclear cells (1.5 × 106 cells/well) from nine healthy control individuals and 13 RA patients, as well as synovial fluid (SF) mononuclear cells from five RA patients, were stimulated with α-GalCer and re-stimulated on day 7 with autologous, α-GalCer pulsed, irradiated PB mononuclear cells in the presence of 2 U/ml IL-2. NKT cell numbers were determined by flow cytometry on day 14. Patients were considered nonresponders when the frequency of Vα24+Vβ11+CD3+ NKT cells derived from PB mononuclear cells was lower than 2% after 14 days of culture. Error bars indicate standard error of the mean. *P < 0.01.
Figure 6 Cytokine profile of α-galactosylceramide (α-GalCer) expanded natural killer T (NKT) cells. Peripheral blood (PB) mononuclear cells (1.5 × 106 cells/well) from five healthy control individuals and five RA patients as well as synovial fluid (SF) mononuclear cells from five RA patients were stimulated with α-GalCer and re-stimulated on day 7 with autologous, α-GalCer pulsed, irradiated PB mononuclear cells in the presence of 2 U/ml IL-2. The cytokine profile was analyzed by intracellular staining and gating on the Vα24+ subset. Error bars indicate standard error of the mean. *P < 0.05.
Figure 7 Cytokine profile of CD4+ and CD4- natural killer T (NKT) cell lines derived from peripheral blood mononuclear cells from rheumatoid arthritis (RA) patients and healthy control individuals. Vα24+ cells of α-galactosylceramide (α-GalCer) stimulated, 14-day-old cultures from nine healthy control individuals and nine RA patients were isolated using biomagnetic selection. The cytokine profile of (a) CD4+ and (b) CD4- NKT cells was assessed by intracellular staining. Error bars indicate standard error of the mean. *P < 0.05.
Table 1 Patient characteristics
Patient Age (years)/sex Disease duration (years) Treatment
1 54/M 5 Azathioprine, methylprednisolone
2 38/F 6 Hydroxychloroquine, salazopyrine
3 64/F 7 NSAID
4 43/F 5 NSAID
5 46/M 1 Salazopyrine
6 46/M <1 Salazopyrine
7 52/M 11 NSAID
8 53/F 11 NSAID
9 49/M 10 NSAID
10 52/F 4 NSAID
11 69/F 36 Salazopyrine
12 65/F <1 Untreated
13 35/M <1 Untreated
14a 57/F 4 Methotrexate
15a 46/M 2 Anti-TNF, salazopyrine
16a 41/M 10 Salazopyrine
17a 41/M 13 Methotrexate
18a 60/M 2 Leflunomide
19a 43/F 5 Methotrexate
20b 63/M 17 Salazopyrine
21a,b 65/F 4 Salazopyrine, hydroxychloroquine
22b 62/F 12 Leflunomide
23b 54/F 17 Methylprednisolone
aSynovial fluid sample. bSynovial tissue sample. F, female; M, male; NSAID, nonsteroidal anti-inflammatory drug; TNF, tumour necrosis factor.
Table 2 T cell receptor Vα24 usage in peripheral blood mononuclear cells, synovial fluid mononuclear cells and synovial tissue from rheumatoid arthritis patients
Vα24 Vβ11
PBMCs SFMCs ST PBMCs SFMCs ST
RA 1 mono mono NA oligo (2) mono NA
RA 2 mono poly oligo (2) mono poly oligo (2)
RA 3 mono oligo (2) NA Poly oligo (2) NA
RA 4 poly poly NA Poly poly NA
RA 5 oligo (3) NA mono mono NA mono
RA 6 poly NA poly poly NA poly
RA 7 oligo (3) NA mono poly NA oligo (3)
The clonality of the T-cell receptor (TCR) Vα24 family was assessed by CDR3 spectratyping of peripheral blood mononuclear cells (PBMCs), synovial fluid mononuclear cells (SFMCs) and synovial tissue (ST) from rheumatoid arthritis (RA) patients. (See Fig. 1 for representative monoclonal [panel a], oligoclonal [panel b] and polyclonal [panel c] profiles.) mono, monoclonal profile; NA, not available; oligo, oligoclonal profile; poly, polyclonal profile.
==== Refs
Godfrey DI Hammond KJ Poulton LD Smyth MJ Baxter AG NKT cells: facts, functions and fallacies Immunol Today 2000 21 573 583 11094262 10.1016/S0167-5699(00)01735-7
Lee PT Benlagha K Teyton L Bendelac A Distinct functional lineages of human V(alpha)24 natural killer T cells J Exp Med 2002 195 637 641 11877486 10.1084/jem.20011908
Gumperz JE Miyake S Yamamura T Brenner MB Functionally distinct subsets of CD1d-restricted natural killer T cells revealed by CD1d tetramer staining J Exp Med 2002 195 625 636 11877485 10.1084/jem.20011786
Wilson MT Singh AK Van Kaer L Immunotherapy with ligands of natural killer T cells Trends Mol Med 2002 8 225 231 12067632 10.1016/S1471-4914(02)02325-0
Brossay L Chioda M Burdin N Koezuka Y Casorati G Dellabona P Kronenberg M CD1d-mediated recognition of an alpha-galactosylceramide by natural killer T cells is highly conserved through mammalian evolution J Exp Med 1998 188 1521 1528 9782129 10.1084/jem.188.8.1521
Nieda M Nicol A Koezuka Y Kikuchi A Takahashi T Nakamura H Furukawa H Yabe T Ishikawa Y Tadokoro K Activation of human Valpha24NKT cells by alpha-glycosylceramide in a CD1d-restricted and Valpha24TCR-mediated manner Hum Immunol 1999 60 10 19 9952023 10.1016/S0198-8859(98)00100-1
Singh N Hong S Scherer DC Serizawa I Burdin N Kronenberg M Koezuka Y Van Kaer L Cutting edge: activation of NK T cells by CD1d and alpha-galactosylceramide directs conventional T cells to the acquisition of a Th2 phenotype J Immunol 1999 163 2373 2377 10452969
Burdin N Brossay L Kronenberg M Immunization with alpha-galactosylceramide polarizes CD1-reactive NK T cells towards Th2 cytokine synthesis Eur J Immunol 1999 29 2014 2025 10382765 10.1002/(SICI)1521-4141(199906)29:06<2014::AID-IMMU2014>3.0.CO;2-G
Wilson SB Kent SC Patton KT Orban T Jackson RA Exley M Porcelli S Schatz DA Atkinson MA Balk SP Extreme Th1 bias of invariant Valpha24JalphaQ T cells in type 1 diabetes Nature 1998 391 177 181 9428763 10.1038/34419
van der Vliet HJ von Blomberg BM Nishi N Reijm M Voskuyl AE van Bodegraven AA Polman CH Rustemeyer T Lips P van den Eertwegh AJ Circulating V(alpha24+) Vbeta11+ NKT cell numbers are decreased in a wide variety of diseases that are characterized by autoreactive tissue damage Clin Immunol 2001 100 144 148 11465942 10.1006/clim.2001.5060
Kojo S Adachi Y Keino H Taniguchi M Sumida T Dysfunction of T cell receptor AV24AJ18+, BV11+ double-negative regulatory natural killer T cells in autoimmune diseases Arthritis Rheum 2001 44 1127 1138 11352245 10.1002/1529-0131(200105)44:5<1127::AID-ANR194>3.0.CO;2-W
Araki M Kondo T Gumperz JE Brenner MB Miyake S Yamamura T T(h)2 bias of CD4(+) NKT cells derived from multiple sclerosis in remission Int Immunol 2003 15 279 288 12578858 10.1093/intimm/dxg029
Singh AK Wilson MT Hong S Olivares-Villagomez D Du C Stanic AK Joyce S Sriram S Koezuka Y Van Kaer L Natural killer T cell activation protects mice against experimental autoimmune encephalomyelitis J Exp Med 2001 194 1801 1811 11748281 10.1084/jem.194.12.1801
Jahng AW Maricic I Pedersen B Burdin N Naidenko O Kronenberg M Koezuka Y Kumar V Activation of natural killer T cells potentiates or prevents experimental autoimmune encephalomyelitis J Exp Med 2001 194 1789 1799 11748280 10.1084/jem.194.12.1789
Furlan R Bergami A Cantarella D Brambilla E Taniguchi M Dellabona P Casorati G Martino G Activation of invariant NKT cells by alphaGalCer administration protects mice from MOG35-55-induced EAE: critical roles for administration route and IFN-gamma Eur J Immunol 2003 33 1830 1838 12811843 10.1002/eji.200323885
Hong S Wilson MT Serizawa I Wu L Singh N Naidenko OV Miura T Haba T Scherer DC Wei J The natural killer T-cell ligand alpha-galactosylceramide prevents autoimmune diabetes in non-obese diabetic mice Nat Med 2001 7 1052 1056 11533710 10.1038/nm0901-1052
Sharif S Arreaza GA Zucker P Mi QS Sondhi J Naidenko OV Kronenberg M Koezuka Y Delovitch TL Gombert JM Activation of natural killer T cells by alpha-galactosylceramide treatment prevents the onset and recurrence of autoimmune Type 1 diabetes Nat Med 2001 7 1057 1062 11533711 10.1038/nm0901-1057
Okai M Nieda M Tazbirkova A Horley D Kikuchi A Durrant S Takahashi T Boyd A Abraham R Yagita H Human peripheral blood Valpha24+ Vbeta11+ NKT cells expand following administration of alpha-galactosylceramide-pulsed dendritic cells Vox Sang 2002 83 250 253 12366768 10.1046/j.1423-0410.2002.00217.x
VanderBorght A Geusens P Raus J Stinissen P The autoimmune pathogenesis of rheumatoid arthritis: role of autoreactive T cells and new immunotherapies Semin Arthritis Rheum 2001 31 160 175 11740797 10.1053/sarh.2001.27736
Moots RJ A fistful of T cells Br J Rheumatol 1998 37 602 611 9667613 10.1093/rheumatology/37.6.602
Weyand CM Goronzy JJ T-cell responses in rheumatoid arthritis: systemic abnormalities-local disease Curr Opin Rheumatol 1999 11 210 217 10328581 10.1097/00002281-199905000-00010
Kusaba M Honda J Fukuda T Oizumi K Analysis of type 1 and type 2 T cells in synovial fluid and peripheral blood of patients with rheumatoid arthritis J Rheumatol 1998 25 1466 1471 9712085
Chiba A Oki S Miyamoto K Hashimoto H Yamamura T Miyake S Suppression of collagen-induced arthritis by natural killer T cell activation with OCH, a sphingosine-truncated analog of alpha-galactosylceramide Arthritis Rheum 2004 50 305 313 14730629 10.1002/art.11489
Arnett FC Edworthy SM Bloch DA McShane DJ Fries JF Cooper NS Healey LA Kaplan SR Liang MH Luthra HS The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis Arthritis Rheum 1988 31 315 324 3358796
van der AA Hellings N Bernard CC Raus J Stinissen P Functional properties of myelin oligodendrocyte glycoprotein-reactive T cells in multiple sclerosis patients and controls J Neuroimmunol 2003 137 164 176 12667661 10.1016/S0165-5728(03)00048-1
VanderBorght A van der AA Geusens P Vandevyver C Raus J Stinissen P Identification of overrepresented T cell receptor genes in blood and tissue biopsies by PCR-ELISA J Immunol Methods 1999 223 47 61 10037234 10.1016/S0022-1759(98)00201-4
Van Kaer L Natural killer T cells as targets for immunotherapy of autoimmune diseases Immunol Cell Biol 2004 82 315 322 15186263 10.1111/j.0818-9641.2004.01252.x
Hammond KJ Godfrey DI NKT cells: potential targets for autoimmune disease therapy? Tissue Antigens 2002 59 353 363 12144618 10.1034/j.1399-0039.2002.590501.x
Wilson MT Van Kaer L Natural killer T cells as targets for therapeutic intervention in autoimmune diseases Curr Pharm Des 2003 9 210 220 10.2174/1381612033392080
Maeda T Keino H Asahara H Taniguchi M Nishioka K Sumida T Decreased TCR AV24AJ18+ double-negative T cells in rheumatoid synovium Rheumatology (Oxford) 1999 38 186 188 10342637
Illes Z Kondo T Newcombe J Oka N Tabira T Yamamura T Differential expression of NK T cell V alpha 24J alpha Q invariant TCR chain in the lesions of multiple sclerosis and chronic inflammatory demyelinating polyneuropathy J Immunol 2000 164 4375 4381 10754338
Matsuda JL Gapin L Fazilleau N Warren K Naidenko OV Kronenberg M Natural killer T cells reactive to a single glycolipid exhibit a highly diverse T cell receptor beta repertoire and small clone size Proc Natl Acad Sci USA 2001 98 12636 12641 11592984 10.1073/pnas.221445298
Kojo S Tsutsumi A Goto D Sumida T Low expression levels of soluble CD1d gene in patients with rheumatoid arthritis J Rheumatol 2003 30 2524 2528 14719188
Koetz K Bryl E Spickschen K O'Fallon WM Goronzy JJ Weyand CM T cell homeostasis in patients with rheumatoid arthritis Proc Natl Acad Sci USA 2000 97 9203 9208 10922071 10.1073/pnas.97.16.9203
Wilson MT Johansson C Olivares-Villagomez D Singh AK Stanic AK Wang CR Joyce S Wick MJ Van Kaer L The response of natural killer T cells to glycolipid antigens is characterized by surface receptor down-modulation and expansion Proc Natl Acad Sci USA 2003 100 10913 10918 12960397 10.1073/pnas.1833166100
Fujii S Shimizu K Kronenberg M Steinman RM Prolonged IFN-gamma-producing NKT response induced with alpha-galactosylceramide-loaded DCs Nat Immunol 2002 3 867 874 12154358 10.1038/ni827
Thomas SY Hou R Boyson JE Means TK Hess C Olson DP Strominger JL Brenner MB Gumperz JE Wilson SB CD1d-restricted NKT cells express a chemokine receptor profile indicative of Th1-type inflammatory homing cells J Immunol 2003 171 2571 2580 12928408
| 15899036 | PMC1174940 | CC BY | 2021-01-04 16:02:36 | no | Arthritis Res Ther. 2005 Feb 18; 7(3):R493-R502 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1695 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar16971589903210.1186/ar1697Research ArticleCitrullinated proteins have increased immunogenicity and arthritogenicity and their presence in arthritic joints correlates with disease severity Lundberg Karin [email protected] Suzanne [email protected] Erik R [email protected] Karin [email protected] Venrooij Walter J [email protected] Lars [email protected] AJW [email protected] Helena Erlandsson [email protected] Rheumatology Unit, Department of Medicine, Karolinska Institutet, Stockholm, Sweden2 Department of Biochemistry, Radboud University Nijmegen, Nijmegen, The Netherlands2005 21 2 2005 7 3 R458 R467 21 10 2004 2 12 2004 16 12 2004 20 1 2005 Copyright © 2005 Lundberg et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Autoantibodies directed against citrulline-containing proteins have an impressive specificity of nearly 100% in patients with rheumatoid arthritis and have been suggested to be involved in the disease pathogenesis. The targeted epitopes are generated by a post-translational modification catalysed by the calcium-dependent enzyme peptidyl arginine deiminase (PAD), which converts positively charged arginine to polar but uncharged citrulline. The aim of this study was to explore the effects of citrullination on the immunogenicity of autoantigens as well as on potential arthritogenicity. Thus, immune responses to citrullinated rat serum albumin (Cit-RSA) and to unmodified rat serum albumin (RSA) were examined as well as arthritis development induced by immunisation with citrullinated rat collagen type II (Cit-CII) or unmodified CII. In addition, to correlate the presence of citrullinated proteins and the enzyme PAD4 with different stages of arthritis, synovial tissues obtained at different time points from rats with collagen-induced arthritis were examined immunohistochemically. Our results demonstrate that citrullination of the endogenous antigen RSA broke immunological tolerance, as was evident by the generation of antibodies directed against the modified protein and cross-reacting with the native protein. Furthermore we could demonstrate that Cit-CII induced arthritis with higher incidence and earlier onset than did the native counterpart. Finally, this study reveals that clinical signs of arthritis precede the presence of citrullinated proteins and the enzyme PAD4. As disease progressed into a more severe and chronic state, products of citrullination appeared specifically in the joints. Citrullinated proteins were detected mainly in extracellular deposits but could also be found in infiltrating cells and on the cartilage surface. PAD4 was detected in the cytoplasm of infiltrating mononuclear cells, from day 21 after immunisation and onwards. In conclusion, our data reveal the potency of citrullination to break tolerance against the self antigen RSA and to increase the arthritogenic properties of the cartilage antigen CII. We also show that citrullinated proteins and the enzyme PAD4 are not detectable in healthy joints, and that the appearance and amounts in arthritic joints of experimental animals are correlated with the severity of inflammation.
==== Body
Introduction
The chronic inflammatory joint disease rheumatoid arthritis (RA) is characterised by synovial inflammation and pannus formation, which can lead to severe destruction of cartilage and bone. Several self proteins have been suggested as disease-driving autoantigens, and the presence of autoantibodies with different specificities in patients with RA (reviewed in [1,2]) supports the hypothesis of an autoimmune aetiology. Rheumatoid factor has for a long time been the best-described RA-associated antibody marker, recognising the Fc part of IgG molecules. However, another class of autoantibodies has lately gained attention, namely antibodies directed against proteins containing the non-standard amino acid citrulline [3,4].
Citrulline is generated by the deimination of arginine, a post-translational modification occurring during apoptosis as well as during the terminal differentiation of cells, in both healthy and arthritic individuals [5,6]. Citrullination is catalysed by a family of calcium-dependent enzymes named peptidyl arginine deiminase (PAD, EC 3.5.3.15) (reviewed in [7]). These enzymes are present in several different cell and tissue types, including inflammatory cells (PAD2 [8-10] and PAD4 [10-12]). PAD4 has been detected in granulocytes infiltrating the synovial tissue in a mouse model of arthritis [13] and this enzyme, together with PAD2, has also been demonstrated in macrophages from synovial fluid of patients with RA [10].
The best-described citrulline-reactive autoantibodies associated with RA are the following: anti-perinuclear factor [14,15] and anti-keratin autoantibodies [16,17], both directed against citrullinated filaggrin [18]; anti-Sa autoantibodies [19] directed against citrullinated vimentin [20]; and antibodies against cyclic citrullinated peptide (anti-CCP) [21,22]. These latter autoantibodies have a sensitivity of up to 80% and a specificity of 98% in patients with RA [1,22]. Besides this high specificity, these markers are present early in disease, even before clinical onset [23,24], and they are synthesised locally by plasma cells in the pannus [25,26]. In addition, the existence of citrulline-reactive antibodies has been associated with a more active and severe disease [27-34] and a strong association with major histocompatibility complex (MHC) shared epitope haplotypes [28,35,36] has also been reported.
The accumulated data point towards a link between citrullinated proteins and the pathogenesis of RA. We therefore considered it to be of interest to explore the effects of citrullination on the immunogenicity of autoantigens and on potential arthritogenicity. In the present study we examined the responses of rat T and B cells to citrullinated rat serum albumin (Cit-RSA) in comparison with those of unmodified rat serum albumin (RSA). To investigate the clinical arthritogenic relevance of citrullination, the cartilage antigen rat collagen type II (CII) was modified and arthritis development was evaluated in the experimental rat model collagen-induced arthritis (CIA). In addition, to correlate the presence of citrullinated proteins with that of PAD4 with different stages of arthritis, we examined synovial tissue immunohistochemically at different time points of CIA.
Our study demonstrates, for the first time, the kinetics of the presence of citrullinated proteins as well as the enzyme PAD4 in arthritic joints from experimental animals. The amounts of citrullinated proteins and the enzyme PAD4 are correlated with severity of inflammation and are not detectable in healthy joints. The study also reveals that the citrullination of proteins can break natural tolerance mechanisms and increase the arthritogenic properties of CII.
Materials and methods
Animals
Lew.1AV1 and Dark Agouti (DA) rats, originating from the Zentralinstitut für Versuchstierzucht, Hannover, Germany, were bred, kept and used under pathogen-free conditions at the animal department of Karolinska University Hospital, Stockholm, Sweden. Animals were fed with standard rodent food and water ad libitum. Experiments were performed with female Lew.1AV1 or male DA rats, aged 9 to 14 weeks, with approval from the Stockholm North Ethical Committee.
Preparation and citrullination of antigen
CII was prepared from Swarm's rat chondrosarcoma by pepsin digestion and purification as previously described [37,38]. CII was stored freeze-dried at -20°C until dissolved in 0.01 M acetic acid and dialysed against the citrullination buffer, 0.1 M Tris-HCl (pH 7.6) containing 10 mM CaCl2 and 5 mM dithiothreitol. RSA (Sigma-Aldrich, Steinheim, Germany) was dissolved directly in citrullination buffer. Proteins were incubated with rabbit skeletal muscle PAD (Sigma), at a concentration of 2 U/mg protein, for 2 hours at 37°C. Citrullination was terminated by adding 20 mM EDTA and subsequently dialysed successively against 2 mM EDTA (pH7.6) and Milli-Q water containing 10 mM Tris-HCl (pH7.6) at 4°C. Control proteins were treated similarly, apart from the addition of PAD. Protein solutions were then freeze-dried and stored at -20°C. For experimental use, freeze-dried proteins were dissolved in appropriate buffer, namely RSA in PBS and CII in 0.01 M acetic acid.
To verify the purity and amount of proteins, aliquots of samples were run in SDS-PAGE gels followed by Coomassie Brilliant Blue staining. Total protein quantification was also performed with a modified version of the Bradford method, using the Coomassie Plus Protein Assay Reagent Kit (Pierce Biotechnology, Rockford, IL, USA) in accordance with the manufacturer's instructions. To verify successful citrullination, immunoblotting was performed with RA3 [39], a human recombinant antibody directed against citrulline (Fig. 1). In brief, 5 μl of citrullinated CII (Cit-CII), CII, Cit-RSA and RSA (concentration 2 mg/ml) were applied to a nitrocellulose membrane (BioTrace NT, product no. 66485; Pall Life Sciences, Ann Arbor, MI, USA). The membrane was incubated for 1 hour in blocking buffer (PBS containing 0.05% Tween 20 with 5% (w/v) non-fat dried milk) then incubated for a further 1 hour with primary antibody RA3, diluted 1:5 in blocking buffer, at room temperature (20–25°C). After incubation with a horseradish peroxidase (HRP)-conjugated donkey anti-human antibody (diluted 1:1,000 in blocking buffer; Amersham Pharmacia Biotech, Little Chalfont, Bucks., UK) for 1 hour at room temperature, the protein–antibody complexes were detected by enhanced chemiluminescence with the ECL system (Amersham Pharmacia Biotech, Uppsala, Sweden).
Induction and evaluation of clinical disease
To increase the likelihood of detecting citrullinated proteins and the enzyme PAD4 in the joints of arthritic animals, we chose to perform the immunohistochemical stainings in the DA rat, because this strain has proven to develop the most severe CIA.
In contrast, the Lew.1AV1 rat strain develops a milder disease and was therefore selected when investigating the additive arthritogenic effects of Cit-CII. Lew.1AV1 rats were also used when studying the effects of citrullination on tolerance to RSA.
Anaesthetised rats were immunised in the base of the tail: male DA rats with 150 μg of CII dissolved in 100 μl of 0.01 M acetic acid, emulsified with an equal volume of Freund's incomplete adjuvant (FIA; Difco, Detroit, MI, USA) (intradermally); female Lew.1AV1 rats with 120 μg of CII or Cit-CII dissolved in 100 μl of 0.01 M acetic acid plus 100 μl of FIA (intradermally) or with 180 μg of RSA or Cit-RSA dissolved in 150 μl of PBS plus 150 μl FIA (subcutaneously). The animals immunised with CII and Cit-CII were monitored for signs of arthritis from day 12 after immunisation, in accordance with a previously described procedure [40]. Each paw was divided into three groups of joints, the interphalangeal joints of the digits, the metacarpophalangeal and wrist joints in the forepaws and the metatarsophalangeal and ankle joints in the hind paws. In brief, 1 point signifies swelling of one group of joints, 2 points signifies two groups of swollen joints, 3 points signifies three groups of swollen joints and 4 points signifies swelling of the entire paw. The maximum score was thus 16 for each rat.
Evaluation of serum anti-RSA and anti-Cit-RSA antibody levels
Individual serum samples from Lew.1AV1 rats immunised with RSA and Cit-RSA were obtained at different time points, namely 12, 24 and 35 days after immunisation (by tail bleeding) as well as at 61 days after immunisation (by heart puncture). ELISA plates (96-well Maxisorp; Nunc, Roskilde, Denmark) were coated with 10 μg/ml RSA or Cit-RSA (diluted in PBS), overnight at 4°C. Plates were washed with PBS-Tween (0.05%) and sera (serially diluted in PBS containing 0.05% Tween) were added in duplicate. After 2 hours of incubation at room temperature, plates were washed as above and incubated with alkaline phosphatase-conjugated goat anti-rat IgG (diluted 1:5,000 in PBS containing 0.05% Tween; Jackson Immunoresearch Lab, West Grove, PA, USA) for a further 2 hours at room temperature. Finally, a phosphatase substrate (Sigma, St Louis, MO, USA) was added and absorbance was determined at 405 nm with an Emax precision microplate reader (Molecular Devices).
To investigate the risk of contamination by PAD in our Cit-RSA sample, giving a false positive result in our Cit-RSA ELISA, we performed an anti-PAD ELISA. Plates were coated with the same amount of PAD expected to contaminate the Cit-RSA sample in the Cit-RSA ELISA, namely 0.0432 μg/ml. The ELISA was performed with the same procedure as described for Cit-RSA and RSA (see above).
In addition, to quantify antibodies specific for anti-citrullinated protein, an in-house CCP ELISA was used. Control reactions with corresponding cyclic arginine-containing peptides were also included. In brief, Streptawell plates were coated with 1 μg of biotinylated cfc1-cyc citrullinated peptide or cf0-cyc control peptide (diluted in PBS, 0.1% BSA) overnight at 4°C. Rat serum samples (diluted 1:50 in PBS containing 0.05% Tween and 1% BSA) were added in duplicate and incubated for 1 hour in a 37°C humid chamber. Plates were washed before the addition of HRP-conjugated rabbit anti-rat IgG (Sigma P216; diluted 1:1,000 in PBS containing 0.05% Tween and 1% BSA), followed by incubation for 1 hour in a humid chamber at 37°C. Bound antibodies were revealed by the addition of 3,3',5,5'-tetramethylbenzidine (Sigma). The enzymic reaction was stopped with 0.5 M H2SO4 after 30 min, and absorbance was determined at 450 nm.
In vitro T cell proliferation in response to stimulation with RSA and Cit-RSA
Inguinal lymph nodes from Lew.1AV1 rats immunised with RSA and Cit-RSA were removed 10 days after immunisation; single-cell suspensions were prepared and resuspended in Dulbecco's modified Eagle's medium supplemented with glutamine, penicillin, streptomycin, HEPES and 10% fetal calf serum (Life Technologies, Paisley, Renfrewshire, UK). Cells were cultured in vitro for 72 hours at 106 cells per well in triplicate (96-well flat-bottomed culture plates; Nunc) in the presence of RSA (10 μg/ml), Cit-RSA (10 μg/ml), PBS or concanavalin A (2 μg/ml; Sigma-Aldrich). [3H]Thymidine (PerkinElmer Life Sciences Inc., Boston, MA, USA), was added (1 μCi per well) for the final 16 hours of culture. Cells were harvested with a Tomtec cell harvester and [3H]thymidine incorporation was measured as counts per minute (c.p.m.) in a Wallac Trillux 1450 microbeta counter. Stimulation was calculated and expressed as stimulation index (C.p.m. after stimulation/C.p.m. of background).
Antibodies
Recombinant anti-citrullinated protein single-chain variable fragment (scFv) antibody RA3, selected from RA-patient-derived phage display libraries and control scFv human anti-U1-70K have been described elsewhere [39]. Rabbit antibodies directed against chemically modified citrulline (anti-MC antibodies), developed by Dr Tatsuo Senshu and colleagues [5,41], were purchased from Upstate Biochemicals (Lake Placid, NY, USA). Polyclonal antibodies recognising PAD4 (SN823) were produced by immunising rabbits with PAD4 isotype-specific peptides (amino acids 210 to 225 and 517 to 531) and by affinity purification with a CNBr-activated Sepharose 4B column, as described previously [10]. Preimmune rabbit IgG was used as control for PAD4.
Immunohistochemical analysis
Zamboni-fixed [42] cryopreserved sections of synovial tissue from CII-immunised DA rats were stained for expression of citrullinated proteins and PAD4. Endogenous peroxidase activity was blocked by treatment for 30 min in darkness at room temperature with 1% hydrogen peroxide and 2% sodium nitride dissolved in PBS. After washing with PBS, 2% normal goat serum was added to block non-specific binding sites. Sections were then incubated with 200 μl of primary antibody (RA3 (undiluted) or anti-PAD4 (diluted 1:24)) overnight in a humid chamber at 4°C. An isotype-matched recombinant human anti-U1-70K antibody and a preimmune rabbit IgG were used as respective controls. Slides were then washed in PBS before incubation with 200 μl of appropriate biotin-labelled secondary antibody (donkey anti-human IgG (Jackson Immuno Research) diluted 1:1,000 or goat anti-rabbit IgG (Vector, Burlingame, CA, USA) diluted 1:800) for 1 hour at room temperature. After further washing, 200 μl of avidin-biotin-HRP (Vectastain; Vector) prepared in accordance with the manufacturer's directions was applied for 60 min at room temperature. A final wash was followed by addition of the substrate diaminobenzidine (Peroxidase Substrate Kit; Vector) for 5 min. The colour reaction was stopped by washes in tap water, after which sections were counterstained with Mayer's haematoxylin. Finally the slides were dried and mounted with buffered glycerol, and evaluation was performed by microscopy with a Polyvar II microscope (Reichert-Jung, Vienna, Austria) connected to a charge-coupled device colour camera (DXC-750P; Sony Corp., Tokyo, Japan).
In addition, anti-MC antibodies were used to confirm the citrulline staining obtained with RA3. Before incubation with primary antibody, sections were treated for 3 hours at 37°C in a chemical modification solution consisting of one part solution A (0.025% (w/v) FeCl3, 4.6 M H2SO4, 3.0 M H3PO4) plus one part solution B (0.5% diacetyl monoxime, 0.25% antipyrine, 0.5 M acetic acid) (anti-Citrulline (Modified) Detection Kit; Upstate Biochemicals). Control sections were incubated in a solution containing one part solution A and one part Milli-Q water. After extensive washing in PBS, slides were incubated for 40 min at room temperature with 1% H2O2, washed in PBS and incubated for 30 min at room temperature with 5% normal goat serum in PBS plus 1% BSA. Primary antibody (anti-MC), diluted 1:1,000 in PBS plus 1% BSA, was added and slides were incubated overnight at room temperature. Slides were then washed in PBS, followed by incubation for 30 min at room temperature with secondary antibody (biotinylated goat anti-rabbit IgG; Vector) diluted 1:800 in PBS plus 1% BSA. The subsequent steps were performed as described above.
Statistical analysis
All data were evaluated with the Mann–Whitney U-test for independent groups, except for arthritis incidence, which was evaluated by Kaplan–Meier survival analysis.
Results
Presence of citrullinated proteins and PAD4 in the arthritic joints correlates with the degree of inflammation
To examine the presence of citrullinated proteins and the enzyme PAD4 in the joints at different stages of experimental arthritis, histological analyses of rat ankle joints were performed by immunohistochemistry. Citrullinated proteins could be detected in the joints of arthritic animals with the RA3 antibody (Fig. 2 and Table 1). The first appearance of protein citrullination was noted after disease onset, at day 21 after immunisation, and increased staining was observed as the disease progressed into a more severe, chronic state (namely 28 and 38 days after immunisation). Unimmunised animals and time points before clinical signs of arthritis (namely 0, 3, 6 and 10 days after immunisation) as well as the time of disease onset (namely 15 days after immunisation) were negative for citrulline. Some infiltrating cells as well as the cartilage surface stained positively for citrulline and the major sources of citrullinated proteins in the arthritic joint were extracellular deposits, presumably fibrin deposits. The occurrence of citrullinated proteins in the joints is specific, because other investigated organs such as lung, ear, spinal cord, spleen, lymph nodes and salivary glands were all negative (data not shown).
For comparison with previous studies in humans [43] and mice [13] we also used antibodies targeting chemically modified citrulline. This method circumvents the risk of epitope blocking by residues flanking citrulline, because the citrulline side-chain becomes so bulky through the chemical modification that antibody recognition cannot be influenced by other amino acids [5,41]. Here we confirmed the results obtained with RA3. A similar staining pattern was observed, with positive cells, cartilage and extracellular deposits, whereas control stainings of non-modified sections were negative (data not shown).
PAD4, which has been reported to be present in mouse and human arthritic synovia [10,13], could also be detected in synovial tissue of this rat arthritis model from 21 days after immunisation (Fig. 3 and Table 1). The number of positive cells localised to the synovial infiltrate increased by days 28 and 38 after immunisation. PAD4 was not evident in healthy synovial tissue, nor in sections from 3, 6 and 10 days after immunisation; although an apparent inflammation was observed, PAD4 could not be detected 15 days after immunisation.
Citrullination of a non-immunogenic antigen breaks B cell tolerance
Lew.1AV1 rats were immunised with the non-immunogenic autoantigen RSA or the modified counterpart Cit-RSA and the differences in induced immune responses were analysed in vitro. The kinetics of the antigen-specific IgG response was investigated and the results revealed not only a response towards the modified protein, but also cross-reactivity to the unmodified form of RSA (Fig. 4). Antibodies were detected from 12 days after immunisation, had increased by day 24 after immunisation and persisted for a further 40 days. Cross-reactivity was demonstrated at all time points. No B cell response was noted in animals immunised with unmodified RSA.
Animals immunised with Cit-RSA and RSA were also tested for anti-PAD IgG responses. The anti-PAD ELISA was negative both for animals immunised with Cit-RSA and for animals immunised with RSA (data not shown), indicating that the amount of PAD contaminating the Cit-RSA sample did not result in any false positive result in the Cit-RSA ELISA, described above.
In addition, by performing both an anti-CCP ELISA and a control ELISA containing cyclic arginine-containing peptides, we could confirm that the animals immunised with Cit-RSA produced antibodies against citrullinated epitopes and also antibodies recognising arginine epitopes (data not shown).
The T cell response was evaluated by [3H]thymidine incorporation 10 days after immunisation, as depicted in Fig. 5. Here, a proliferative response was demonstrated in animals immunised with unmodified RSA, both when stimulated with the same antigen and when stimulated with the citrullinated antigen. Cells from animals immunised with Cit-RSA showed a stronger proliferation than the RSA-immunised rats in response to the modified protein. The same tendency was observed when using RSA as a stimulus, although it was not statistically significant.
Immunisation with Cit-CII increases arthritis incidence and accelerates clinical onset of disease
Unmodified or citrullinated rat CII was, together with adjuvant, injected into Lew.1AV1 rats; arthritis development was monitored by blinded macroscopic evaluation. The citrullinated form of CII induced arthritis with significantly higher incidence and earlier onset than did the same amount of unmodified CII (Fig. 6). The incidence was 35% higher in the Cit-RSA group during the early phase of the disease and 15% higher during the end stage. Disease onset in this group occurred 13 days after immunisation, compared with 16 days after immunisation in the control group. A trend towards higher mean arthritis score was also observed in the affected animals in the Cit-CII group compared with the CII group, although this was not statistically significant.
Discussion
The production of anti-citrullinated protein antibodies is almost 100% specific for patients with RA, indicating an important role for citrullinated proteins in the pathogenesis of RA. We therefore considered it of interest to investigate the ability to break immunological tolerance by the citrullination of endogenous proteins. Our data reveal that citrullination of the non-immunogenic antigen RSA induced an antibody response against the citrullinated protein with cross-reactivity to the unmodified protein at all time points investigated. The observation that anti-RSA IgG titres declined by 61 days after immunisation, as opposed to anti-cit-RSA IgG titres, might be due to affinity maturation of the B cell response with bias towards the citrullinated 'neoepitope'. Another possibility could be the formation of RSA antibodies and RSA immune complexes, which would increase their clearance from the circulation.
We could record an ex vivo proliferative T cell response to RSA as well as to Cit-RSA in rats immunised with modified protein and also in rats immunised with the unmodified protein, as demonstrated by stimulation index values over 1.0. When comparing the proliferative responses to Cit-RSA, cells from animals immunised with the citrullinated protein demonstrated a significantly stronger response than cells from animals immunised with the unmodified protein (P < 0.05). A similar trend was detected when comparing the responses of T cells to RSA, but the difference did not reach statistical significance. This indicates that RSA, together with a strong adjuvant, can induce an autoreactive T cell response, although this is not sufficient to induce B cell help. In contrast, immunisation with Cit-RSA induced a stronger T cell response that could confer B cell help, as depicted in Fig. 4.
Furthermore, Cit-CII induced arthritis with earlier onset and higher incidence than unmodified CII after immunisation. Considering the low arginine content of CII, our results indicate that protein modification by the conversion of arginine to citrulline is a highly potent mechanism for increased autoreactivity. Previous studies have demonstrated that anti-CII antibody titres do not reflect disease severity in this model of CIA and that stimulation in vitro with homologous CII does not induce T cell proliferation. We therefore chose not to analyse the humoral or cellular responses to CII in these animals.
Several post-translational protein modifications, especially those related to apoptosis [44,45], are associated with autoimmunity [46-49]. Inefficient clearance of these modified proteins in an inflammatory environment, conveying 'danger signals' to the immune system, might in combination with the appropriate MHC haplotype override tolerance mechanisms and activate autoreactive T cells. Citrullination interferes with organised protein structure, contributes to protein unfolding and to increased susceptibility to digesting enzymes [50,51], which could lead to altered antigen uptake, processing and presentation. Indeed, citrullinated peptides have been reported to bind more efficiently to the RA-associated HLA-DRB1*0401 haplotype than do non-citrullinated peptides [52]. As tolerance has not been established to such peripherally modified self-peptides, citrullination might increase the risk of activating pathogenic T cells. Our data therefore demonstrate not only the potency of citrullinated proteins to break tolerance, as evident from our experiments with RSA, but also the increased arthritogenic properties of Cit-CII.
Because anti-citrullinated protein antibodies can be detected in RA sera before clinical signs of disease [24], we speculate that citrullination of synovial proteins might occur at time points before RA becomes manifest. It is not ethically possible to examine the synovial tissue of healthy individuals at risk of developing RA. However, this becomes feasible with the use of an experimental model. In the present study we examined, for the first time, the kinetics of the presence of citrullinated proteins and the citrullinating enzyme PAD4 in joints during arthritis development. In this experimental model of homologous CIA with 100% disease incidence, citrullinated proteins and PAD4 were not detectable before clinical signs of arthritis. Rather, as inflammation proceeded, increasing amounts of citrullinated proteins and PAD4 were detected specifically in the joints. Extracellular deposits constitute the major source of citrullinated proteins in the inflamed joint, although both cells and cartilage also contribute to the positive staining. Although PAD4 has been described to have a nuclear localisation [12], our stainings clearly demonstrate the presence of this enzyme in the cytoplasm of infiltrating mononuclear cells. Our inability to detect nuclear staining of PAD4 might be due to the absence of permeabilising agents during staining procedures. Our finding that citrullinated synovial proteins could not be encountered before disease onset might explain the failure in detecting anti-citrullinated protein antibodies in animal models. Perhaps the presence of citrullinated proteins in these arthritis models is not disease-initiating, but rather a result of inflammation. This is also supported by the observation that citrullinated proteins can be detected in inflamed synovial tissue of non-RA patients lacking the anti-citrulline-specific B cell response [53,54].
Taking together our findings and those of others, we hypothesise that in individuals with the appropriate genetic background a subclinical inflammation that induces citrullination of endogenous proteins will generate a pathogenic immune response against the modified proteins and that through the production of autoantibodies a chronic inflammation will be established. An immune response towards any citrullinated protein is clearly not enough to induce clinical arthritis, as demonstrated by the lack of signs of disease in the Cit-RSA-immunised animals. Both genetic and environmental factors most probably contribute to the development of arthritis. The initiation of citrulline reactivity does not necessarily need to occur in the joints. For example, smoking has been reported to be an environmental risk factor for RA in individuals with MHC shared epitope [55]. It is plausible that smoking could induce such an inflammatory reaction in lungs with the formation of citrullinated proteins. Why an autoreactivity to citrullinated proteins would precipitate in arthritis and not in other organ-specific inflammatory diseases is currently unknown. However, several reports indicate that joints are especially sensitive to inflammatory stimuli. The experimental model oil-induced arthritis demonstrates that administration of the non-antigenic adjuvant mineral oil at a distant location (namely the tail base) induces arthritis and no other disease in rats [56]. Additionally, systemic overexpression of proinflammatory cytokine TNF generates a joint-specific inflammation [57]. Similarly, interleukin-1Ra knockout mice develop arthritis [58].
Conclusion
Our data reveal the potency of citrullination to break tolerance against the ubiquitous systemic self antigen RSA and to increase the arthritogenicity of the tissue-specific protein CII. Furthermore, we have demonstrated that the amounts of citrullinated proteins and the enzyme PAD4 in the arthritic joints of experimental animals correlated with the severity of inflammation, while no citrullinated proteins or PAD4 could be detected in healthy joints. On the basis of our new data and previous findings on this topic we hypothesise that a subclinical inflammation might induce citrullination of endogenous proteins. This citrullination does not necessarily need to occur in joints but can occur elsewhere in the body. In combination with a certain genetic context, anti-citrullinated protein antibodies will be generated, which will potentiate an inflammatory trigger in joints and thereby initiate RA.
Abbreviations
Anti-CCP = anti-cyclic citrullinated peptide; anti-MC antibodies = antibodies against chemically modified citrulline; BSA = bovine serum albumin; CIA = collagen-induced arthritis; CII = collagen type II; Cit-CII = citrullinated CII; Cit-RSA = citrullinated RSA; DA = Dark Agouti; ELISA = enzyme-linked immunosorbent assay; FIA = Freund's incomplete adjuvant; HRP = horseradish peroxidase; MHC = major histocompatibility complex; PAD = peptidyl arginine deiminase; PBS = phosphate-buffered saline; RA = rheumatoid arthritis; RSA = rat serum albumin.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
KL was responsible for most of the experiments and data analysis as well as drafting the manuscript. HEH, together with KL, was responsible for study design coordination and compilation of the manuscript. SN and ERV performed the citrullination of proteins and the production of antibodies (RA3, anti-U1-70K, anti-PAD4 and preimmune rabbit sera). KP performed the immunisation, scoring and sectioning of rat ankle joints of the DA rats. WJV, LK and AJWZ contributed to interpretation and discussion of data. All authors read and approved the final manuscript.
Acknowledgements
The authors thank Kalok Cheung for performing the CCP ELISA and Associate Professor Robert A Harris for linguistic advice. This work was supported by grants from the Swedish Science council, the Swedish Association Against Rheumatism, the Foundation of King Gustav V, the Börje Dahlin Foundation, the Nanna Svartz Foundation, the Af Ugglas Foundation, the Netherlands Foundation for Chemical Research and the Netherlands Technology Foundation (grant 349–5077), Het Nationaal Reumafonds of The Netherlands (The Dutch League against Rheumatism, grant 00-2-402) and the Netherlands Foundation for Medical Research (NOW grant 940-35-037).
Figures and Tables
Figure 1 Verification of citrullination of collagen type II (CII) and rat serum albumin (RSA) by immunoblotting. Using the recombinant anti-citrullinated protein antibody RA3, citrulline was detected in the citrullinated CII (Cit-CII) sample (upper left) as well as in the citrullinated RSA (Cit-RSA) sample (upper right), while uncitrullinated CII (lower left) and uncitrullinated RSA (lower right) were negative for citrulline.
Figure 2 Citrullinated proteins are present in the arthritic joint. Positive citrulline staining was found in extracellular deposits (a,e), cartilage (c) and infiltrating cells (d). Unimmunised animals were negative for citrulline (f). Immunohistochemical staining was performed with RA3, a human recombinant anti-citrullinated protein antibody. Control staining was performed with an isotyped-matched recombinant human anti-U1-70K antibody (b). (Original magnifications: ×100 (a,b); ×250 (c); ×400 (d); ×40 (e,f)).
Figure 3 Peptidyl arginine deiminase 4 (PAD4) is present in the arthritic joint. PAD4 was detected in infiltrating cells (a), localised to the cytoplasm of mononuclear cells (c,e). Unimmunised animals were negative for PAD4 staining (f). Immunohistochemical stainings were performed with a rabbit anti-PAD4 antibody. Control staining was performed with preimmune rabbit sera (b,d) (Original magnifications: ×40 (a,b,f); ×200 (c-e)).
Figure 4 Immunisation with citrullinated rat serum albumin (Cit-RSA) breaks immunological tolerance on the B cell side. Immunisation with Cit-RSA (white bars) induced an antibody response against Cit-RSA (a), cross-reacting with unmodified rat serum albumin (RSA) (b) at all time points investigated, whereas immunisation with RSA (black bars) did not. Sera were collected 12, 24, 35 and 61 days after immunisation, and total IgG was measured by ELISA as OD at 405 nm. Results are means ± SD (n = 7 animals per group), representative of two replicate experiments. **P < 0.01 at all time points.
Figure 5 In vitro proliferative responses to rat serum albumin (RSA) and to citrullinated RSA (Cit-RSA). Stronger T cell responses to Cit-RSA (statistically significant) and to RSA (a trend) were observed in animals immunised with Cit-RSA (white bars) than in animals immunised with unmodified RSA (black bars). As previously shown (Fig. 4), the animals immunised with Cit-RSA developed a B cell response to RSA and to Cit-RSA, whereas the animals immunised with RSA did not. The dotted line suggests a hypothetical threshold that it is necessary to reach to induce B cell help and subsequent antibody production. Single cell suspensions were prepared from inguinal lymph nodes 10 days after immunisation, and stimulation index (S.I.) was calculated after 72 hours of culture in the presence of RSA or Cit-RSA (10 μg/ml). Results are means ± SD (n = 7 animals per group), representative of two replicate experiments. *P < 0.05.
Figure 6 Citrullination of collagen type II (CII) increases its arthritogenic properties. Lew.1AV1 rats developed disease with higher incidence (a) and earlier onset as well as a trend towards higher mean arthritic score among the affected animals (b) when immunised with citrullinated CII (Cit-CII) (open squares) than when immunised with unmodified CII (filled circles). Rats were immunised with Cit-CII or unmodified CII and monitored for signs of arthritis for the next 30 days. Incidence (a) and mean score of sick animals (b) were calculated. Data include 20 animals per group from two pooled experiments. *P < 0.05.
Table 1 Correlation of presence of citrullinated proteins and peptidyl arginine deiminase 4 with degree of inflammation
Days after immunisation Arthritis score Cell infiltration Cartilage destruction Citrulline-positive deposits Citrulline-positive cells Citrulline-positive cartilage PAD4-positive cells
0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
3 0.0 0.0 0.0 0.0 0.0 0.0 0.0
6 0.0 0.0 0.0 0.0 0.0 0.0 0.0
10 0.0 0.0 0.0 0.0 0.0 0.0 0.0
15 3.3 ± 2.4 1.3 ± 0.5 0.0 0.0 0.0 0.0 0.0
21 9.8 ± 3.3 1.8 ± 0.5 0.5 ± 0.6 1.0 0.5 ± 0.6 0.8 ± 1.0 1.0 ± 0.5
28 11.0 ± 3.5 3.0 2.5 ± 0.6 1.8 ± 1.0 1.3 ± 0.5 1.5 ± 1.3 1.5 ± 0.4
38 11.8 ± 2.6 3.0 3.0 2.7 ± 0.6 1.8 ± 0.5 0.5 ± 0.6 2.2 ± 0.5
PAD4, peptidyl arginine deiminase 4. Arthritic score, cell infiltration, cartilage destruction and presence of citrullinated proteins in extracellular deposits, cells and cartilage were evaluated at different time points of collagen-induced arthritis. All data are from four rats, evaluated in a blinded manner by two independent evaluators and are means ± SD. Arthritis score ranges from 0 to 16 as explained in the Materials and methods section. Arbitrary units were used for the other features, ranging from 0, indicating no cell infiltration, cartilage destruction, presence of citrullinated proteins or PAD4, to 3, indicating extensive cell infiltration, cartilage destruction, presence of citrullinated proteins and PAD4.
==== Refs
van Boekel MA Vossenaar ER van den Hoogen FH van Venrooij WJ Autoantibody systems in rheumatoid arthritis: specificity, sensitivity and diagnostic value Arthritis Res 2002 4 87 93 11879544 10.1186/ar395
Steiner G Smolen J Autoantibodies in rheumatoid arthritis and their clinical significance Arthritis Res 2002 4 Suppl 2 S1 S5 12110150 10.1186/ar551
Schellekens GA de Jong BA van den Hoogen FH van de Putte LB van Venrooij WJ Citrulline is an essential constituent of antigenic determinants recognized by rheumatoid arthritis-specific autoantibodies J Clin Invest 1998 101 273 281 9421490
Girbal-Neuhauser E Durieux JJ Arnaud M Dalbon P Sebbag M Vincent C Simon M Senshu T Masson-Bessiere C Jolivet-Reynaud C The epitopes targeted by the rheumatoid arthritis-associated antifilaggrin autoantibodies are posttranslationally generated on various sites of (pro)filaggrin by deimination of arginine residues J Immunol 1999 162 585 594 9886436
Senshu T Akiyama K Kan S Asaga H Ishigami A Manabe M Detection of deiminated proteins in rat skin: probing with a monospecific antibody after modification of citrulline residues J Invest Dermatol 1995 105 163 169 7543546 10.1111/1523-1747.ep12317070
Senshu T Kan S Ogawa H Manabe M Asaga H Preferential deimination of keratin K1 and filaggrin during the terminal differentiation of human epidermis Biochem Biophys Res Commun 1996 225 712 719 8780679 10.1006/bbrc.1996.1240
Vossenaar ER Zendman AJ van Venrooij WJ Pruijn GJ PAD, a growing family of citrullinating enzymes: genes, features and involvement in disease BioEssays 2003 25 1106 1118 14579251 10.1002/bies.10357
Watanabe K Akiyama K Hikichi K Ohtsuka R Okuyama A Senshu T Combined biochemical and immunochemical comparison of peptidylarginine deiminases present in various tissues Biochim Biophys Acta 1988 966 375 383 3416014
Nagata S Senshu T Peptidylarginine deiminase in rat and mouse hemopoietic cells Experientia 1990 46 72 74 2404790
Vossenaar ER Radstake TR Van Der Heijden A Van Mansum MA Dieteren C De Rooij DJ Barrera P Zendman AJ Van Venrooij WJ Expression and activity of citrullinating peptidylarginine deiminase enzymes in monocytes and macrophages Ann Rheum Dis 2004 63 373 381 15020330 10.1136/ard.2003.012211
Asaga H Nakashima K Senshu T Ishigami A Yamada M Immunocytochemical localization of peptidylarginine deiminase in human eosinophils and neutrophils J Leukoc Biol 2001 70 46 51 11435484
Nakashima K Hagiwara T Yamada M Nuclear localization of peptidylarginine deiminase V and histone deimination in granulocytes J Biol Chem 2002 277 49562 49568 12393868 10.1074/jbc.M208795200
Vossenaar ER Nijenhuis S Helsen MM van der Heijden A Senshu T van den Berg WB van Venrooij WJ Joosten LA Citrullination of synovial proteins in murine models of rheumatoid arthritis Arthritis Rheum 2003 48 2489 2500 13130468 10.1002/art.11229
Hoet RM Boerbooms AM Arends M Ruiter DJ van Venrooij WJ Antiperinuclear factor, a marker autoantibody for rheumatoid arthritis: colocalisation of the perinuclear factor and profilaggrin Ann Rheum Dis 1991 50 611 618 1718228
Nienhuis RL Mandema E A new serum factor in patients with rheumatoid arthritis; the antiperinuclear factor Ann Rheum Dis 1964 23 302 305 14178016
Simon M Girbal E Sebbag M Gomes-Daudrix V Vincent C Salama G Serre G The cytokeratin filament-aggregating protein filaggrin is the target of the so-called 'antikeratin antibodies,' autoantibodies specific for rheumatoid arthritis J Clin Invest 1993 92 1387 1393 7690781
Young BJ Mallya RK Leslie RD Clark CJ Hamblin TJ Anti-keratin antibodies in rheumatoid arthritis Br Med J 1979 2 97 99 111762
Sebbag M Simon M Vincent C Masson-Bessiere C Girbal E Durieux JJ Serre G The antiperinuclear factor and the so-called antikeratin antibodies are the same rheumatoid arthritis-specific autoantibodies J Clin Invest 1995 95 2672 2679 7539459
Despres N Boire G Lopez-Longo FJ Menard HA The Sa system: a novel antigen-antibody system specific for rheumatoid arthritis J Rheumatol 1994 21 1027 1033 7932409
Vossenaar ER Despres N Lapointe E Van Der Heijden A Lora M Senshu T Van Venrooij WJ Menard HA Rheumatoid arthritis specific anti-Sa antibodies target citrullinated vimentin Arthritis Res Ther 2004 6 R142 R150 15059278 10.1186/ar1149
Schellekens GA Visser H de Jong BA van den Hoogen FH Hazes JM Breedveld FC van Venrooij WJ The diagnostic properties of rheumatoid arthritis antibodies recognizing a cyclic citrullinated peptide Arthritis Rheum 2000 43 155 163 10643712 10.1002/1529-0131(200001)43:1<155::AID-ANR20>3.0.CO;2-3
van Venrooij WJ Hazes JM Visser H Anticitrullinated protein/peptide antibody and its role in the diagnosis and prognosis of early rheumatoid arthritis Neth J Med 2002 60 383 388 12607587
Kurki P Aho K Palosuo T Heliovaara M Immunopathology of rheumatoid arthritis. Antikeratin antibodies precede the clinical disease Arthritis Rheum 1992 35 914 917 1379430
Rantapaa-Dahlqvist S de Jong BA Berglin E Hallmans G Wadell G Stenlund H Sundin U van Venrooij WJ Antibodies against cyclic citrullinated peptide and IgA rheumatoid factor predict the development of rheumatoid arthritis Arthritis Rheum 2003 48 2741 2749 14558078 10.1002/art.11223
Masson-Bessiere C Sebbag M Durieux JJ Nogueira L Vincent C Girbal-Neuhauser E Durroux R Cantagrel A Serre G In the rheumatoid pannus, anti-filaggrin autoantibodies are produced by local plasma cells and constitute a higher proportion of IgG than in synovial fluid and serum Clin Exp Immunol 2000 119 544 552 10691929 10.1046/j.1365-2249.2000.01171.x
Reparon-Schuijt CC van Esch WJ van Kooten C Schellekens GA de Jong BA van Venrooij WJ Breedveld FC Verweij CL Secretion of anti-citrulline-containing peptide antibody by B lymphocytes in rheumatoid arthritis Arthritis Rheum 2001 44 41 47 11212174 10.1002/1529-0131(200101)44:1<41::AID-ANR6>3.0.CO;2-0
Kirstein H Mathiesen FK Antikeratin antibodies in rheumatoid arthritis. Methods and clinical significance Scand J Rheumatol 1987 16 331 338 2446384
Bas S Perneger TV Mikhnevitch E Seitz M Tiercy JM Roux-Lombard P Guerne PA Association of rheumatoid factors and anti-filaggrin antibodies with severity of erosions in rheumatoid arthritis Rheumatology (Oxford) 2000 39 1082 1088 11035127
Paimela L Gripenberg M Kurki P Leirisalo-Repo M Antikeratin antibodies: diagnostic and prognostic markers for early rheumatoid arthritis Ann Rheum Dis 1992 51 743 746 1377463
Paimela L Palosuo T Aho K Lukka M Kurki P Leirisalo-Repo M von Essen R Association of autoantibodies to filaggrin with an active disease in early rheumatoid arthritis Ann Rheum Dis 2001 60 32 35 11114279 10.1136/ard.60.1.32
Forslin K Vincent C Serre G Svensson B Antifilaggrin antibodies in early rheumatoid arthritis may predict radiological progression Scand J Rheumatol 2001 30 221 224 11578017 10.1080/030097401316909567
van Jaarsveld CH ter Borg EJ Jacobs JW Schellekens GA Gmelig-Meyling FH van Booma-Frankfort C de Jong BA van Venrooij WJ Bijlsma JW The prognostic value of the antiperinuclear factor, anti-citrullinated peptide antibodies and rheumatoid factor in early rheumatoid arthritis Clin Exp Rheumatol 1999 17 689 697 10609067
Vencovsky J Machacek S Sedova L Kafkova J Gatterova J Pesakova V Ruzickova S Autoantibodies can be prognostic markers of an erosive disease in early rheumatoid arthritis Ann Rheum Dis 2003 62 427 430 12695154 10.1136/ard.62.5.427
Meyer O Labarre C Dougados M Goupille P Cantagrel A Dubois A Nicaise-Roland P Sibilia J Combe B Anticitrullinated protein/peptide antibody assays in early rheumatoid arthritis for predicting five year radiographic damage Ann Rheum Dis 2003 62 120 126 12525380 10.1136/ard.62.2.120
Goldbach-Mansky R Lee J McCoy A Hoxworth J Yarboro C Smolen JS Steiner G Rosen A Zhang C Menard HA Rheumatoid arthritis associated autoantibodies in patients with synovitis of recent onset Arthritis Res 2000 2 236 243 11056669 10.1186/ar93
van Gaalen FA van Aken J Huizinga TW Schreuder GM Breedveld FC Zanelli E van Venrooij WJ Verweij CL Toes RE deVries RR Association between HLA class II genes and autoantibodies to cyclic citrullinated peptides (CCPs) influences the severity of rheumatoid arthritis Arthritis Rheum 2004 50 2113 2121 15248208 10.1002/art.20316
Miller EJ Structural studies on cartilage collagen employing limited cleavage and solubilization with pepsin Biochemistry 1972 11 4903 4909 4565026 10.1021/bi00776a005
Andersson M Holmdahl R Analysis of type II collagen-reactive T cells in the mouse. I. Different regulation of autoreactive vs. non-autoreactive anti-type II collagen T cells in the DBA/1 mouse Eur J Immunol 1990 20 1061 1066 1694131
Raats JM Wijnen EM Pruijn GJ van den Hoogen FH van Venrooij WJ Recombinant human monoclonal autoantibodies specific for citrulline-containing peptides from phage display libraries derived from patients with rheumatoid arthritis J Rheumatol 2003 30 1696 1711 12913924
Åkerlund K Erlandsson Harris H Tracey KJ Wang H Fehniger T Klareskog L Andersson J Andersson U Anti-inflammatory effects of a new tumour necrosis factor-alpha (TNF-alpha) inhibitor (CNI-1493) in collagen-induced arthritis (CIA) in rats Clin Exp Immunol 1999 115 32 41 9933418 10.1046/j.1365-2249.1999.00750.x
Senshu T Sato T Inoue T Akiyama K Asaga H Detection of citrulline residues in deiminated proteins on polyvinylidene difluoride membrane Anal Biochem 1992 203 94 100 1524220 10.1016/0003-2697(92)90047-B
Ahmed M Bjurholm A Schultzberg M Theodorsson E Kreicbergs A Increased levels of substance P and calcitonin gene-related peptide in rat adjuvant arthritis. A combined immunohistochemical and radioimmunoassay analysis Arthritis Rheum 1995 38 699 709 7538298
Masson-Bessiere C Sebbag M Girbal-Neuhauser E Nogueira L Vincent C Senshu T Serre G The major synovial targets of the rheumatoid arthritis-specific antifilaggrin autoantibodies are deiminated forms of the alpha- and beta-chains of fibrin J Immunol 2001 166 4177 4184 11238669
Doyle HA Mamula MJ Posttranslational protein modifications: new flavors in the menu of autoantigens Curr Opin Rheumatol 2002 14 244 249 11981321 10.1097/00002281-200205000-00009
Utz PJ Gensler TJ Anderson P Death, autoantigen modifications, and tolerance Arthritis Res 2000 2 101 114 11094420 10.1186/ar75
Arentz-Hansen H Korner R Molberg O Quarsten H Vader W Kooy YM Lundin KE Koning F Roepstorff P Sollid LM The intestinal T cell response to alpha-gliadin in adult celiac disease is focused on a single deamidated glutamine targeted by tissue transglutaminase J Exp Med 2000 191 603 612 10684852 10.1084/jem.191.4.603
Neugebauer KM Merrill JT Wener MH Lahita RG Roth MB SR proteins are autoantigens in patients with systemic lupus erythematosus. Importance of phosphoepitopes Arthritis Rheum 2000 43 1768 1778 10943867 10.1002/1529-0131(200008)43:8<1768::AID-ANR13>3.0.CO;2-9
Zamvil SS Mitchell DJ Moore AC Kitamura K Steinman L Rothbard JB T-cell epitope of the autoantigen myelin basic protein that induces encephalomyelitis Nature 1986 324 258 260 2431317 10.1038/324258a0
Andrade F Casciola-Rosen L Rosen A Apoptosis in systemic lupus erythematosus. Clinical implications Rheum Dis Clin North Am 2000 26 215 227 10768210
Tarcsa E Marekov LN Mei G Melino G Lee SC Steinert PM Protein unfolding by peptidylarginine deiminase. Substrate specificity and structural relationships of the natural substrates trichohyalin and filaggrin J Biol Chem 1996 271 30709 30716 8940048 10.1074/jbc.271.48.30709
Lamensa JW Moscarello MA Deimination of human myelin basic protein by a peptidylarginine deiminase from bovine brain J Neurochem 1993 61 987 996 7689646
Hill JA Southwood S Sette A Jevnikar AM Bell DA Cairns E Cutting edge: the conversion of arginine to citrulline allows for a high-affinity peptide interaction with the rheumatoid arthritis-associated HLA-DRB1*0401 MHC class II molecule J Immunol 2003 171 538 541 12847215
Chapuy-Regaud S Sebbag M Baeten D Clavel C De Keyser F Serre G The presence of deiminated fibrin in the synovial membrane is not specific for rheumatoid arthritis Arthritis Res Ther 2004 6 Suppl 1 20 10.1186/ar1062
Vossenaar ER Smeets TJ Kraan MC Raats JM Van Venrooij WJ Tak PP The presence of citrullinated proteins is not specific for rheumatoid synovial tissue Arthritis Rheum 2004 50 3485 3494 15529392 10.1002/art.20584
Padyukov L Silva C Stolt P Alfredsson L Klareskog L A gene-environment interaction between smoking and shared epitope genes in HLA-DR provides a high risk of seropositive rheumatoid arthritis Arthritis Rheum 2004 50 3085 3092 15476204 10.1002/art.20553
Kleinau S Erlandsson H Holmdahl R Klareskog L Adjuvant oils induce arthritis in the DA rat. I. Characterization of the disease and evidence for an immunological involvement J Autoimmun 1991 4 871 880 1812893 10.1016/0896-8411(91)90050-M
Butler DM Malfait AM Mason LJ Warden PJ Kollias G Maini RN Feldmann M Brennan FM DBA/1 mice expressing the human TNF-alpha transgene develop a severe, erosive arthritis: characterization of the cytokine cascade and cellular composition J Immunol 1997 159 2867 2876 9300710
Horai R Saijo S Tanioka H Nakae S Sudo K Okahara A Ikuse T Asano M Iwakura Y Development of chronic inflammatory arthropathy resembling rheumatoid arthritis in interleukin 1 receptor antagonist-deficient mice J Exp Med 2000 191 313 320 10637275 10.1084/jem.191.2.313
| 15899032 | PMC1174941 | CC BY | 2021-01-04 16:02:35 | no | Arthritis Res Ther. 2005 Feb 21; 7(3):R458-R467 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1697 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar16981589904410.1186/ar1698Research ArticleSynovial histopathology of psoriatic arthritis, both oligo- and polyarticular, resembles spondyloarthropathy more than it does rheumatoid arthritis Kruithof Elli [email protected] Dominique [email protected] Rycke Leen [email protected] Bernard [email protected] Dirk [email protected] Johannes [email protected]ñete Juan D [email protected] Annemieke M [email protected] Eric M [email protected] Keyser Filip [email protected] Department of Rheumatology, Ghent University Hospital, Ghent, Belgium2 Department of Pediatrics, University of Münster, Münster, Germany3 Institute of Experimental Dermatology, University of Münster, Münster, Germany4 Department of Rheumatology, Hospital Clinic de Barcelona, Universitat de Barcelona, Barcelona, Spain5 Department of Pharmacology, Organon NV, Oss, The Netherlands2005 3 3 2005 7 3 R569 R580 17 9 2004 21 10 2004 18 1 2005 Copyright © 2005 Kruithof et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
At present only few biological data are available to indicate whether psoriatic arthritis (PsA) is part of the spondyloarthropathy (SpA) concept, whether it is a separate disease entity or a heterogeneous disease group with oligoarticular/axial forms belonging to SpA and polyarticular forms resembling rheumatoid arthritis (RA). To address this issue with regard to peripheral synovitis, we compared the synovial characteristics of PsA with those of ankylosing spondylitis (AS)/undifferentiated SpA (USpA) and RA, and compared the synovium of oligoarticular versus polyarticular PsA. Synovial biopsies were obtained from patients with RA, nonpsoriatic SpA (AS + USpA), and oligoarticular and polyarticular PsA. The histological analysis included examination(s) of the lining layer thickness, vascularity, cellular infiltration, lymphoid aggregates, plasma cells and neutrophils. Also, we performed immunohistochemical assessments of CD3, CD4, CD8, CD20, CD38, CD138, CD68, CD163, CD83, CD1a, CD146, αVβ3, E-selectin, intercellular adhesion molecule-1, vascular cell adhesion molecule-1, S100A12, intracellular citrullinated proteins and major histocompatibility complex (MHC)–human cartilage (HC) gp39 peptide complexes. Comparing SpA (PsA + AS + USpA) with RA, vascularity, and neutrophil and CD163+ macrophage counts were greater in SpA (P < 0.05), whereas lining layer thickness and the number of CD83+ dendritic cells were greater in RA (P < 0.05). In RA, 44% of samples exhibited positive staining for intracellular citrullinated proteins and 46% for MHC–HC gp39 peptide complexes, whereas no staining for these markers was observed in SpA samples. We excluded influences of disease-modifying antirheumatic drug and/or corticosteroid treatment by conducting systematic analyses of treated and untreated subgroups. Focusing on PsA, no significant differences were observed between PsA and nonpsoriatic SpA. In contrast, vascularity (P < 0.001) and neutrophils were increased in PsA as compared with RA (P = 0.010), whereas staining for intracellular citrullinated proteins and MHC–HC gp39 peptide complexes was exclusively observed in RA (both P = 0.001), indicating that the same discriminating features are found in PsA and other SpA subtypes compared with RA. Exploring synovial histopathology between oligoarticular and polyarticular PsA, no significant differences were noted. Moreover, intracellular citrullinated proteins and MHC–HC gp39 peptide complexes, which are specific markers for RA, were observed in neither oligoarticular nor polyarticular PsA. Taken together, these data indicate that the synovial histopathology of PsA, either oligoarticular or polyarticular, resembles that of other SpA subtypes, whereas both groups can be differentiated from RA on the basis of these same synovial features, suggesting that peripheral synovitis in PsA belongs to the SpA concept.
==== Body
Introduction
Psoriatic arthritis (PsA) is a chronic inflammatory autoimmune disease that is characterized by inflammatory arthritis and skin psoriasis. It is commonly classified as a subtype of the spondyloarthropathy (SpA) concept, together with ankylosing spondylitis (AS), reactive arthritis, arthritis associated with inflammatory bowel disease, and undifferentiated spondyloarthropathy (USpA) [1]. Indeed, PsA shares a number of phenotypic characteristics with the other SpA subtypes, such as asymmetrical synovitis, enthesitis and familial aggregation. Typical SpA features such as sacroiliitis, presence of HLA-B27, uveitis and chronic inflammatory gut lesions are also found in PsA, albeit at a lower frequency than in other SpA subtypes [2-4]. Interestingly, the peripheral joint involvement is mostly asymmetrical and oligoarticular, as in SpA, but can also mimic rheumatoid arthritis (RA) with eventually destructive involvement of multiple small hand and feet joints [5]. Finally, skin and nail involvement, total ankylosis of destructed peripheral joints, predominant distal interphalyngeal joint involvement and arthritis mutilans sets PsA apart from both SpA and RA. Therefore, there is at present no clear consensus regarding whether PsA belongs to the SpA concept, whether it is a separate disease entity, or whether it is a heterogeneous disease group with oligoarticular and axial forms belonging to the SpA concept on the one hand and destructive, polyarticular forms resembling RA on the other [6,7]. This question will probably only be answered once the pathogenesis of the disease is fully elucidated, leading to an aetiological rather than phenotypical disease classification.
In previous histopathological synovial studies we considered PsA as part of the SpA concept and compared it as such with other inflammatory joint diseases. Although some early data are available on PsA synovial histology based on small studies of surgical or blind needle specimens, only few studies systematically evaluated PsA synovium obtained from actively inflamed joints [8,9]. Moreover, most of those studies used RA as a control and did not compare PsA and nonpsoriatic SpA, or oligoarticular and polyarticular PsA. In this context it must be highlighted that the few reports dealing with PsA synovium did not specify whether these patients had oligoarticular or polyarticular disease. The present study is the first to address these issues systematically and quantitatively in a large set of actively inflamed synovium samples by (1) conducting a detailed analysis of PsA synovitis in comparison with AS–USpA on the one hand and RA on the other, using a large panel of histopathological features; (2) comparing oligoarticular versus polyarticular PsA with respect to synovial features that discriminate between RA and SpA; and (3) analyzing the expression of S100A12, which was reported to be an interesting marker in PsA [10].
Materials and methods
Patients
In accordance with the aims of the study, three different patient cohorts were included. Cohort 1 was considered in the comparison of synovial histopathology between PsA, AS–USpA and RA. It included 22 patients with PsA having asymmetrical peripheral synovitis and skin psoriasis, and moreover fulfilling the European Spondyloarthropathy Study Group (ESSG) criteria [11], 28 nonpsoriatic SpA patients fulfilling the ESSG criteria (including 13 AS patients diagnosed according to the modified New York criteria [12] and 15 USpA patients), and 52 RA patients fulfilling the American College of Rheumatology (ACR) criteria [13]. Of the 22 PsA patients, six had a swollen joint count of more than 5 but with predominant involvement of the lower limbs and presence of other characteristics of SpA (i.e. enthesitis, sacroiliitis, familial history, etc.). None of them fulfilled the ACR criteria for RA [13]. Demographic and clinical features are summarized in Table 1.
Cohort 2 included a large number of PsA patients (n = 45), permitting us to compare synovial histopathology between oligoarticular and polyarticular disease. Of these patients 28 had oligoarticular disease with involvement of fewer than 5 peripheral joints, and the remaining 17 patients had polyarticular disease (≥ 5 joints). Demographic and clinical features are summarized in Table 2.
Cohort 3 was considered in the analysis of S100A12 by immunohistochemistry of synovial tissue. It included eight PsA patients (fulfilling the ESSG criteria [11]), 12 nonpsoriatic SpA patients (according to the ESSG criteria) and 20 RA patients fulfilling the ACR criteria [13]. The demographic and clinical features of these patients are summarized in Table 3.
Since we demonstrated previously that synovial histopathology is influenced by local disease activity in SpA and RA, all included patients were required to have a clinical effusion of at least one knee joint to perform needle arthroscopy with synovial biopsy sampling [14]. Written informed consent was obtained from all patients before their inclusion in the study, which was approved by the Ethics Committee of the Ghent University Hospital.
Synovial histopathology
Synovial biopsies (16 per patient) were obtained by needle arthroscopy of the knee as described previously [15]. Eight biopsies were stored in formaldehyde and embedded in paraffin, and eight biopsies were snap frozen and mounted in Jung tissue freezing medium (Leica Instruments, Nussloch, Germany) and utilized for immunohistochemistry. The procedure for histological and immunohistochemical analysis of the different markers was extensively described and validated previously [14,16-20]. Briefly, paraffin-embedded biopsies were stained with haematoxylin and eosin for histological analysis, including mean synovial lining layer thickness (score: 1 = mean of 1–2 cell layers, 2 = mean of 3–5 cell layers, 3 = mean of >5 cell layers), vascularity of the sublining layer, global cellular infiltration of the sublining layer, and presence of lymphoid aggregates, plasma cells and polymorphonuclear cells (PMCs). Frozen sections of the synovial biopsies were stained with the following antibodies (from Dako [Glostrup, Denmark], unless otherwise stated): anti-CD3 (T cells, clone UCHT1), anti-CD4 (T-helper cells, clone MT310), anti-CD8 (cytotoxic T cells, clone DK25), anti-CD20 (B cells, clone L26), anti-CD38 (plasma cells, clone AT13/5), anti-CD138 (plasma cells, clone CBL455; Chemicon, Temecula, CA, USA), anti-CD68 (pan-macrophage marker expressed on monocytes and macrophages, clone EBM11), anti-CD163 (scavenger receptor expressed on resident tissue macrophages, clone Ber-MAC3), anti-CD83 (dendritic cells, clone HB15A; Immunotech SA, Marseille, France), anti-CD1a (interdigitating dendritic cells, clone NA1/34), anti-CD146 (endothelial cells, clone P1H12; Chemicon), anti-αVβ3 (CD51/CD61, integrin expressed on endothelial cells, fibroblasts, osteoclasts, etc., clone LM609; Chemicon), anti-E-selectin (CD62E, endothelial leucocyte adhesion molecule 1, clone 1.2B6), anti-ICAM-1 (CD54, intercellular adhesion molecule [ICAM]-1, clone 6.5B5), anti-VCAM-1 (CD106, vascular cell adhesion molecule [VCAM]-1, clone 1.4C3), the rabbit polyclonal anti-L-citrulline antibody (citrullinated peptides; Biogenesis, Poole, UK), and Mab-12A (detecting major histocompatibility complex [MHC] class II–human cartilage [HC] gp39 peptide complexes; NV Organon, Oss, The Netherlands) [18,21]. Immunostaining for S100A12 (calgranulin C) was performed with polyclonal affinity-purified rabbit antisera against human S100A12 (a S100A12) [22].
It should be noted that the scavenger receptor CD163 identifies a subset of CD68+ macrophages that is specifically increased in SpA synovium and gut, and which could play an important role in innate immune inflammation of these tissues [16,19].
After incubation with the primary antibody, sections were sequentially incubated with a biotinylated second antibody, a streptavidine–horseradish peroxidase link, and finally with amino-ethyl-carbazole substrate as chromogen. Parallel sections were incubated with irrelevant isotype and concentration matched monoclonal antibody as negative control. Sections were coded and analyzed semiquantitatively on a 4-point scale (0–3) by two independent observers who were blinded to diagnosis and clinical data. Because the numbers of positive cells per synovial section for citrullinated peptides and Mab-12A were too small to be scored semiquantitatively, these parameters were scored as present or absent. The analysis included all areas of the eight biopsies and a global score was given for each parameter, using a semiquantitative 4-point scale (0 = lowest level of expression, 3 = highest level of expression) [14,16-21]. Because some histological markers are more abundant than others, the scoring system was calibrated for each marker separately by examining a representative number of samples. The scores obtained by the two observers were concordant in more than 95% of cases. In the event of discordant scores, which differed by a maximum of 1 point, the mean of the two scores was used.
Statistics
Because the semiquantitative data are nonparametric, these data are presented as median (range). Differences between groups were analyzed using the nonparametric Mann–Whitney U-test. For markers scored as present or absent, the χ2 test was used. Bonferroni adjustment of alpha was performed where indicated. P < 0.05, after Bonferroni correction, was considered statistically significant.
Results
Synovial histopathology of spondyloarthropathy versus rheumatoid arthritis
We previously established the differences between SpA (including PsA fulfilling ESSG criteria) and RA in terms of global synovial histology and infiltrating cell populations [14,16]. Therefore, in patient cohort 1 we compared the following synovial features between the pooled SpA group (50 samples) and the RA group (52 patients): lining layer thickness, vascularity, global cellular infiltration, lymphoid follicles, plasma cells, PMCs, CD3, CD4, CD8, CD20, CD38, CD138, CD68, CD163, CD1a, CD83, CD146, αVβ3, E-selectin, ICAM-1, VCAM-1, intracellular citrullinated proteins and MHC–HC gp39 peptide complexes.
As shown in Table 4, the lining layer thickness (P = 0.006) and number of CD83+ dendritic cells (P = 0.006) were significantly greater in RA than in SpA. There was also a trend toward greater CD38+ plasma cell infiltration in RA than in SpA (P = 0.060). On the contrary, vascularity (P < 0.001), infiltration with PMCs (P = 0.008), and the presence of CD163+ macrophages in the lining layer (P = 0.033) and sublining layer (P = 0.031) were higher in SpA (although the medians for PMC infiltration did not differ, reflecting that only a subset of SpA samples was characterized by high degree of PMC infiltration). In agreement with previous studies [16,19], the increase in the CD163+ macrophage subset in SpA was not parallalled by changes in the global number of macrophages, as detected by the panmacrophage marker CD68. ICAM-1 staining in the synovial lining layer (P = 0.025), but not in the sublining or on the vascular endothelium, was also higher in SpA than in RA. Intracellular citrullinated proteins and MHC–HC gp39 peptide complexes were observed only in RA patients (44% positive for citrullinated proteins and 46% positive for MHC–HC gp39 peptide complexes), with absence of staining for these markers in the SpA samples (for both: P < 0.001).
There were no other significant differences between both groups. Taken together these data confirm our previous observations and indicate that the samples selected for the study are representative of the global histopathological picture of SpA and RA. The differences are illustrated in Fig. 1.
Influence of disease-modifying antirheumatic drug and steroid treatment on synovial histopathology
Because the studied disease groups in cohort 1 received different therapies (Table 1) and because previous longitudinal studies with paired biopsy sampling have demonstrated the effect of disease-modifying antirheumatic drugs (DMARDs) and systemic corticosteroids on synovial histopathology [23-27], we assessed whether DMARD and/or systemic corticosteroid therapy could bias a cross-sectional histopathological evaluation of clinically involved joints.
As shown in more detail in Table 1, six out of 22 PsA patients, eight out of 28 AS–USpA patients, and 25 out of 52 RA patients were treated with DMARDs and/or systemic corticosteroids. For the RA group, we compared patients with versus those without DMARD treatment, as well as patients with versus those without corticosteroid treatment with respect to the synovial features mentioned above. Only endothelial VCAM-1 (higher in RA with than in RA without DMARD treatment: 2.0 [0–3.0] versus 0 [0–3.0]; P = 0.035) and endothelial E-selectin (lower in RA with than in RA without corticosteroid treatment: 0.5 [0–1.5] versus 1.5 [0–3.0]; P = 0.017) were significantly different as a function of treatment.
For the SpA group (PsA + AS + USpA), we compared 13 DMARD-treated patients with 37 patients who received no DMARD treatment. Only ICAM-1 expression was significantly different as a function of DMARD treatment (lower in SpA with than in SpA without DMARD treatment: 2.5 [1.0–3.0] versus 3.0 [1.5–3.0] for the lining layer [P = 0.033], and 1.0 [0–2.5] versus 2.5 [0–3.0] for the sublining layer [P = 0.003]).
Focusing on the PsA group, a comparison of patients with versus those without DMARD treatment for the same parameters revealed no significant differences.
Synovial histopathology of psoriatic arthritis versus rheumatoiad asthritis and versus ankylosing spondylitis–undifferentiated spondyloarthropathy
Having confirmed histopathological differences between RA and SpA synovitis and having excluded a systemic bias introduced by therapy, we then focused more specifically on PsA. Using the same histological markers, in patient cohort 1 we compared 22 PsA with 52 RA synovia samples. The degree of synovial vascularity was higher in PsA (median 2.5 [range 1.0–3.0]) than in RA (1.5 [1.0–3.0]; P < 0.001). Similarly, the presence of PMCs was more pronounced in PsA (0.5 [0–3.0]) than in RA (0 [0–2.5]; P = 0.010). Intracellular citrullinated proteins and MHC–HC gp39 peptide complexes were not observed in PsA, whereas 44% of RA samples were positive for citrullinated proteins and 46% were positive for MHC–HC gp39 peptide complexes (for both: P = 0.001). There was no significant difference for the other investigated markers, including E-selectin and CD68, which were previously reported to be different between both diseases [8]. Performing a similar analysis in function of AS–USpA, we compared the 22 PsA samples with 28 nonpsoriatic SpA samples (13 AS and 15 USpA); none of the investigated markers was significantly different between both groups.
Oligoarticular versus polyarticular psoriatic arthritis
It was previously suggested that PsA could comprise different subsets, including an oligoarticular subtype that resembles SpA and a polyarticular form mimicking RA. To address this issue, we constituted a new PsA cohort (patient cohort 2) including 28 patients with oligoarticular disease (<5 swollen joints) and 17 patients with polyarticular PsA (≥ 5 swollen joints). We compared the PsA subgroups with respect to those synovial features that appeared to discriminate best between SpA and RA in the previous experiments or that have been reported in the literature [8]: lining layer hyperplasia, vascularity, PMCs, CD163+ and CD68+ macrophages, and expression of E-selectin. As shown in Table 5, no significant differences in these parameters were observed between oligoarticular and polyarticular PsA. Again, intracellular citrullinated proteins and MHC–HC gp39 peptide complexes – two markers specific for RA – were not observed in either the oligoarticular or polyarticular PsA samples.
S100A12
Finally, we investigated whether, apart from the well known histopathological markers evaluated in the previous experiments, additional specific immunohistochemical stainings were helpful in the evaluation of PsA synovitis compared with nonpsoriatic SpA and RA. Based on the greater infiltration with PMCs in SpA than in RA, and a recent report on the granulocyte calcium-binding protein S100A12 in PsA [10], we assessed the expression of S100A12 in synovium of patients with PsA (n = 8), nonpsoriatic SpA (n = 12) and RA (n = 20; patient cohort 3; Table 3). Cellular staining for S100A12 was observed in the sublining layer in all three patient subgroups, whereas staining in the lining layer was only occasionally observed (Fig. 2). Confirming the findings of the previous study [10], S100A12 expressing cells were found essentially in perivascular infiltrates, either adjacent to small blood vessels or adhering to endothelial cells. However, this staining pattern was observed in all groups without clear distinction between PsA, nonpsoriatic SpA and RA. This difference may be accounted for by the fact that, in the study conducted by Foell and coworkers [10], the disease duration of the patients at study entry was shorter, and none of the PsA-patients received any DMARDs or steroids. Moreover, there was no significant difference in the degree of staining between PsA (median 1 [range 0–3]) and nonpsoriatic SpA (1, [0–2]; P = 1.000) or RA (median 0 [0–3]; P = 0.430). Pooling of the PsA and nonpsoriatic SpA samples in one SpA group did not result a difference between SpA and RA either (P = 0.096).
Discussion
Because the synovial membrane is one of the primary tissue targets in PsA, detailed histopathological analysis of PsA synovitis might be a useful approach to gain new insights into this disease and to analyze biological similarities with or differences from other chronic inflammatory joint disorders such as RA and SpA. This is of particular importance because PsA is thought to resemble either RA or SpA, depending on the clinical pattern of peripheral joint involvement, with respectively polyarticular symmetrical disease and oligoarticular asymmetrical involvement. In this context, a detailed analysis of the synovial histopathology in PsA not only may provide new insights in the disease process but also may help to provide a biological rationale for the classification of PsA. However, until now such analyses are lacking.
Synovial histopathology in PsA is generally characterized by neovascularization, and inflammatory infiltration with predominantly mononuclear cells (T lymphocytes, B lymphocytes and plasma cells, and macrophages), although PMCs can also be detected [28]. Mild to moderate synovial lining hyperplasia is observed in a considerable percentage of cases. Whereas these characteristics are found as well in other types of inflammatory arthritis, the availability of synovial biopsies from needle arthroscopy and new histopathological tools (monoclonal antibodies, RNA probes) have permitted a more detailed comparison of PsA and RA synovitis. One of the most prominent differences was the higher degree and typical morphology of synovial vascularity, which appeared to be related to angiogenic factors such as vascular endothelial growth factor, matrix metalloproteinase (MMP)-9 and angiopoietins, and was accompanied by a higher expression of E-selectin [8,29,30]. Lining layer hyperplasia appeared to be more pronounced in RA, both by altered apoptosis of lining cells as by increased presence of CD68+ macrophage-like synoviocytes, although macrophage-derived pro-inflammatory cytokines such as tumour necrosis factor-α were also found abundantly in PsA synovium [8,31-33]. Apart from a difference in CD68+ macrophages and the somewhat unexpected presence of PMCs in PsA synovitis [8,28], no major differences in infiltrating cell populations were described between PsA and RA. Although these differences between PsA and RA are of interest, it should be highlighted that these reports did not clearly distinguish between different clinical subtypes of PsA and did not compare PsA with SpA.
Considering that the oligoarticular form of PsA might belong to the SpA concept, both the present study and previous reports of our group indicate similar differences between SpA and RA as those described between PsA and RA: higher lining layer thickness in RA, but more pronounced vascularity, presence of polymorphonuclear cells and presence of the CD163+ macrophage subset in SpA.[14,16,19]. However, we could not confirm earlier reported differences in the presence of CD68+ macrophages between the disease groups [8]. These findings of the present study are in agreement with our previous reports in SpA [16,19], in which we indicated that the increased synovial infiltration with CD163+ macrophages – a subset that may play an important role in innate immune inflammation – is not paralleled by an increase in global macrophage number, as detected using the pan-macrophage marker CD68. This is also in agreement with a recent independent study [34] comparing PsA with RA, in which no difference was found in the number of CD68+ macrophages between the diseases.
In order to confirm whether similarities between SpA and PsA exist, as suggested by these independent studies, the present study provides a direct comparison of oligoarticular PsA with RA and with nonpsoriatic SpA. Our findings clearly demonstrate the resemblance between oligoarticular PsA and SpA synovium, whereas oligoarticular PsA synovium is significantly different from RA in terms of lining layer hyperplasia and PMC infiltration. Having previously indicated the importance of local disease activity but not disease duration in the assessment of synovial histopathology, the present study furthermore provides evidence that the obtained results are not biased by differences in treatment between the groups. Confirming the observations of other recent studies [19,35,36], these data certainly do not suggest that DMARD and/or corticosteroid treatment have no influence on synovial histopathology [23-27] but they indicate that global synovial histopathology is not altered in the case of persistent synovitis despite DMARD and/or systemic corticosteroid treatment.
Bearing in mind that subclassification of PsA is based on the pattern of peripheral joint involvement [6,7], it is surprising that, until now, no synovial histology data were available detailing oligoarticular PsA, which is thought to be related to SpA, and polyarticular PsA, which can mimic RA. Using histological parameters that discriminate between SpA and RA, such as lining layer thickness, degree of vascularity, presence of PMCs and CD163 expression, we were unable to demonstrate any significant differences between oligoarticular and polyarticular PsA. In addition, earlier reported markers distinguishing PsA from RA, such as CD68 and E-selectin, appeared not to be differentially expressed between oligoarticular and polyarticular PsA [8,37]. Considering the discrepancies with regard to published data on CD68 and E-selectin expression between PsA and RA, one should consider potential biases in study populations or in treatment schedules. Moreover, intracellular citrullinated proteins and Mab-12A staining, which are highly specific for RA and were found in approximately half of the RA synovia [17,18], were not detected in the PsA cohorts. These data provide the first clear evidence that polyarticular PsA does not exhibit RA-specific synovial features and that the synovial histopathology of PsA, either oligoarticular or polyarticular, is more closely related to SpA than to RA.
Apart from global histological features, a number of studies have analyzed more specific molecular systems in PsA synovium, including S100 proteins. The calcium-binding granulocyte protein S100A12 was recently described in PsA synovium [10]: although the number of analyzed samples was too small to allow statistical comparison, a specific expression in PsA as compared with RA was suggested. In the present study we confirmed both the presence and the staining pattern of S100A12 in PsA synovium, but neither synovial tissue analysis nor synovial fluid measurements identified differences between PsA, SpA and RA. Two other S100 proteins (S100A8 and S100A9, respectively named myeloid-related protein [MRP]8 and MRP14), which are found in both infiltrating PMCs and monocytes, were recently described in different types of inflammatory arthritis [36-40]. Similar to the findings of the present study, the expressions of these MRPs in synovium and synovial fluid were different between SpA and RA but not between PsA and nonpsoriatic SpA [36]. Furthermore, we recently provided evidence that the effect of tumour necrosis factor-α blockade with infliximab on MRPs, as well as on global histological features, was similar in PsA and in nonpsoriatic SpA, further supporting the concept that peripheral joint disease is closely related in both diseases [41,42].
Nevertheless, the data provided by the present study and the resulting hypothesis require some critical comments. First, it should be noted that all patients included in the present study had an active knee synovitis, which could bias the study population by excluding polyarticular PsA with predominant or exclusive involvement of hand and feet joints. In addition, DMARD therapy has been shown to represent a confounding factor in terms of the pattern and number of swollen joints in established PsA; patients with polyarticular joint involvement may therefore be under-represented in our patient cohort [7]. Moreover, because occasional evolution to polyarticular disease suggests that PsA is a more systemic disease than is nonpsoriatic SpA, it should be further investigated whether there are specific histological alterations in clinically uninvolved joints in PsA versus AS and USpA. Second, more sensitive scoring (digital image analysis versus semiquantitative scoring) or detection methods might have revealed small differences that could have been overlooked in the present study. However, such small differences should be considered in the light of their likelihood of being reproduced on the one hand and their biological relevance on the other [43].
A third criticism of the present study is that it addressed a wide range of histopathological markers that are known to discriminate either PsA or SpA from RA, but the resulting concept requires confirmation by studies of cellular and molecular players, which are crucially involved in the pathogenesis of peripheral synovitis. Therefore, novel mediators identified in vitro or in animal models of PsA should be further evaluated. Considering the specific radiological features of peripheral joint involvement in PsA, an interesting issue – which was not addressed in the present study – is the synovial mechanism involved in cartilage and/or bone destruction. However, we recently investigated a set of metalloproteinases and their inhibitors in a similar study [35] that indicated that there was no difference in their synovial expression between SpA and RA. Surprisingly, the minor differences between PsA and nonpsoriatic SpA indicated less pronounced expression of MMP-1 and MMP-2 in PsA.
Another recent study [44] demonstrated that synovial osteoclasts and the receptor activator of nuclear factor-κB (RANK)/RANK ligand system, which are known to play a crucial role in bone degradation in RA, were also clearly present in PsA synovium. However, a systematic comparison of synovial osteoclasts and RANK/RANK ligand in PsA, SpA and RA remains to be conducted. Taken together, these three issues indicate that the absence of histopathological differences between PsA and non-psoriatic SpA in the present study does not indicate that these conditions are similar. Further studies are needed to address these issues in greater detail.
Conclusion
The present study is the first to provide detailed comparisons of PsA synovitis with both RA and SpA, and of oligoarticular with polyarticular PsA synovitis. It indicates that the synovial histopathology of PsA, either oligoarticular or polyarticular, resembles SpA more than it does RA. These biological data do not support the clinical subclassification of PsA in a polyarticular RA-like and an oligoarticular SpA-like subtypes, and more importantly, they emphasize the need of more specific analyses of cellular and molecular characteristics, especially those that are involved in cartilage and bone pathology, to unravel the pathogenetic and phenotypic differences and similarities.
Abbreviations
ACR = American College of Rheumatology; AS = ankylosing spondylitis; DMARD = disease-modifying antirheumatic drug; ESSG = European Spondyloarthropathy Study Group; HC = human cartilage; ICAM = intercellular adhesion molecule; MHC = major histocompatibility complex; MMP = matrix metalloproteinase; MRP = myeloid-related protein; PMC = polymorphonuclear cell; PsA = psoriatic arthritis; RA = rheumatoid arthritis; RANK = receptor activator of nuclear factor-κB; SpA = spondyloarthropathy; USpA = undifferentiated spondyloarthropathy; VCAM = vascular cell adhesion molecule.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
EK collected the samples, participated in the immunohistochemistry, performed statistical analysis and the interpretation of the study, and prepared the manuscript. DB designed the study, participated in its coordination, participated in the interpretation of the results, and drafted the manuscript. LDR participated in the collection of the samples, in the immunohistochemistry and in the interpretation of the results. BV participated in the collection of the samples and in the interpretation of the results. DF provided the polyclonal affinity-purified rabbit antisera against human S100A12 and participated in the interpretation of the results. JR provided the polyclonal affinity-purified rabbit antisera against human S100A12 and participated in the interpretation of the results. JC participated in the collection of the samples and in the interpretation of the results. AB provided the antibody detecting the MHC class II-HC gp39 peptide complexes and participated in the interpretation of the results. EMV supervised the collection of the samples as well as the design of the study. FDK participated in the design of the study and in its coordination, and participated in the interpretation of the results. All authors read and approved the final manuscript.
Acknowledgements
The authors wish to thank Virgie Baert for excellent technical assistance.
Dominique Baeten is a Senior Clinical Investigator of the Fund for Scientific Research-Flanders (FWO-Vlaanderen). The work of Leen De Rycke was supported by the 'Vlaams instituut voor de bevordering van het wetenschappelijk-technologisch onderzoek in de industrie' (IWT/SB/11127). Bernard Vandooren is a Research Fellow of the Fund for Scientific Research-Flanders (FWO-Vlaanderen). Juan Cañete is supported by the Fondo de Investigaciones Sanitarias grant G03/152.
Figures and Tables
Figure 1 Synovial histology in patients with rheumatoid arthritis (RA), psoriatic arthritis (PsA) and nonpsoriatic spondyloarthropathy (ankylosing spondylitis [AS] + undifferentiated spondyloarthropathy [USpA]). Synovial biopsies from RA, PsA and spondyloarthropathy (SpA; AS/USpA) patients were scored on a semiquantitative scale (0–3) by two independent observers. Representative sections of RA and PsA and nonpsoriatic SpA synovium are shown, and the corresponding semiquantitative score for each picture is indicated. The evaluated parameters included synovial lining layer thickness, CD83+ dendritic cells, CD38+ plasma cells, degree of vascularity, number of neutrophils (polymorphnuclear neutrophils [pmn]) and CD163+ macrophages.
Figure 2 Synovial expression of S100A12 in psoriatic arthritis (PsA), nonpsoriatic spondyloarthropathy (SpA; ankylosing spondylitis [AS] + undifferentiated spondyloarthropathy [USpA]) and rheumatoid arthritis (RA). (a) PsA: lining = 0, sublining = 1 (original magnification 640×); (b) nonpsoriatic SpA: lining = 0, sublining = 2 (original magnification 640×); and (c) RA: lining = 0, sublining = 2 (original magnification 640×).
Table 1 Demographic and clinical features of patients with PsA, nonpsoriatic SpA (AS + USpA) and RA.
Parameter PsA Nonpsoriatic SpA RA
Number of patient 22 28 52
Sex (male/female) 14/8 22/6 25/27
Age (years) 45.6 ± 11.0 43.3 ± 13.7 54.0 ± 14.9
Disease duration (years) 10.1 ± 11.5 8.4 ± 9.8 4.9 ± 6.2
Number of swollen joints 4.1 ± 4.0 3.7 ± 4.0 10.5 ± 6.8
Serum CRP (mg/dl) 2.1 ± 1.9 4.0 ± 5.8 6.0 ± 7.4
ESR (mm/hour) 23.6 ± 22.2 30.1 ± 30.5 35.7 ± 27.1
Rheumatoid factor (+/-) 0/21 1/26 36/13
NSAID (+/-) 15/7 25/3 38/14
DMARD (+/-) 6/16 7/21 21/31
Corticosteroids (+/-) 1/21 2/26 12/40
The nonpsoriatic SpA subgroup includes 13 patients with AS and 15 with USpA. Values are expressed as mean ± standard deviation or as numbers. AS, ankylosing spondylitis; CRP, C-reactive protein; DMARD, disease-modifying antirheumatic drug; ESR, erythrocyte sedimentation rate; NSAID, nonsteroidal anti-inflammatory drug; PsA, psoriatic arthritis; RA, rheumatoid arthritis; SpA, spondyloarthropathy; USpA, undifferentiated spondyloarthropathy.
Table 2 Demographic and clinical features of patients with oligoarticular and polyarticular PsA.
Parameter Oligoarticular PsA Polyarticular PsA
Number of patients 28 17
Sex (male/female) 20/8 9/8
Age (years) 42.6 ± 10.9 43.7 ± 10.7
Disease duration (years) 7.4 ± 8.4 11.9 ± 11.1
Number of swollen joints 1.7 ± 1.0 9.2 ± 4.2
Serum CRP (mg/dl) 2.3 ± 3.6 3.4 ± 2.2
ESR (mm/hour) 19.2 ± 18.2 40.8 ± 21.1
Rheumatoid factor (positive/negative) 0/28 0/17
Sacroiliitis (≥grade II; +/-) 2/15 4/12
Enthesitis (+/-) 6/14 10/6
Hand joint involvement (+/-) 9/12 14/3
Symmetrical joint involvement (+/-) 4/19 10/7
DIP joint involvement (+/-) 8/11 10/7
NSAID (+/-) 21/6 15/1
DMARD (+/-) 8/19 10/6
Corticosteroids (treated/untreated) 2/25 3/13
Demographic and clinical features of patients with oligoarticular and polyarticular PsA. Values are expressed as mean ± standard deviation or as numbers. CRP, C-reactive protein; DIP, distal interphalangeal joint; DMARD, disease-modifying antirheumatic drugs; ESR, erythrocyte sedimentation rate; NSAID, nonsteroidal anti-inflammatory drug.
Table 3 Demographic and clinical features of patients with PsA, nonpsoriatic SpA and RA
Parameter PsA Nonpsoriatic SpA RA
Number of patients 8 12 20
Sex (male/female) 4/4 9/3 8/12
Age (years) 45.1 ± 7.1 39.7 ± 8.9 58.9 ± 17.2
Disease duration (years) 6.0 ± 9.7 6.6 ± 6.0 4.5 ± 7.9
Number of swollen joints 3.0 ± 1.9 2.6 ± 3.1 11.7 ± 8.3
Serum CRP (mg/dl) 6.2 ± 6.5 4.3 ± 5.1 7.3 ± 7.4
ESR (mm/hour) 36.4 ± 34.7 23.8 ± 25.2 44.2 ± 28.9
Rheumatoid factor (+/-) 0/8 0/11 14/6
DMARD (+/-) 3/3 8/4 8/12
Corticosteroids (+/-) 1/5 0/12 9/11
The nonpsoriatic SpA subgroup includes six AS and six USpA patients. Values are expressed as mean ± standard deviation. AS, ankylosing spondylitis; CRP, C-reactive protein; DMARD, disease-modifying antirheumatic drug; ESR, erythrocyte sedimentation rate; NSAID, nonsteroidal anti-inflammatory drug; PsA, psoriatic arthritis; RA, rheumatoid arthritis; USpA, undifferentiated spondyloarthropathy.
Table 4 Comparison of histological markers between SpA (PsA + AS + USpA) synovium and RA synovium
Marker SpA (n = 50) RA (n = 52) P
Lining layer thickness 1.0 (1.0–2.5) 2.0 (0–3.0) 0.006
Vascularity 2.0 (1.0–3.0) 1.5 (1.0–3.0) <0.001
Cellular infiltration 1.5 (0–3.0) 2.0 (0–3.0) 0.120
Lymphoid follicles 0 (0–3.0) 0 (0–3.0) 0.232
Plasma cells 0.5 (0–3.0) 0.5 (0–3.0) 0.497
Polymorphonuclear cells 0 (0–3.0) 0 (0–2.5) 0.008
CD3 1.5 (0–3.0) 2.0 (0–3.0) 0.152
CD4 1.0 (0–3.0) 1.0 (0–3.0) 0.192
CD8 1.0 (0–3.0) 1.5 (0–3.0) 0.737
CD20 1.0 (0–3.0) 2.0 (0–3.0) 0.231
CD38 1.5 (0–3.0) 2.0 (0–3.0) 0.060
CD138 1.0 (0–3.0) 2.0 (0–3.0) 0.447
CD68 lining 1.0 (0–3.0) 1.0 (0–3.0) 0.631
CD68 sublining 1.5 (0–3.0) 1.5 (0.5–3.0) 0.957
CD163 lining 2.0 (0–3.0) 1.5 (0–3.0) 0.033
CD163 sublining 1.5 (0–3.0) 1.0 (0–2.5) 0.031
CD1a 0 (0–3.0) 1.0 (0–3.0) 0.121
CD83 0 (0–1.0) 0 (0–3.0) 0.006
CD146 2.0 (0–3.0) 0 (0–3.0) 0.075
αVβ3 lining 1.0 (0–3.0) 1.0 (0–3.0) 0.397
αVβ3 sublining 1.0 (0–3.0) 1.0 (0–3.0) 0.306
E-selectin blood vessels 1.0 (0–3.0) 1.5 (0–3.0) 0.760
ICAM-1 lining 3.0 (1.0–3.0) 2.5 (0–3.0) 0.025
ICAM-1 sublining 2.0 (0–3.0) 2.0 (0–3.0) 0.654
ICAM-1 blood vessels 2.5 (0.5–3.0) 2.5 (0–3.0) 0.652
VCAM-1 lining 3.0 (0–3.0) 3.0 (0–3.0) 0.077
VCAM-1 sublining 1.5 (0–3.0) 1.5 (0–3.0) 0.473
VCAM-1 blood vessels 1.0 (0–3.0) 1.5 (0–3.0) 0.997
Intracellular citrullinated proteins 0% 44% <0.001
Mab-12A 0% 46% <0.001
The Mann–Whitney U-test was used to evaluate markers scored semiquantitatively, and the χ2 test was used to evaluate markers scored as present or absent. Results are expressed as median (range), and markers evaluated as present or absent are expressed as percentage of positive results. P < 0.05 was considered statistically significant. AS, ankylosing spondylitis; ICAM, intercellular adhesion molecule; PsA, psoriatic arthritis; RA, rheumatoid arthritis; USpA, undifferentiated spondyloarthropathy; VCAM, vascular cell adhesion molecule.
Table 5 Comparison of histological markers between oligoarticular and polyarticular PsA
Marker Oligoarticular PsA (n = 28) Polyarticular PsA (n = 17) P
Lining layer thickness 1.0 (1.0–3.0) 1.0 (1.0–2.0) 0.492
Vascularity 1.5 (0–3.0) 2.0 (0–3.0) 0.661
Polymorphonuclear cells 0 (0–3.0) 0 (0–2.5) 0.126
CD68 lining 2.0 (0–3.0) 2.0 (0–2.0) 0.478
CD68 sublining 1.5 (0–3.0) 1.0 (0–3.0) 0.845
CD163 lining 2.0 (0–3.0) 2.0 (0.5–3.0) 0.768
CD163 sublining 1.5 (0–2.5) 1 (0–2.5) 0.935
E-selectin blood vessels 0 (0–3.0) 0.25 (0–2.5) 0.594
Intracellular citrullinated proteins 0% 0% 1.000
Mab-12A 0% 0% 1.000
The Mann–Whitney U-test was used to evaluate markers scored semiquantitatively, and the χ2 test was used to evaluate markers scored as present or absent. Results are expressed as median (range), and markers evaluated as present or absent are expressed as percentage of positive results. P < 0.05 was considered statistically significant. PsA, psoriatic arthritis.
==== Refs
Dougados M Hochberg MC Why is the concept of spondyloarthropathies important? Best Pract Res Clin Rheumatol 2002 16 495 505 12406423 10.1016/S1521-6942(02)90253-6
Williamson L Dockerty JL Dalbeth N McNally E Ostlere S Wordsworth BP Clinical assessment of sacroiliitis and HLA-B27 are poor predictors of sacroiliitis diagnosed by magnetic resonance imaging in psoriatic arthritis Rheumatology 2004 43 85 88 13130147 10.1093/rheumatology/keg475
Eastmond CJ Psoriatic arthritis. Genetics and HLA antigens Baillieres Clin Rheumatol 1994 8 263 276 8076387
Schatteman L Mielants H Veys EM Cuvelier C De Vos M Gyselbrecht L Elewaut D Goemaere S Gut inflammation in psoriatic arthritis: a prospective ileocolonoscopic study J Rheumatol 1995 22 680 683 7791163
Queiro-Silva R Torre-Alonso JC Tinture-Eguren T Lopez-Lagunas I A polyarticular onset predicts erosive and deforming disease in psoriatic arthritis Ann Rheum Dis 2003 62 68 70 12480674 10.1136/ard.62.1.68
Veale D Rogers S Fitzgerald O Classification of clinical subsets in psoriatic arthritis Br J Rheumatol 1994 33 133 138 8162477
Kane D Stafford L Bresnihan B FitzGerald O A classification study of clinical subsets in an inception cohort of early psoriatic peripheral arthritis: 'DIP or no DIP revisited' Rheumatology 2003 42 1469 1476 12867577 10.1093/rheumatology/keg445
Veale D Yanni G Rogers S Barnes L Bresnihan B Fitzgerald O Reduced synovial membrane macrophage numbers, ELAM-1 expression, and lining layer hyperplasia in psoriatic arthritis as compared with rheumatoid arthritis Arthritis Rheum 1993 36 893 900 7686370
Ceponis A Konttinen YT Imai S Tamulaitiene M Li TF Xu JW Hietanen J Santavirta S Fassbender HG Synovial lining, endothelial and inflammatory mononuclear cell proliferation in synovial membranes in psoriatic and reactive arthritis: a comparative quantitative morphometric study Br J Rheumatol 1998 37 170 178 9569072 10.1093/rheumatology/37.2.170
Foell D Kane D Bresnihan B Vogl T Nacken W Sorg C Fitzgerald O Roth J Expression of the pro-inflammatory protein S100A12 (EN-RAGE) in rheumatoid and psoriatic arthritis Rheumatology 2003 42 1383 1389 12832707 10.1093/rheumatology/keg385
Dougados M van der Linden S Juhlin R Huitfeldt B Amor B Calin A Cats A Dijkmans B Olivieri I Pasero G The European Spondyloarthropathy Study Group preliminary criteria for the classification of spondyloarthropathy Arthritis Rheum 1991 34 1218 1227 1930310
van der Linden S Valkenburg HA Cats A Evaluation of diagnostic criteria for ankylosing spondylitis. A proposal for modification of the New York criteria Arthritis Rheum 1984 27 361 368 6231933
Arnett FC Edworthy SM Bloch DA McShane DJ Fries JF Cooper NS Healey LA Kaplan SR Liang MH Luthra HS The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis Arthritis Rheum 1988 31 315 324 3358796
Baeten D Demetter P Cuvelier C Van Den Bosch F Kruithof E Van Damme N Verbruggen G Mielants H Veys EM De Keyser F Comparative study of the synovial histology in rheumatoid arthritis, spondyloarthropathy, and osteoarthritis: influence of disease duration and activity Ann Rheum Dis 2000 59 945 953 11087697 10.1136/ard.59.12.945
Baeten D Van den Bosch F Elewaut D Stuer A Veys EM De Keyser F Needle arthroscopy of the knee with synovial biopsy sampling: technical experience in 150 patients Clin Rheumatol 1999 18 434 441 10638766 10.1007/s100670050134
Baeten D Demetter P Cuvelier CA Kruithof E Van Damme N De Vos M Veys EM De Keyser F Macrophages expressing the scavenger receptor CD 163: a link between immune alterations of the gut and synovial inflammation in spondyloarthropathy J Pathol 2002 196 343 350 11857499 10.1002/path.1044
Baeten D Peene I Union A Meheus L Sebbag M Serre G Veys EM De Keyser F Specific presence of intracellular citrullinated proteins in rheumatoid arthritis synovium: relevance to antifilaggrin autoantibodies Arthritis Rheum 2001 44 2255 2262 11665966 10.1002/1529-0131(200110)44:10<2255::AID-ART388>3.0.CO;2-#
Baeten D Steenbakkers PGA Rijnders AMW Boots AM Veys EM De Keyser F Detection of MHC/HC gp-39 complexes in rheumatoid arthritis synovium as a specific and independent histological marker Arthritis Rheum 2004 50 444 451 14872486 10.1002/art.20012
Baeten D Moller HJ Delanghe J Veys EM Moestrup SK De Keyser F Association of CD163+ macrophages and local production of soluble CD163 with lower lymphocyte activation in spondyloarthropathy synovitis Arthritis Rheum 2004 50 1611 1623 15146432 10.1002/art.20174
Baeten D Kruithof E De Rycke L Vandooren B Wyns B Boullart L Hoffman IE Boots AM Veys EM De Keyser F Diagnostic classification of spondylarthropathy and rheumatoid arthritis by synovial histopathology: a prospective study in 154 consecutive patients Arthritis Rheum 2004 50 2931 2941 15457462 10.1002/art.20476
Steenbakkers PG Baeten D Rovers E Veys EM Rijnders AW Meijerink J De Keyser F Boots AM Localization of MHC class II/human cartilage glycoprotein-39 complexes in synovia of rheumatoid arthritis patients using complex-specific monoclonal antibodies J Immunol 2003 170 5719 5727 12759455
Vogl T Propper C Hartmann M Strey A Strupat K van den Bos C Sorg C Roth J S100A12 is expressed exclusively by granulocytes and acts independently from MRP8 and MRP14 J Biol Chem 1999 274 25291 25296 10464253 10.1074/jbc.274.36.25291
Dolhain RJ Tak PP Dijkmans BA De Kuiper P Breedveld FC Miltenburg AM Methotrexate reduces inflammatory cell numbers, expression of monokines and of adhesion molecules in synovial tissue of patients with rheumatoid arthritis Br J Rheumatol 1998 37 502 508 9651076 10.1093/rheumatology/37.5.502
Kraan MC Reece RJ Barg EC Smeets TJ Farnell J Rosenburg R Veale DJ Breedveld FC Emery P Tak PP Modulation of inflammation and metalloproteinase expression in synovial tissue by leflunomide and methotrexate in patients with active rheumatoid arthritis. Findings in a prospective, randomized, double-blind, parallel-design clinical trial in thirty-nine patients at two centers Arthritis Rheum 2000 43 1820 1830 10943872 10.1002/1529-0131(200008)43:8<1820::AID-ANR18>3.0.CO;2-D
Youssef PP Triantafillou S Parker A Coleman M Roberts-Thomson PJ Ahern MJ Smith MD Effects of pulse methylprednisolone on cell adhesion molecules in the synovial membrane in rheumatoid arthritis. Reduced E-selectin and intercellular adhesion molecule 1 expression Arthritis Rheum 1996 39 1970 1979 8961901
Youssef PP Haynes DR Triantafillou S Parker A Gamble JR Roberts-Thomson PJ Ahern MJ Smith MD Effects of pulse methylprednisolone on inflammatory mediators in peripheral blood, synovial fluid, and synovial membrane in rheumatoid arthritis Arthritis Rheum 1997 40 1400 1408 9259419
Kane D Gogarty M O'leary J Silva I Bermingham N Bresnihan B Fitzgerald O Reduction of synovial sublining layer inflammation and proinflammatory cytokine expression in psoriatic arthritis treated with methotrexate Arthritis Rheum 2004 50 3286 3295 15476228 10.1002/art.20518
Konig A Krenn V Gillitzer R Glockner J Janssen E Gohlke F Eulert J Muller-Hermelink HK Inflammatory infiltrate and interleukin-8 expression in the synovium of psoriatic arthritis: an immunohistochemical and mRNA analysis Rheumatol Int 1997 17 159 168 9440147 10.1007/s002960050028
Fraser A Fearon U Reece R Emery P Veale DJ Matrix metalloproteinase 9, apoptosis, and vascular morphology in early arthritis Arthritis Rheum 2001 44 2024 2028 11592363 10.1002/1529-0131(200109)44:9<2024::AID-ART351>3.0.CO;2-K
Fearon U Griosios K Fraser A Reece R Emery P Jones PF Veale DJ Angiopoietins, growth factors, and vascular morphology in early arthritis J Rheumatol 2003 30 260 268 12563678
Ceponis A Hietanen J Tamuliatiene M Partsch G Patiala H Konttinen YT A comparative quantitative morphometric study of cell apoptosis in synovial membranes in psoriatic, reactive and rheumatoid arthritis Rheumatology 1999 38 431 440 10371281 10.1093/rheumatology/38.5.431
Danning CL Illei GG Hitchon C Greer MR Boumpas DT McInnes IB Macrophage-derived cytokine and nuclear factor kappaB p65 expression in synovial membrane and skin of patients with psoriatic arthritis Arthritis Rheum 2000 43 1244 1256 10857783 10.1002/1529-0131(200006)43:6<1244::AID-ANR7>3.0.CO;2-2
Ritchlin C Haas-Smith SA Hicks D Cappuccio J Osterland CK Looney RJ Patterns of cytokine production in psoriatic synovium J Rheumatol 1998 25 1544 1552 9712099
Salvador G Sanmarti R Garcia-Peiro A Rodriguez-Cros JR Muñez-Gomez J Cañete JD p53 expression in rheumatoid and psoriatic arthritis synovial tissue and association with joint damage Ann Rheum Dis 2005 64 183 187 15647425 10.1136/ard.2004.024430
Vandooren B Kruithof E Yu DT Rihl M Gu J De Rycke L Van Den Bosch F Veys EM De Keyser F Baeten D Involvement of matrix metalloproteinases and their inhibitors in peripheral synovitis and down-regulation by tumor necrosis factor α blockade in spondylarthropathy Arthritis Rheum 2004 50 2942 2953 15457463 10.1002/art.20477
De Rycke L Baeten D Foell D Kruithof E Veys EM Roth J Differential expression and response to anti-TNFalpha treatment of specific macrophage subsets in rheumatoid arthritis and spondyloarthropathy: MRP8 and MRP14 reflect synovial inflammation J Pathol
Jones SM Dixey J Hall ND McHugh NJ Expression of the cutaneous lymphocyte antigen and its counter-receptor E-selectin in the skin and joints of patients with psoriatic arthritis Br J Rheumatol 1997 36 748 757 9255108 10.1093/rheumatology/36.7.748
Kane D Roth J Frosch M Vogl T Bresnihan B FitzGerald O Increased perivascular synovial membrane expression of myeloid-related proteins in psoriatic arthritis Arthritis Rheum 2003 48 1676 1685 12794836 10.1002/art.10988
Youssef P Roth J Frosch M Costello P Fitzgerald O Sorg C Bresnihan B Expression of myeloid related proteins (MRP) 8 and 14 and the MRP8/14 heterodimer in rheumatoid arthritis synovial membrane J Rheumatol 1999 26 2523 2528 10606357
Frosch M Strey A Vogl T Wulffraat NM Kuis W Sunderkotter C Harms E Sorg C Roth J Myeloid-related proteins 8 and 14 are specifically secreted during interaction of phagocytes and activated endothelium and are useful markers for monitoring disease activity in pauciarticular-onset juvenile rheumatoid arthritis Arthritis Rheum 2000 43 628 637 10728757 10.1002/1529-0131(200003)43:3<628::AID-ANR20>3.0.CO;2-X
Baeten D Kruithof E Van den Bosch F Demetter P Van Damme N Cuvelier C De Vos M Mielants H Veys EM De Keyser F Immunomodulatory effects of anti-tumor necrosis factor alpha therapy on synovium in spondyloarthropathy: histologic findings in eight patients from an open-label study Arthritis Rheum 2001 44 186 195 11212159 10.1002/1529-0131(200101)44:1<186::AID-ANR25>3.0.CO;2-B
Kruithof E Baeten D Van den Bosch F Mielants H Veys EM De Keyser F Histological evidence that infliximab treatment leads to down-regulation of inflammation and tissue remodelling of the synovial membrane in spondyloarthropathy Ann Rheum Dis 2004, 23 September
Kraan M Haringman J Ahern M Breedveld F Smith M Tak PP Quantification of the cell infiltrate in synovial tissue by digital image analysis Rheumatology 2000 39 43 49 10662872 10.1093/rheumatology/39.1.43
Ritchlin CT Haas-Smith SA Li P Hicks DG Schwarz EM Mechanisms of TNF-alpha- and RANKL-mediated osteoclastogenesis and bone resorption in psoriatic arthritis J Clin Invest 2003 111 821 831 12639988 10.1172/JCI200316069
| 15899044 | PMC1174942 | CC BY | 2021-01-04 16:02:35 | no | Arthritis Res Ther. 2005 Mar 3; 7(3):R569-R580 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1698 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar16991589903310.1186/ar1699Research ArticleThe P2X7 receptor is a candidate product of murine and human lupus susceptibility loci: a hypothesis and comparison of murine allelic products Elliott James I [email protected] John H [email protected] Christopher F [email protected] MRC Clinical Sciences Centre, Faculty of Medicine, Imperial College, Hammersmith Hospital Campus, London, UK2005 21 2 2005 7 3 R468 R475 20 10 2004 30 11 2004 18 1 2005 21 1 2005 Copyright © 2005 Elliott et al., licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Systemic lupus erythematosus and its murine equivalent, modelled in the New Zealand Black and New Zealand White (NZB × NZW)F1 hybrid strain, are polygenic inflammatory diseases, probably reflecting an autoimmune response to debris from cells undergoing programmed cell death. Several human and murine loci contributing to disease have been defined. The present study asks whether the proinflammatory purinergic receptor P2X7, an initiator of a form of programmed cell death known as aponecrosis, is a candidate product of murine and human lupus susceptibility loci. One such locus in (NZB × NZW)F1 mice is lbw3, which is situated at the distal end of NZW chromosome 5. We first assess whether NZB mice and NZW mice carry distinct alleles of the P2RX7 gene as expressed by common laboratory strains, which differ in sensitivity to ATP stimulation. We then compare the responses of NZB lymphocytes, NZW lymphocytes and (NZB × NZW)F1 lymphocytes to P2X7 stimulation. NZB and NZW parental strains express the distinct P2X7-L and P2X7-P alleles of P2RX7, respectively, while lymphocytes from these and (NZB × NZW)F1 mice differ markedly in their responses to P2X7 receptor stimulation. NZB mice and NZW mice express functionally distinct alleles of the proinflammatory receptor, P2X7. We show that current mapping suggests that murine and human P2RX7 receptor genes lie within lupus susceptibility loci lbw3 and SLEB4, and we argue that these encode a product with the functional characteristics consistent with a role in lupus. Furthermore, we argue that aponecrosis as induced by P2X7 is a cell death mechanism with characteristics that potentially have particular relevance to disease pathogenesis.
==== Body
Introduction
Systemic lupus erythematosus (SLE) is a polygenic disease, although the genes contributing towards the disease are unknown. Several human susceptibility loci have been identified, with eight of the strongest candidates mapping to 1q23, 1q25-31, 1q41-42, 2q35-37, 4p16-15.2, 6p11-21, 12q24 and 16q12 [1]. Of the murine models, the New Zealand Black and New Zealand White (NZB × NZW)F1 hybrid strain is widely studied due to its similarity to human disease and its female preponderance. As with human SLE, the disorder of (NZB × NZW)F1 mice is polygenic with a contribution from both parents. In a study of (NZB × NZW)F2 mice, eight susceptibility loci were identified [2]. In the case of the locus lbw3, at the distal region of chromosome 5, homozygosity for the NZW-derived locus was associated with increased mortality at 12 months. Although originally mapped to 88 cM on murine chromosome 5 [2], more recent data locate the microsatellite used to define lbw3 at 81 cM (discussed later).
We have studied the properties of the proinflammatory purinergic receptor P2X7, encoded by a gene within the human SLE locus SLEB4 [3] at 12q24 (Ensembl Genome Browser: ) and by the murine lbw3 region (Ensembl Genome Browser: ), and discuss its potential role in disease. The P2X7 receptor belongs to a family of ion channels gated by extracellular ATP, but unlike other P2X receptors it is largely restricted to haematopoietic cells. The P2X7 receptor has been proposed to play a role in a variety of immune functions including the secretion of leaderless cytokines and the shedding of the lymphocyte homing receptor CD62L [4]. As P2X7-deficient mice exhibit resistance to antibody-induced arthritis and impaired CD62L shedding and IL-1β secretion [5], stimulation of this receptor is proinflammatory – suggesting a potential role in autoimmune disease.
In this respect, SLE is of particular interest. Not only is SLE an inflammatory disorder, but it probably reflects, at least in part, an immune response to debris of cells undergoing programmed cell death (PCD). As P2X7 stimulation is proinflammatory and induces PCD, functional polymorphisms in this gene would be predicted to affect lupus susceptibility. Moreover, PCD stimulated through the P2X7 receptor belongs to a category that bears many of the hallmarks of 'classic' caspase-dependent apoptosis, but also to other categories such as cytoplasmic vacuolization more often associated with necrosis. Such cell death has sometimes been termed 'aponecrosis' [6]. Whereas removal of 'classic' apoptotic cells is believed to be immunologically silent, necrotic cell debris is proinflammatory [7]. The effect of intermediate forms of PCD such as aponecrosis, for which clearance mechanisms have not been defined, is unknown, yet such material potentially plays a significant role in the pathogenesis of SLE (discussed later). Finally, P2X7 stimulation results in rapid translocation of phosphatidylserine (PS) from the inner to the outer leaflet of the plasma membrane, which is reversible if stimulation is brief (and thus independent of cell death). As PS and associated proteins are major targets of autoantibodies in SLE [8], cells stimulated via the P2X7 receptor may be a significant source of autoantigen in this disease.
An allelic variation (P451L) of the cytoplasmic domain of the P2X7 receptor in commonly used mouse strains is associated with significant differences in its sensitivity to the ATP ligand [9]. These allelic forms with proline (P2X7-P) and leucine (P2X7-L) at position 451 confer high sensitivity and low sensitivity to stimulation by ATP, respectively. While the NZW strain has been shown to express the more responsive allele of P2X7 (P2X-P) [9], that expressed by the NZB strain is unknown. We show in the present article that NZB mice and NZW mice express different alleles of the proinflammatory receptor P2X7, and furthermore that NZW lymphocytes are markedly more responsive to P2X7 stimulation than those from NZB mice. Lymphocytes from (NZB × NZW)F1 mice exhibited intermediate sensitivities to P2X7-induced PS translocation and to PCD, but were as sensitive to induction of CD62L shedding as those from NZW mice, indicating a comparatively complex phenotypic penetration. The results indicate that P2X7 is a strong candidate for being the product of the murine lbw3 locus. As the human P2RX7 gene maps close to SLEB4, we hypothesize that similar polymorphisms may also contribute towards human disease.
Methods
Mice
Male mice were purchased from Harlan-Olac (Bicester, UK) and used at between 8 and 14 weeks. Institute guidelines for care of laboratory animals were followed. All studies received ethical review approval.
P2X7 PCRs
PCR amplification of the NZB mouse genomic sequence encompassing the T1352C polymorphism [9] was performed using the forward and reverse primers CCTGTCTAGGCTGTCCCTAT and GCTTATGGAAGAGCTTGGAG for 30 cycles. PCR products were cloned using the TOPO TA cloning system (Invitrogen, Paisley, UK). Forty-three independent clones were sequenced using an ABI PRISM Big Dye terminator ready reaction kit (Applied Biosystems, Warrington, UK) and were analysed on a 3700 DNA Analyser (Applied Biosystems). Nucleotide and amino acid substitutions were numbered using the cDNA sequence accession number NM-011027.
Reagents
Matrix metalloproteinase inhibitor III was from Calbiochem (Nottingham, UK). Other reagents were from Sigma (Poole, UK), unless stated. Diluents had no effect in any assay used.
Flow cytometry
Mesenteric lymphocyte cells (107/ml) in phenol red-free DMEM were stained with a combination of CD4APC, CD4CYCHROME, CD4PE, CD8APC, CD8CYCHROME, CD8PE, CD8FITC and CD62LFITC antibodies (Becton Dickinson, Oxford, UK) as indicated, washed and resuspended in DMEM. Cells were equilibrated with annexin VFITC or annexin VCY5 (Becton Dickinson) to assess cell surface PS exposure, or with propidium iodide for 4 min to assess cell death, and were analysed by flow cytometry on a FACScalibur machine using CellQuest (Becton Dickinson) or Flowjo (Tree Star, Ashland, OR, USA) software. Baseline fluorescence was established for approximately 1 min prior to addition of 150 μM (unless otherwise stated) 2'-3'-O-(4-benzoylbenzoyl)-adenosine 5'-triphosphate (BzATP). Cells were monitored for PS exposure or CD62L shedding continuously in real time for up to 9 min or were monitored for uptake of propidium iodide, as indicated. All results are representative of at least three independent experiments.
IL-1β secretion
Spleens from NZW mice and NZB mice were disaggregated and erythrocytes were lysed (Puregene RBC lysis solution; Gentra Ltd, Minneapolis, MN, USA). Splenocytes (5 × 106/ml) were then resuspended in DMEM supplemented with 10% FCS (Helena Biosciences, Sunderland, UK), and were stimulated with 2 μg/ml lipopolysaccharide. After 6 hours at 37°C the medium was removed and replaced with DMEM. Cells were then incubated at 37°C for 30 min with BzATP added as indicated. The supernatant was then collected and the cells and particulate matter were removed by centrifugation. Supernatants were then frozen at -20°C. IL-1β was quantified by ELISA (Quantikine mouse IL-1β kit; R&D Systems, Minneapolis, MN, USA) in accordance with the manufacturer's instructions. Statistical significance was measured by Student's t test.
Results
Real-time comparison of P2X7-stimulated PS translocation and CD62L shedding by NZW, (NZB × NZW)F1 and NZB lymphocytes
We initially confirmed that NZW mice are homozygous for the P2X-P allele of P2RX7 [9] associated with high sensitivity to stimulation, and we showed that NZB mice are homozygous for the low sensitivity allele P2X-L (data not shown). These forms differ at a single amino acid (451). P2X7 activation, stimulated by BzATP (Fig. 1), results in rapid externalization of PS and shedding of CD62L. We therefore developed real-time flow cytometric assays to directly compare the responses to P2X7 simulation of NZB lymphocytes, NZW lymphocytes and (NZB × NZW)F1 lymphocytes.
PS translocation
PS is largely confined to the inner leaflet of the plasma membrane in healthy cells. Loss of lipid asymmetry, as evidenced by surface exposure of PS occurring prior to membrane breakdown, is generally assumed to be a marker of PCD. To enable the direct comparison of responses of cells from NZB mice, NZW mice and (NZB × NZW)F1 mice in a single tube, lymphocytes from these three strains were stained with anti-CD4CYCHROME, anti-CD4PE and anti-CD4APC, respectively, mixed and equilibrated with annexin VFITC (to detect exposed PS). Labelled cells could thus subsequently be distinguished by flow cytometric gating. The rates of P2X7-stimulated PS exposure by lymphocyte subsets derived from the three mouse strains were directly compared by real-time flow cytometry (Fig. 1). Baseline fluorescence was established, and cells were stimulated with the P2X7 agonist BzATP at the time indicated. The order of responsiveness to P2X7 stimulation, as evidenced by the percentage of cells translocating PS, was consistently NZW > (NZB × NZW)F1 > NZB. Therefore, consistent with the greater sensitivity of P2X7-P, lymphocytes from NZW mice show greater sensitivity to P2X7 stimulation, and show co-dominance with the NZB-derived allele (P2X7-L) in F1 hybrid mice.
CD62L shedding
Shedding of CD62L from T cells is a key event in lymphocyte migration to inflammatory sites [10] and is known to be induced by P2X7 stimulation. Lymphocytes from NZB mice, NZW mice and (NZB × NZW)F1 mice were differentially stained as already stated but were labelled with FITC-conjugated anti-CD62L in place of annexin VFITC to allow direct comparison of the rate of CD62L shedding in a single tube. P2X7-stimulated CD62L shedding was apparent as a decrease in fluorescence in the FL-1 channel. While the rates of P2X7-stimulated CD62L shedding were high and low in NZW lymphocytes and NZB lymphocytes, respectively (Fig. 2), interestingly the rate of CD62L shedding by (NZB × NZW)F1 lymphocytes was indistinguishable from that by NZW cells. The high NZW response of P2X7 thus appears dominant with respect to CD62L shedding, indicating that factors downstream of P2X7 stimulation contribute to this phenotype. That loss of CD62L reflects shedding and not decreased cell surface expression through other mechanisms is evidenced by its blockade by an inhibitor of matrix metalloproteinase [11] (Fig. 2c).
P2X7-induced secretion of IL-1β
Stimulation of the P2X7 receptor on lipopolysaccharide-stimulated monocytes and macrophages promotes secretion of the proinflammatory cytokine IL-1β [5,12], which may therefore be expected to differ between mice bearing P2X7-L or P2X7-P receptors. Indeed IL-1β secretion by NZW splenocytes stimulated in vitro with lipopolysaccharide and BzATP exceeded that by cells from NZB mice (P < 0.05 at 50, 100 and 150 μM BzATP; Fig. 3). Elevated IL-1β secretion by NZW splenocytes was apparent even in the absence of BzATP (although slightly below statistical significance), suggesting that inadvertent stimulation of the high (but not low) sensitivity P2X7 receptor may have occurred through cell death and the consequent release of ATP during cell preparation (spleen disaggregation and erythrocyte lysis). In one NZW splenocyte preparation exhibiting particularly high IL-1β secretion, cells were refractory to further stimulation of the P2X7 receptor in vitro.
P2X7-induced PCD of NZW and NZB lymphocytes
Several lines of evidence indicate that lupus reflects an autoimmune response to debris from cells undergoing PCD. Although stimulation of P2X7 results in rapid PS translocation, the effects are reversible if exposure to the agonist is brief [4]. Only prolonged treatment with agonist results in PCD. Translocation of PS following P2X7 activation cannot therefore be used as a direct measure of irreversible commitment to PCD. To measure PCD following prolonged P2X7 stimulation, we therefore compared the rate of terminal membrane breakdown (indicated by propidium iodide uptake) following BzATP treatment of NZW lymphocytes, NZB lymphocytes and (NZB × NZW)F1 lymphocytes (Fig. 4). P2X7 stimulation resulted in significant PCD in all populations tested, with the order of sensitivity NZW > (NZB × NZW)F1 > NZB, consistent with the high responder status of the NZW cells and the dominance of the NZW-derived P2RX7 allele in this response.
Discussion
Stimulation of the proinflammatory haematopoietic P2X7 receptor [5] results in IL-1β secretion, in high rates of PCD [4] and in CD62L shedding [13], each of which is associated with human SLE [14-16]. The P2X7 receptor therefore has the characteristics of a candidate lupus susceptibility gene product. Moreover, the gene encoding human P2X7 is located within a region (12q24; Ensembl Genome Browser: ) recently identified and confirmed in Hispanic and European-American Families as a lupus susceptibility locus, designated SLEB4 [3].
A polymorphism in the cytoplasmic domain of the P2X7 receptor of common mouse strains is associated with differential responsiveness [9]. While most strains, including NZW mice [9], possess proline in amino acid position 451, we showed that NZB mice express P2X7 with lysine at this position and that the variant confers markedly decreased sensitivity to P2X7 stimulation. Notably, the murine P2RX7 gene is encoded by a gene on chromosome 5 within a region designated lbw3 due to the identification of a NZW-derived susceptibility locus conferring increased mortality at 12 months [2]. While susceptibility regions are broad, the microsatellite marker D5Mit101 (defining lbw3 [2]) was in the original study mapped to 88 cM on chromosome 5, which may have discouraged identification of P2RX7 as a candidate susceptibility gene. Current mapping data, however, show this marker located at 81 cM – Mouse Genome Informatics and Ensembl Genome Browser . As the marker D5Mit118 that is adjacent to P2RX7 is located at 67 cM – Ensembl Genome Browser and Mouse Genome Informatics – the two are approximately 14 cM (or 19 Mb) apart (120 Mb versus 139 Mb), easily within the 20 cM distance used by Kono and colleagues [2] to define coverage by markers in their study.
Although gene polymorphisms may have unpredicted effects, other than P2RX7 there appear to be few candidate susceptibility genes (based on lymphoid expression and protein activity) in the region described by lbw3. However, other candidates might include those encoding: lnk, an adaptor protein in T-cell signalling (65.0 cM) [17]; P2X4, a purinergic receptor (65.0 cM) whose activity is assumed primarily to be neuronal, but which is also expressed (at least at the level of mRNA) in lymphocytes [18]; shp2, a tyrosine phosphatase [19] (~66 cM [118.6 Mb]); and FLT3 (CD135, 82.0 cM), a tyrosine kinase expressed in haemaotopoietic cells [20]. None of these has been reported to be polymorphic between NZB mice and NZW mice.
It is widely thought that lupus reflects an autoimmune response to cells undergoing PCD. Aberrant responses to such debris may reflect qualitative or quantitative abnormalities; for example, if its handling is defective and/or following exposure to increased levels of 'apoptotic' material. Both have been reported to contribute to human SLE [15,16]. That prolonged stimulation of the P2X7 receptor induces PCD is therefore of particular note. However, multiple pathways of PCD exist. While 'apoptosis' and 'PCD' are frequently used as synonyms, 'apoptosis' is often used to imply caspase-dependent cell death. Nevertheless, caspase involvement is not a good indicator of the physiologic importance, or 'programming', of a cell death pathway, and consequently classic 'apoptosis' may describe one end of a continuum of active PCD mechanisms [21]. Hence, in principle, a defect in one of many PCD pathways, rather than increased susceptibility to PCD per se, may be sufficient to increased the burden of cellular debris and hence the susceptibility to lupus. Indeed, that SLE may reflect an autoimmune response to debris from 'apoptotic' cells, despite clearance of such material being thought generally immunologically silent [7], has been a conundrum.
To reconcile these findings it has been suggested that, in SLE, mechanisms for removing apoptotic debris are overloaded, with remaining cells undergoing secondary necrosis, and/or that apoptotic cells have some immunostimulatory properties [7]. We suggest the additional possibility that different forms of PCD may give rise to debris with different degrees of immunogenicity. It is therefore necessary to dissect distinct PCD pathways to assess the potential effects that defects have on the disease process. It is attractive to speculate that P2X7-induced aponecrotic debris, perhaps due to the catastrophic nature of its generation or the apparent differences in cell dismantling, may be more necrotic than apoptotic in character and thus be immunostimulatory. Such material may either promote responses to surrounding 'apoptotic' cells and/or directly stimulate autoimmune responses to itself (if lupus autoantigens are appropriately packaged in P2X7-induced PCD). P2X7 receptor-induced PCD is therefore potentially a source of lupus autoantigens or may represent a catastrophic form of cell death that overwhelms the host's ability to clear such material.
We therefore suggest the following involvement of the P2X7 receptor in SLE. ATP exists at very high concentrations in normal cells (5–10 mM), and is released upon cell death before its rapid breakdown by ATPases. Consequently, extracellular concentrations of ATP, although normally low, are transiently increased at sites of tissue damage. Stimulation of P2X7 occurs at sufficient concentrations of ATP, resulting in secretion of IL-1β and in CD62L shedding within minutes. P2X7 stimulation thus acts to promote the inflammatory response. The resulting lymphoid infiltration leads to additional lymphocyte-mediated cell death, and to consequent ATP release, exacerbating the P2X7-driven inflammatory cycle. Indeed, given sufficient tissue damage, prolonged stimulation of P2X7 itself induces PCD, further adding to the cycle of ATP release and destruction. Release of autoantigens within P2X7-stimulated aponecrotic debris may also contribute to a breakdown in self-tolerance and initiation of autoimmunity.
While one must make the proviso that little is known at the moment about the level of ATP released at sites of tissue damage, its rate of decay and how these may vary between pathological conditions including SLE, we suggest it is reasonable to hypothesize that polymorphisms within P2X7 can influence the pathogenesis of lupus. Importantly, there are a number of polymorphisms within P2X7 that affect its activity. The Ile-568 to Asn [22], Arg307 to Gln [23], and Glu496 to Ala [24] polymorphisms therefore all result in reduced function of human P2X7, and might each be hypothesized to result in decreased severity of SLE.
Conclusions
In summary, we have shown that polymorphism of the P2X7 receptor between NZW and NZB strains is associated with marked differences in P2X7-stimulated proinflammatory responses, consistent with high responsiveness and low responsiveness previously reported for the two alleles. We also show that current genetic mapping indicates that the P2RX7 gene is located within the region defined as lbw3 and is a therefore a strong candidate for being the product of this lupus susceptibility locus. Furthermore, as the human gene maps very close to SLEB4, we hypothesize that polymorphisms within P2RX7 may also contribute to human disease. Stimulation of the P2X7 receptor is proinflammatory and induces a form of cell death known as aponecrosis, which exhibits several characteristics of apoptosis. We therefore suggest that the P2X7 receptor and gene have the functional and positional characteristics suggestive of a role in the pathogenesis in SLE, and that the potential of the cell death mechanism aponecrosis to contribute to disease warrants study.
Abbreviations
BzATP = 2'-3'-O-(4-benzoylbenzoyl)-adenosine 5'-triphosphate; DMEM = Dulbecco's modified Eagle's medium; ELISA = enzyme-linked immunosorbent assay; FCS = foetal calf serum; FITC = fluorescein isothiocyanate; IL = interleukin; NZB = New Zealand Black; NZW = New Zealand White; PCD = programmed cell death; PCR = polymerase chain reaction; PS = phosphatidylserine; SLE = systemic lupus erythematosus.
Competing interests
The author(s) declare there are no competing interests.
Authors' contributions
JIE conceived the study, carried out the flow cytometric and IL-1β secretion experiments, and wrote the first draft of the manuscript. JHM designed allele-specific primers and typed the P2RX7 genes of NZB mice and NZW mice, and contributed to drafting of the manuscript. CFH contributed to the design of experiments and drafting of the manuscript. All authors read and approved the final manuscript.
Acknowledgements
This work was supported by the Medical Research Council of Great Britain. The authors would like to thank Dr T Vyse for helpful discussions.
Figures and Tables
Figure 1 P2X7-stimulated phosphatidylserine (PS) exposure on lymphocytes. P2X7-dependent exposure of PS on New Zealand Black (NZB) lymphocytes, New Zealand White (NZW) lymphocytes and (NZB × NZW)F1 (NZB/W) lymphocytes. To enable the direct comparison of responses of cells from NZB mice, NZW mice and (NZB × NZW)F1 mice in a single tube, lymphocytes from these strains were stained with anti-CD4CYCHROME, anti-CD4PE and anti-CD4APC, respectively, mixed and equilibrated with annexin VFITC. Thus labelled, cells could subsequently be distinguished by flow cytometric gating. Cells were stimulated with the P2X7 agonist 2'-3'-O-(4-benzoylbenzoyl)-adenosine 5'-triphosphate (BzATP) at the time indicated by the arrow in (a). (a) Density plots of the rate of extracellular PS exposure in each cell population, as indicated by increased binding of annexin VFITC. (b) Corresponding percentage of cells bearing exposed PS in each population at a single timepoint (indicated by boxes in (a)).
Figure 2 P2X7-stimulated shedding of CD62L by lymphocytes. To enable the direct comparison of responses of cells from New Zealand Black (NZB) mice, New Zealand White (NZW) mice and (NZB × NZW)F1 (NZB/W) mice in a single tube, lymphocytes from these strains were stained with anti-CD4PE, anti-CD4CYCHROME and anti-CD4APC, respectively, mixed and stained with anti-CD62LFITC. Thus labelled, cells could subsequently be distinguished by flow cytometric gating. Cells were stimulated with the P2X7 agonist 2'-3'-O-(4-benzoylbenzoyl)-adenosine 5'-triphosphate (BzATP) at the time indicated by the arrow in (a). (a) Density plots of the rate of CD62L shedding in each cell population, as indicated by decreased binding of anti-CD62LFITC. (b) Corresponding levels of cell surface CD62L in each population (NZB, red line; NZW, green line; NZB/W, black line) immediately preceding P2X7 stimulation (indicated by left-hand gates in (a)) or 7 min after P2X7 stimulation (indicated by right-hand gates in (a)). (c) Effect of a broad inhibitor of metalloproteinases on loss of CD62L. Lymphocytes from NZW mice were stained with anti-CD4CYCHROME and anti-CD62LPE, and the rate of loss of CD62L was assessed by flow cytometry. Cells were stimulated with BzATP in the presence or absence of 10 μM metalloproteinase inhibitor at the time indicated by an arrow. Shedding of CD62L is indicated by decreased binding of anti-CD62LPE. MMP, matrix metallopoteinase.
Figure 3 P2X7-stimulated secretion of IL-1β. Splenocytes were primed in vitro with lipopolysaccharide and were then stimulated with 2'-3'-O-(4-benzoylbenzoyl)-adenosine 5'-triphosphate (BzATP) as indicated. The graph shows IL-1β secretion by cells from New Zealand Black mice (open squares) and New Zealand White mice (filled squares). Each line represents IL-1β secretion by cells from a single mouse.
Figure 4 P2X7-stimulated lymphocyte programmed cell death (PCD). Lymphocytes from New Zealand Black (NZB) mice, New Zealand White (NZW) mice and (NZB × NZW)F1 (NZB/W) mice were stained with anti-CD4APC and anti-CD8FITC, and were equilibrated with propidium iodide (PI). Panels show the mean percentage (± standard deviation, n = 5) of dead cells (those taking up PI) as assessed by flow cytometry before (t = 0), and at 15-min intervals subsequent to, stimulation of the P2X7 receptor with 2'-3'-O-(4-benzoylbenzoyl)-adenosine 5'-triphosphate. (a) CD4+ cells, and (b) CD8+ cells. NZB, open squares; NZW, open circles; NZB/W, open diamonds.
==== Refs
Tsao BP Update on human systemic lupus erythematosus genetics Curr Opin Rheumatol 2004 16 513 521 15314487 10.1097/01.bor.0000132648.62680.81
Kono DH Burlingame RW Owens DG Kuramochi A Balderas RS Balomenos D Theofilopoulos AN Lupus susceptibility loci in New Zealand mice Proc Natl Acad Sci 1994 91 10168 10172 7937857
Nath SK Quintero-Del-Rio AI Kilpatrick J Feo L Ballesteros M Harley JB Linkage at 12q24 with systemic lupus erythematosus (SLE) is established and confirmed in Hispanic and European American families Am J Hum Genet 2004 74 73 82 14658095 10.1086/380913
MacKenzie A Wilson HL Kiss-Toth E Dower SK North RA Surprenant A Rapid secretion of interleukin-1beta by microvesicle shedding Immunity 2001 15 825 835 11728343 10.1016/S1074-7613(01)00229-1
Labasi JM Petrushova N Donovan C McCurdy S Lira P Payette MM Brissette W Wicks JR Audoly L Gabel CA Absence of the P2X7 receptor alters leukocyte function and attenuates an inflammatory response J Immunol 2002 168 6436 6445 12055263
Formigli L Papucci L Tani A Schiavone N Tempestini A Orlandini GE Capaccioli S Orlandini SZ Aponecrosis: morphological and biochemical exploration of a syncretic process of cell death sharing apoptosis and necrosis J Cell Physiol 2000 182 41 49 10567915 10.1002/(SICI)1097-4652(200001)182:1<41::AID-JCP5>3.0.CO;2-7
Savill J Dransfield I Gregory C Haslett C A blast from the past: clearance of apoptotic cells regulates immune responses Nat Rev Immunol 2002 2 965 975 12461569 10.1038/nri957
McClain MT Arbuckle MR Heinlen LD Dennis GJ Roebuck J Rubertone MV Harley JB James JA The prevalence, onset, and clinical significance of antiphospholipid antibodies prior to diagnosis of systemic lupus erythematosus Arthritis Rheum 2004 50 1226 1232 15077305 10.1002/art.20120
Adriouch S Dox C Welge V Seman M Koch-Nolte F Haag F Cutting edge: a natural P451L mutation in the cytoplasmic domain impairs the function of the mouse P2X7 receptor J Immunol 2002 169 4108 4112 12370338
Gallatin WM Weissman IL Butcher EC A cell-surface molecule involved in organ-specific homing of lymphocytes Nature 1983 304 30 34 6866086 10.1038/304030a0
Preece G Murphy G Ager A Metalloproteinase-mediated regulation of L-selectin levels on leucocytes J Biol Chem 1996 271 11634 11640 8662605 10.1074/jbc.271.20.11634
Grahames CB Michel AD Chessell IP Humphrey PP Pharmacological characterization of ATP- and LPS-induced IL-1beta release in human monocytes Br J Pharmacol 1999 127 1915 1921 10482924 10.1038/sj.bjp.0702732
Gu B Bendall LJ Wiley JS Adenosine triphosphate-induced shedding of CD23 and L-selectin (CD62L) from lymphocytes is mediated by the same receptor but different metalloproteases Blood 1998 92 946 951 9680363
Sfikakis PP Charalambopoulos D Vaiopoulos G Mavrikakis M Circulating P- and L-selectin and T-lymphocyte activation and patients with autoimmune rheumatic diseases Clin Rheumatol 1999 18 28 32 10088945 10.1007/s100670050047
Ren Y Tang J Mok MY Chan AW Wu A Lau CS Increased apoptotic neutrophils and macrophages and impaired macrophage phagocytic clearance of apoptotic neutrophils in systemic lupus erythematosus Arthritis Rheum 2003 48 2888 2897 14558095 10.1002/art.11237
Emlen W Niebur J Kadera R Accelerated in vitro apoptosis of lymphocytes from patients with systemic lupus erythematosus J Immunol 1994 152 3685 3692 8144943
Brown SB Clarke MC Magowan L Sanderson H Savill J Constitutive death of platelets leading to scavenger receptor-mediated phagocytosis. A caspase-independent cell clearance program J Biol Chem 2000 275 5987 5996 10681593 10.1074/jbc.275.8.5987
Wang L Jacobsen SE Bengtsson A Erlinge D P2 receptor mRNA expression profiles in human lymphocytes, monocytes and CD34+stem and progenitor cells BMC Immunol 2004 5 16 15291969 10.1186/1471-2172-5-16
Gjorloff-Wingren A Saxena M Han S Wang X Alonso A Renedo M Oh P Williams S Schnitzer J Mustelin T Subcellular localization of intracellular protein tyrosine phosphatases in T cells Eur J Immunol 2000 30 2412 2421 10940933 10.1002/1521-4141(2000)30:8<2412::AID-IMMU2412>3.0.CO;2-J
Masten BJ Olson GK Kusewitt DF Lipscomb MF Flt3 ligand preferentially increases the number of functionally active myeloid dendritic cells in the lungs of mice J Immunol 2004 172 4077 4083 15034019
Jaattela M Tschopp J Caspase-independent cell death in T lymphocytes Nat Immunol 2003 4 416 423 12719731 10.1038/ni0503-416
Wiley JS Dao-Ung LP Li C Shemon AN Gu BJ Smart ML Fuller SJ Barden JA Petrou S Sluyter R An Ile-568 to Asn polymorphism prevents normal trafficking and function of the human P2X7 receptor J Biol Chem 2003 278 17108 17113 12586825 10.1074/jbc.M212759200
Gu BJ Sluyter R Skarratt KK Shemon AN Dao-Ung LP Fuller SJ Barden JA Clarke AL Petrou S Wiley JS An Arg307 to Gln polymorphism within the ATP-binding site causes loss of function of the human P2X7 receptor J Biol Chem 2004 279 31287 31295 15123679 10.1074/jbc.M313902200
Sluyter R Dalitz JG Wiley JS P2X7 receptor polymorphism impairs extracellular adenosine 5' -triphosphate-induced interleukin-18 release from human monocytes Genes Immun 2004 5 588 591 15306849 10.1038/sj.gene.6364127
| 15899033 | PMC1174943 | CC BY | 2021-01-04 16:02:36 | no | Arthritis Res Ther. 2005 Feb 21; 7(3):R468-R475 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1699 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17001589903510.1186/ar1700Research ArticleA web tool for finding gene candidates associated with experimentally induced arthritis in the rat Andersson Lars [email protected] Greta [email protected] Per [email protected]åhl Fredrik [email protected] Department of Cell and Molecular Biology – Genetics, Goteborg University, Sweden2 School of Health Sciences, University College of Borås, Borås, Sweden2005 18 2 2005 7 3 R485 R492 2 12 2004 4 1 2005 20 1 2005 24 1 2005 Copyright © 2005 Andersson et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Rat models are frequently used for finding genes contributing to the arthritis phenotype. In most studies, however, limitations in the number of animals result in a low resolution. As a result, the linkage between the autoimmune experimental arthritis phenotype and the genomic region, that is, the quantitative trait locus, can cover several hundred genes. The purpose of this work was to facilitate the search for candidate genes in such regions by introducing a web tool called Candidate Gene Capture (CGC) that takes advantage of free text data on gene function. The CGC tool was developed by combining genomic regions in the rat, associated with the autoimmune experimental arthritis phenotype, with rat/human gene homology data, and with descriptions of phenotypic gene effects and selected keywords. Each keyword was assigned a value, which was used for ranking genes based on their description of phenotypic gene effects. The application was implemented as a web-based tool and made public at . The CGC application ranks gene candidates for 37 rat genomic regions associated with autoimmune experimental arthritis phenotypes. To evaluate the CGC tool, the gene ranking in four regions was compared with an independent manual evaluation. In these sample tests, there was a full agreement between the manual ranking and the CGC ranking for the four highest-ranked genes in each test, except for one single gene. This indicates that the CGC tool creates a ranking very similar to that made by human inspection. The exceptional gene, which was ranked as a gene candidate by the CGC tool but not in the manual evaluation, was found to be closely associated with rheumatoid arthritis in additional literature studies. Genes ranked by the CGC tools as less likely gene candidates, as well as genes ranked low, were generally rated in a similar manner to those done manually. Thus, to find genes contributing to experimentally induced arthritis, we consider the CGC application to be a helpful tool in facilitating the evaluation of large amounts of textual information.
==== Body
Introduction
Rheumatoid arthritis (RA) is an autoimmune disease characterised by chronic inflammation of the joints. The prevalence of RA is 0.5 to 1% in many populations [1] and is about 2.5 times higher in women [2]. RA has a very complex genetic basis, and the combination of genetic and environmental causative factors makes it hard to study. The genetic contribution to RA susceptibility is estimated to be between 30% and 50%, of which the major histocompatibility complex accounts for about one-third [3].
Animal models provide a valuable tool for finding genes contributing to the susceptibility to and severity of RA. Rats are very useful for this purpose because autoimmune experimental arthritis phenotypes can be induced in susceptible strains by several agents, such as collagen, pristane, oil, streptococcal cell wall and even adjuvant alone [4-6]. Intercrosses of such susceptible rat strains with resistant strains are used for establishing linkage between genetic markers and quantitative traits distinguishing the arthritis phenotype. Statistically valid linkage between such genomic regions and measurements of quantitative traits are called quantitative trait loci (QTLs). More than 40 QTLs that regulate experimentally induced arthritis have been identified in different rat crosses [7]. Most of these QTLs are several megabases in size, containing many possible gene candidates. Several experimental strategies are used to narrow these regions, and these attempts almost always are combined with the retrieval of potential candidate genes found in different databases.
Information about RA and related genome data is available in several different forms, from raw data to descriptive text. One important difference between raw data and data based on human evaluation is that human evaluation often yields an interpretation that gives meaning to the data. Thus, human considerations bring an added value to genome data, which makes textual description an important source for investigating gene function. However, the amount of free text about RA is growing very fast, so there is an increasing need for developing a tool to help scientists distinguish relevant information from background noise. To facilitate this kind of data mining, we have created a tool, the Candidate Gene Capture (CGC) application, that makes keyword-based searches on textual information for genes situated within selected human chromosomal intervals that are homologous to a given rat QTL. Depending on the connection to RA, the keywords are allocated different values. The values for all matching keywords are summarised for each gene, the final values indicating which genes might be good candidates for contributing to the arthritis phenotype. When evaluated, this approach produces similar rankings to those done manually. In addition, this approach also manages to predict several candidate genes that are already established in the literature. Thus, the CGC application is a helpful tool for finding candidate genes associated with experimentally induced arthritis in rat.
Materials and methods
The focus of this work is the development of a web-based tool that facilitates the identification of potential gene candidates that contribute to experimentally induced autoimmune arthritis. The application, called CGC, was created by combining QTL regions in rat with human gene homology data, descriptions of phenotypic gene effects and selected keywords.
QTL data
Data describing 37 experimentally induced autoimmune arthritis QTLs in rat were obtained from the RatMap database [7]. These data were originally collected from experimentally induced inflammatory arthritis in rat strains susceptible to the following inducing agents: pristane, collagen, streptococcal cell wall, oil or adjuvant alone. Accordingly, the resulting QTLs are named Pristane-induced arthritis (Pia), Collagen-induced arthritis (Cia), Streptococcal cell wall-induced arthritis (Scwia), Oil-induced arthritis (Oia) and Adjuvant-induced arthritis (Aia).
The QTL data retrieved from RatMap include the locus symbol, a QTL description, the chromosomal position and flanking markers defining the borders of the QTL. The range of each QTL was based on the LOD score thresholds suggested in the corresponding papers. These data were stored in a MySQL table labelled 'QTL'.
Gene homology data
Human gene data were assembled primarily from National Centre for Biotechnology Information (NCBI) [8] and the University of California Santa Cruz genome browser [9]. The genome information from NCBI consisted of official gene symbol, chromosome number, Locus Link ID, Online Mendelian Inheritance in Man (OMIM) ID, human Genome Database (GDB) accession ID and Refseq ID. Sequence positions were obtained exclusively from the University of California Santa Cruz genome browser, comprising transcript start/stop, codon start/stop, exon start/stop and number of exons in each gene. From this set of data, a table of human genes ordered by codon start was generated and labelled 'HsRn'.
To find orthologous gene pairs between rat and human, 1,464 chromosomally localised rat genes were obtained from RatMap. About 1,000 of these genes had a known homologous gene mapped in human. The orthologous rat/human gene pairs were characterised by the human data already present in table 'HsRn' together with the official rat gene symbol, rat chromosome number and RatMap ID.
Two flanking markers define each QTL used in this study. To find a human sequence homologous to a rat QTL region, an integrated linkage map containing rat genes and polymorphic DNA markers was used . For each QTL a pair of rat genes (obtained from the integrated linkage map) that were localised at, or close to, the two markers flanking the QTL and orthologous to human genes, was selected. The human chromosomal interval defined by these two orthologous genes was expected to contain a sequence homologous to the rat QTL. Because the homologous QTL interval often contained segments from more than one human chromosome, all orthologous rat/human gene pairs within each QTL were used to find smaller human chromosomal segments to comprise the total list of human genes confined within the homologous region. Information on rat and human gene symbols, chromosomal positions and codon start for all genes included in the homologous interval (obtained from table 'HsRn') was stored in QTL-specific tables labelled with the same symbol as the corresponding QTL.
Downloading gene function data
The OMIM database [10] contains a comprehensive record of gene function and clinical data, which was used as a source for keyword querying in the CGC application. For each human gene within the selected intervals, gene function information was downloaded from OMIM and stored in a table labelled 'OMIMdata'.
Selecting keywords and running the application
The querying process in this application is divided into four steps: finding a QTL of interest, displaying the rat/human homologous QTL region, selecting and ranking keywords, and searching OMIM text for selected keywords.
Finding a QTL of interest
The first step in finding candidate genes for a specific QTL is to choose a QTL of interest. To make this possible, we simply made the QTL database table directly available through a web interface. In this way, the user can access all QTLs in our database by searching for the locus symbol, the chromosome number and/or a descriptive text. The resulting QTLs are presented, together with a brief description obtained from the QTL table.
Displaying the rat/human homologous QTL region
Next, the user can select the preferred QTL. The resulting web page presents all rat/human gene pairs within the chosen rat QTL region, together with all human genes in the homologous human genomic region that are found in OMIM. These data are obtained from the corresponding 'QTL-specific' table.
Thus, all rat genes within a selected QTL and all genes within the homologous human genomic region are displayed. Because the human genome is better characterised than the rat genome, more human genes are usually displayed.
Selecting and ranking of keywords
For all arthritis QTLs a total of 49 default keywords were chosen. Most keywords were obtained by selecting all terms found directly under the MeSH (Medical Subject Headings) terms 'autoimmune diseases' and 'rheumatoid arthritis' in the PubMed MeSH-term database [11]. Some of these terms were truncated to optimise the querying process. In addition, a set of keywords frequently used in arthritis-related literature was added to the default keyword list.
To estimate the relative importance of the default keywords in relation to arthritis, each keyword was given a value depending on its relevance to arthritis. This relevance index was calculated as the number of PubMed abstracts containing both the keyword and the word 'arthritis' divided by the total number of abstracts containing the keyword alone. The relevance indices were multiplied by 100 to generate the final keyword values as percentages.
The application also allows the user to add up to 10 keywords of his or her own choice, and the corresponding keyword values are automatically generated on the basis of the same principle as for the default keyword values. Optionally, the user can overrule all keyword values, including the default ones.
Searching OMIM text for selected keywords
When searching a QTL for all the default keywords, alternatively deselecting unwanted ones and/or adding new ones, the keyword values for all keywords found within each OMIM text (locally stored in the table 'OMIMdata') will be summarised. To take advantage of the large amount of knowledge concerning the human genome, records in OMIM for all genes within the human homologous segment are used in the search, including genes not present in the rat gene list. For each gene, the total sum of all keyword values will be displayed, which indicates its relevance as a candidate gene. Each keyword is only counted once, independently of the number of times it occurs within a given OMIM text.
Results
In the CGC application presented in this paper, all known rat genes within a selected QTL, along with all human genes within the homologous interval, are retrieved and displayed from a table that has the same name as the selected QTL. A list with an array of 49 selectable arthritis related keywords is presented together with their respective keyword values. Up to 10 additional keywords can be added and their keyword values are automatically calculated. When performing a search, the textual information for each human gene stored in the table 'OMIMdata' is scanned for all selected keywords. The genes and all keywords found in the accompanying text are displayed, together with the sum of all matching keyword values.
To estimate whether the CGC application was able to rank candidate genes in fashion similar to human evaluations, gene descriptions for four randomly selected QTL regions (Cia4, Cia10, Cia14 and Cia17) were surveyed manually. For all genes within the selected QTL regions, we compared the outcome of the CGC gene ranking with our own manual evaluation of each OMIM text. The manual rating was made without knowledge of the CGC ranking. To put the application and the manual inspection at a similar level, we tried to base our evaluation on the written OMIM texts only, without taking other information into account. In the manual inspection the OMIM texts were divided into five different classes: (1) obvious gene candidate, (2) likely gene candidate, (3) possible gene candidate, (4) unlikely gene candidate and (5) gene without relevance.
In addition, the genes that were ranked as high by the CGC application were further scrutinised in an extensive analysis of related papers not found in the OMIM reference lists. Finally, the NCF1 gene was studied in detail.
Cia4
In total, 12 genes were ranked by the CGC tool. IFNG was rated as the top candidate by the CGC application and it was also considered to be the most appropriate gene candidate for collagen-induced arthritis within this QTL according to the manual inspection. IL22 was considered the next highest gene candidate both by the CGC application and the manual inspection.
IFNG (interferon-γ), CGC points 291.1, CGC ranking 1, manual rating 1
IFNG was identified by the CGC application on the basis of 10 different keywords: 'rheumatoid', 'HLA', 'sjogren', 'T cell', 'mhc', 'lymphocyte', 'antigen', 'cytokine', 'arthritis' and 'infecti'. IFNG has been shown to be closely associated with RA. In a study of 99 patients with RA of different severity, susceptibility to, and severity of, RA was shown to be related to a microsatellite polymorphism within the first intron of the gene encoding interferon-γ [12].
IL22 (interleukin-22), CGC points 14.1, CGC ranking 2, manual rating 2
IL22 was selected by the keywords 'inflam', 'T cell', 'lymphocyte' and 'cytokine'. IL22 activates three different STAT genes: STAT1, STAT3 and STAT5 [13]. RA synovial fibroblasts are relatively resistant to apoptosis and exhibit dysregulated growth. Retrovirus-mediated gene transfer of dominant-negative mutant STAT3 genes blocks the endogenous STAT3 expression in synovial fibroblasts from patients with RA, leading to failure of growth in the cell culture and apoptosis [14].
A middle group of two genes was selected with the CGC application: MYC (CGC points 10.9, CGC ranking 3, manual rating 3) and HMGIC (CGC points 10.5, CGC ranking 4, manual rating 4).
Cia10
In total, 35 genes were ranked by the CGC tool. RPL7 and NKFB1 were ranked as the two top candidates by the CGC application. These two genes were also manually considered to be the most appropriate gene candidates for collagen-induced arthritis within this QTL.
NFKB1 (nuclear factor κB 1), CGC points 219.7, CGC ranking 1, manual rating 1
The very high point that NFKB1 obtained from the keyword query was in part due to the word 'arthritis' appearing in the corresponding OMIM text. Twelve other keywords were also found to be making a substantial contribution. According to the OMIM record, NFKB1 is a very strong gene candidate because the inappropriate activation of NKFB1 is known to be linked to inflammatory events associated with autoimmune arthritis [15].
RPL7 (ribosomal protein L7), CGC points 37.3, CGC ranking 2, manual rating 1
The RPL7 gene was rated second by the CGC application mainly because of the keywords 'autoimmune', 'lupus' and 'erythematosus'. The RPL7 protein is reported to be a major autoantigen in systemic autoimmune arthritis [16].
A middle group of five genes was rated as relatively high by the CGC application: COL6A3 (CGC points 24.2, CGC ranking 3, manual rating 3), CSF1 (CGC points 17.4, CGC ranking 4, manual rating 3), EDG1 (CGC points 12.5, CGC ranking 5, manual rating 5), VCAM1 (CGC points 11.3, CGC ranking 6, manual rating 2) and PAPSS1 (CGC points 9.3, CGC ranking 7, manual rating 3). Among these genes, CSF1 is a possible gene candidate because recent studies have shown that synovial tissue in RA joints secretes CSF1 together with several other cytokines, which increases the osteoclast activity [17]. VCAM1 might also be a potential gene candidate because it is expressed in endothelial cells of the blood vessels, facilitating the adhesion of leucocytes [18]. EDG1 was a false prediction because the term 'HLA' matched an author (Hla T. Maciag T. J Biol Chem 1990;265:9308-13) and the term 'T cell' matched 'mutant cell'.
Cia14
In total, 16 genes were ranked by the CGC tool. The two top ranked genes according to the CGC application (IL15 and HMOX1 ) were also the highest-rated genes in the manual inspection.
IL15 (interleukin-15), CGC points 27.3, CGC ranking 1, manual rating 1
IL15 was ranked in first place by the CGC application. In the corresponding OMIM text, IL15 is associated with the keywords 'autoimmun', 'inflam', 'T cell', 'lymphocyte', 'antigen', 'cytokine' and 'infecti', but not 'arthritis'. In a recent paper it was shown that increased serum levels of IL15 are found in patients with long-term RA [19].
HMOX1 (haem oxidase 1), CGC points 13.5, CGC ranking 2, manual rating 1
HMOX1 was ranked second by the CGC application with the keywords 'anemia', 'hemolytic', 'inflam' and 'T cell'. HMOX1 has been shown to be involved in the treatment of RA with gold(I)-containing compounds. Gold(I) drugs selectively activate a transcription factor (Nrf2/small Maf heterodimer), which induces the transcription of anti-oxidative stress genes, including HMOX1, and inhibits inflammation [20].
A middle group of four genes were rated as relatively high by the CGC application: ITK (CGC points 9.7, CGC ranking 3, manual rating 2), NFATC3 (CGC points 9.7, CGC ranking 3, manual rating 3), AARS (CGC points 9.2, CGC ranking 5, manual rating 3) and KARS (CGC points 9.2, CGC ranking 5, manual rating 3).
Cia17
In total, 30 genes were ranked by the CGC tool (only one member of the PCDH gene family was included). In the manual inspection, no 'obvious' candidate gene was found. However, four genes were considered to be 'likely' gene candidates. One of these, CD74, also received the highest keyword sum in the CGC application. Another gene among the likely gene candidates, SLC26A2, was ranked second by the CGC application.
CD74, CGC points 27.7, CGC ranking 1, manual rating 3
The CD74 gene was ranked in first place by the CGC application because of results from six different keywords: 'antigen', 'HLA', 'immunoglobulin', 'T cell', 'MHC' and 'inflam'. In a recent paper by Leng and colleagues [21], not present in the OMIM text, CD74 is reported to be required for macrophage migration inhibitory factor (MIF)-induced activation of the extracellular signal-regulated kinase-1/2 mitogen-activated protein kinase cascade, cell proliferation, and prostaglandin E2 production. MIF is an upstream activator of monocytes/macrophages and is centrally involved in the pathogenesis of RA and other inflammatory conditions.
SLC26A2 (solute carrier family 26 member 2), CGC points 24.2, CGC ranking 2, manual rating 2
SLC26A2 was associated with the keyword 'joint'. SLC26A2 is an anion transporter responsible for four recessively inherited chondrodysplasias: multiple epiphyseal dysplasia (MED) [22], diastrophic dysplasia (DTD) [23], atelosteogenesis Type II (AO2) [24] and achondrogenesis type IB (ACG1B) [25]. However, although other forms of chondrodysplasias such as progressive pseudorheumatoid chondrodysplasia show symptoms similar to those of RA, no clear link between SLC26A2 and RA can be concluded.
A middle group of four genes were ranked in positions 3 to 6 by the CGC application: NR3C1 (CGC points 16.5, CGC ranking 3, manual rating 2), SPINK5 (CGC points 14.2, CGC ranking 4, manual rating 3), IK (CGC points 14.1, CGC ranking 5, manual rating 3) and CD14 (CGC points 12.8, CGC ranking 6, manual rating 2). Two of these genes might be related to RA. NR3C1 is significantly overexpressed in untreated patients with RA and in several clinical studies of inflammatory conditions, such as RA [26]. CD14 has been reported to be associated with significantly elevated serum levels in patients with RA [27,28].
NCF1 (neutrophilic cytosolic factor 1)
The gene NCF1 is covered by both the Cia12 and Pia4 QTLs and was assigned a total point of 238.9 by the CGC application. This suggests that NCF1 is a strong gene candidate for RA. Indeed, NCF1 has been identified as a gene that has a naturally occurring polymorphism regulating arthritis severity in rats [29]. On looking at the OMIM text for NCF1, it is clear that most of the points come from the part of the text describing these particular findings. To evaluate the ability of the tool to predict genes that are reported to be related to the arthritis phenotype, the OMIM text was used in the form in which it existed before NCF1 was shown to be associated with arthritis; that is, the part of the OMIM text describing the association between NCF1 and arthritis was deleted before running the application. The resulting keyword sum was, as expected, much lower, with a total point of 10.8. However, these points were still sufficient to rank NCF1 as the top candidate of Cia12 and Pia4 . Recently, the gene GUSB was updated at OMIM, resulting in a total point of 30.7.
Discussion
A common feature of many genetically orientated RA studies is to find genes responsible for, or contributing to, one or several RA-related phenotypes. Typically, a genomic region might be known to be associated with a phenotype, but still there are usually many genes within such a region that might be possible candidates. Specifically, when employing QTL analysis in rats, selecting gene candidates has become a recurrent part of the data analysis. An important part of the search for candidate genes is checking the available bioinformatic resources; most often the written information describing gene function is very informative. The aim of this study was to facilitate this data mining by generating a web-based tool called Candidate Gene Capture (CGC), whose purpose is to identify potential candidate genes associated with experimentally induced arthritis phenotypes in rats.
In brief, the CGC application makes it possible to retrieve a large number of QTL regions previously described in the literature. For each rat QTL, the homologous genomic region in humans is automatically displayed. All genes included in the corresponding human genomic interval can be queried for up to 49 default keywords and up to 10 keywords selected by the user. Each keyword is given a value based on an algorithm that estimates how closely related a keyword is to the term 'arthritis' according to their simultaneous occurrence in PubMed abstracts. OMIM records for human genes in a selected genomic region are ranked by their total keyword values; that is, the sum of the values for all keywords that hit a record. The higher the total keyword sum is, the more likely it is to be a gene candidate. The application can be accessed from the RatMap home page [7] or directly at .
Comparison of manual evaluation with CGC ranking
To estimate the ability of the CGC application to rank candidate genes in a fashion similar to human evaluation, an independent manual inspection was made. Four randomly selected collagen-induced arthritis QTLs were used (Cia4, Cia10, Cia14 and Cia17 ). The OMIM records used in the CGC prediction were surveyed manually and rated on a scale from 1 to 5. Comparing the manual and CGC ratings, it was found that the two highest-ranked candidate genes in the CGC application for all QTLs studied were rated as high in the manual evaluation, with the exception of one gene, CD74 in Cia17 . However, CD74 turned out to be a very likely gene candidate when additional literature was surveyed (see below).
In an extended literature search for the two highest CGC-ranked genes of Cia4, Cia10, Cia14 and Cia17, it was confirmed that seven of eight genes were clearly associated with RA. Literature not covered by the OMIM reference lists revealed that three of these genes (IL5, CD74 and HMOX1 ) had a strong association with RA. Many different keywords fitted each of the OMIM records associated with these three genes. Although none of these keywords had a very high keyword value (ranging from 1.6 to 9.7), the resulting keyword sums (IL15, 27.3; CD74, 22.3; HMOX1, 13.5) still clearly diverged from the keyword sums of other genes within the same QTLs. Thus, the CGC application is able to predict candidate genes from OMIM records even though the association with RA is not explicitly mentioned in the text.
In addition to the two highest-ranked genes in the four QTLs evaluated, we also designated a middle group of candidate genes that were ranked in positions 3 to 6 by the CGC application (except for Cia4, in which the middle group comprised genes ranked in positions 3 and 4). The remaining genes for each investigated QTL formed a separate group (the low group). Comparing the mean values of the CGC ranking with the manual ratings for these three groups (the two highest, the middle group and the low group), a general agreement was found in the ranking of candidate genes (Table 1). The only exception was the relatively low manually rated 'best two' group for Cia17, which is fully explained by the low manual rating of CD74 . As described above, on closer inspection the manual rating of CD74 turned out to be too cautious.
Finally, gene records without any keyword hits at all were not found to be associated with RA in the manual inspection.
Thus, when the CGC prediction is compared with manual inspection, the conclusion is that the application makes a reliable evaluation of the OMIM records for the four QTLs studied in detail. For three genes (IL5, CD74 and HMOX1 ) the CGC application estimated the gene records as being more interesting than the manual inspection, an estimation confirmed by recent papers not yet included in the OMIM reference list. This shows that the CGC application is a very helpful tool for finding gene candidates contributing to RA. Furthermore, the CGC application also seems to follow our manual interpretation for genes that might be of interest (referred to as the 'middle group') as well as for genes with no evident connection to RA.
Keywords
No clear-cut connection can be made between the absolute sum of keyword values and the relevance of candidate genes. However, our evaluation of the four Cia QTLs implies that the ranking of the genes within each QTL based on the keyword sums provides a good prediction of the best candidate genes. For example, in QTL region Cia12, NCF1 has been shown by Olofsson and colleagues to be involved in the regulation of arthritis severity in rats [29]. As expected, NCF1 also obtains a very high keyword sum (225.6), mainly because of the description of Olofsson's findings in the OMIM text. When this description is excluded from the OMIM record, the NCF1 keyword sum decreases to 10.8. This still made NCF1 the highest-ranked gene in this QTL region. As exemplified above, the CGC application is able to find candidate genes even though their relatedness to RA is not explicitly mentioned in the text investigated. In the paper describing Olofsson's findings, the authors stated that they found the candidate gene approach distracting, even though they were facing a region that contained a small set of genes. This could very well be so, but when analysing the genes within a QTL it seems reasonable to start with the most likely candidate genes rather than with randomly picked ones, especially if the region contains a large number of genes. The CGC application makes an unbiased evaluation of genes within a region, indicating which are the most favourable ones to start analysing. Looking at the NCF1 example retrospectively, CGC would in fact have suggested NCF1 as the most probable candidate gene, although this might be a fortunate case.
Among the selected keywords, occasionally there were a few that gave false positives. One example is the word 'joint' (point 24.2), which at times referred to other terms, such as 'joint maximum LOD score'. For example, this caused the gene KEL to be ranked highest (28.7) for the Aia2 QTL. Another example is 'T cell' (points 2.8), which can produce results such as mutant cell or that cell, as found in the OMIM record for EDG1 (Cia10 ). In addition, it was found that some keywords can be misinterpreted as author names. EDG1, for example, was falsely predicted as a candidate gene partly because the term 'HLA' matched an author (Hla T. Maciag T. J Biol Chem 1990;265:9308-13).
Forty-nine keywords were selected, based on PubMed MeSH terms and other terms frequently found in the literature on RA. However, this might not be a completely exhaustive set of keywords and a user of the CGC tool might want to extend or exchange parts of this keyword list. To make this possible, the user can add up to 10 keywords of his or her own and can automatically obtain the corresponding keyword values calculated. These keywords can be used alone or together with the whole or parts of the default keyword list. It should be emphasised that there is really no harm in using a large number of keywords, because irrelevant keywords, such as 'and' or 'is', will get almost no keyword values, thus not disturbing the selecting process. In addition, the user is allowed to overrule all keyword values if preferred and enter values of his or her own choice.
Comparison with related databases
To our knowledge there are three databases other than CGC that address the problem of finding candidate genes for complex disorders.
GeneSeeker is a web-based tool that permits the user to search different databases simultaneously, given a known human genetic location and an expression or phenotypic pattern(s) [30]. Moreover, data from syntenic regions in mouse can be included in the queries. The tool is a general instrument that has its strength in the range of databases covered. However, GeneSeeker has no means for prioritizing between the genes retrieved. Because the CGC tool is specifically adapted for arthritis models, much more keywords relevant to this phenotype are available here although both applications permit the user to enter his or her own keywords.
POCUS (Prioritizing Of Candidate genes Using Statistics) is an application that rates genes on the basis of their similarity to a set of genes generally considered to be associated with a given complex trait [31]. The similarity is quantified by measuring the number of functional annotations (Gene Onthology terms or InterPro domain ID) and/or expression pattern terms and IDs in common (Unigene or NCBI). Although POCUS prioritizes between the gene candidates, the strategy is different from that used for CGC. The genes associated with a given trait are not restricted to a specific genomic region. However, the authors claim that the application might be extended to work in such a way. POCUS is not a web-based tool but can be downloaded.
G2D (candidate Genes To inherited Diseases) is another database accessible from the web [32]. G2D is built on a strategy resembling that of CGC. In brief, chemical terms have here been given scores calculated in a similar fashion to that in CGC; that is, the simultaneous occurrence of chemical terms (MeSH-C) and pathological conditions (MeSH-D) in PubMed. For a given disease several pathological conditions were selected on the basis of a set of representative papers. These pathological conditions were then related to functional descriptions (Gene Ontology terms) by using RefSeq annotations (RefSeq-NCBI) as mediating links, and the degree of relatedness were represented by 'GO-scores'. A gene can be related to a given disease by calculating the average GO-score annotated for that gene. In many ways this approach resembles that described in this paper, although G2D depends on Gene Onthology terms instead of a full text. Moreover, G2D uses the mean GO-score for rating genes rather than calculating the sum. As a consequence, a gene with a GO-score based on just a single Gene Ontology term is rated higher than a gene that is annotated for the same term together with additional Gene Ontology terms with lower scores. Furthermore, in contrast to CGC, the GD2 database is a static database in which no data input from the user is possible, and at present no information on RA is available.
Future developments
As our next step we plan to evolve the CGC application to include other text-based resources, such as PubMed abstracts, Swiss-Prot descriptions and, as a complement, Gene Ontology terms. In addition, we are currently extending the CGC tool to include rat QTLs for metabolic disorders, mainly focused on diabetes mellitus type II. The long-term goal is that the CGC tool will be able to predict candidate genes for any given type of rat QTL, such as multiple sclerosis, blood pressure or obesity. The strategy used in CGC could also be applied on QTLs in other species, such as mouse or human.
Conclusion
We conclude that the excellent agreement between our manual evaluation and the rankings made by the CGC application for the four different QTLs tested (Cia4, Cia10, Cia14 and Cia17 ), as well as the prediction of the NCF1 gene, clearly show that this tool makes very reliable predictions. Consequently, we believe that the CGC tool can be of great use in facilitating the finding of gene candidates related to the arthritis phenotype.
Abbreviations
Aia = Adjuvant-induced arthritis; CGC = Candidate Gene Capture; Cia = Collagen-induced arthritis; NCBI = National Centre for Biotechnology Information; OMIM = Online Mendelian Inheritance in Man; Pia = Pristane-induced arthritis; QTL = quantitative trait locus; RA = rheumatoid arthritis.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
LA performed the programming of the CGC application, contributed original ideas on assigning keyword values and drafted the manuscript. GP created the rat/human comparative database, implemented it in the CGC application and drafted the manuscript. PJ had main responsibility for all supporting functions of the application and was involved in the theoretical basis of the work. FS supervised the project, contributed with original ideas and took full part in the preparation of the manuscript. All authors read and approved the final manuscript.
Acknowledgements
This work was supported in part by the Swedish Medical Research Council, the SWEGENE Foundation, the Sven and Lilly Lawski Foundation, the Royal Society of Arts and Sciences in Goteborg, the Wilhelm and Martina Lundgren Research Foundation and the Royal Hvitfeldtska Foundation.
Figures and Tables
Table 1 Comparison between manual evaluation and Candidate Gene Capture (CGC) rating
Best two Middle group Low group
QTL CGC Manual CGC Manual CGC Manual
Cia4 152.6 1.5 10.7 3.5 2.3 3.9
Cia10 128.5 1.0 14.9 2.0 3.6 3.9
Cia14 20.4 1.0 9.5 2.7 4.6 3.4
Cia17 26.0 2.5 14.4 2.5 4.9 3.8
Mean values of keyword sums and manual ratings for genes in three groups are shown, on the basis of their ranking by the CGC application. QTL, quantitative trait locus.
==== Refs
Felson DT Koopman WJ Epidemiology of rheumatic diseases Arthritis and Allied Conditions – A Textbook of Rheumatology 1997 Baltimore, MD: Williams & Williams 3 10
Wilder RL Schumacher HR Jr, Klippel JH, Koopman WJ Rheumatoid arthritis: epidemiology, pathology, and pathogenesis Primer on the Rheumatic Diseases 1993 10 Atlanta: Arthritis Foundation 86 89
Deighton CM Walker DJ Griffiths ID Roberts DF The contribution of HLA to rheumatoid arthritis Clin Genet 1989 36 178 182 2676268
Wilder RL Griffiths MM Cannon GW Caspi R Remmers EF Susceptibility to autoimmune disease and drug addiction in inbred rats. Are there mechanistic factors in common related to abnormalities in hypothalamic–pituitary–adrenal axis and stress response function? Ann NY Acad Sci 2000 917 784 796 11268408
Griffiths MM Remmers EF Genetic analysis of collagen-induced arthritis in rats: a polygenic model for rheumatoid arthritis predicts a common framework of cross-species inflammatory/autoimmune disease loci Immunol Rev 2001 184 172 183 12086311 10.1034/j.1600-065x.2001.1840116.x
Holmdahl R Dissection of the genetic complexity of arthritis using animal models J Autoimmun 2003 21 99 103 12935777 10.1016/S0896-8411(03)00096-9
RatMap, Rat Genome Database, Dept for Cell and Molecular Biology, Goteborg University, Sweden
Human Genome Resources, National Center for Biotechnology Information, National Library of Medicine (Bethesda, MD)
Genome Bioinformatics Group at University of California Santa Cruz (UCSC)
Online Mendelian Inheritance in Man, OMIM™. McKusick-Nathans Institute for Genetic Medicine, Johns Hopkins University (Baltimore, MD) and National Center for Biotechnology Information, National Library of Medicine (Bethesda, MD)
PubMed, National Center for Biotechnology Information, National Library of Medicine (Bethesda, MD)
Khani-Hanjani A Lacaille D Hoar D Chalmers A Horsman D Anderson M Balshaw R Keown PA Association between dinucleotide repeat in non-coding region of interferon-gamma gene and susceptibility to, and severity of, rheumatoid arthritis Lancet 2000 356 820 825 11022930 10.1016/S0140-6736(00)02657-X
Xie MH Aggarwal S Ho WH Foster J Zhang Z Stinson J Wood WI Goddard AD Gurney AL Interleukin (IL)-22, a novel human cytokine that signals through the interferon receptor-related proteins CRF2-4 and IL-22R J Biol Chem 2000 275 31335 31339 10875937 10.1074/jbc.M005304200
Krause A Scaletta N Ji JD Ivashkiv LB Rheumatoid arthritis synoviocyte survival is dependent on Stat3 J Immunol 2002 169 6610 6616 12444174
Chen F Castranova V Shi X Demers LM New insights into the role of nuclear factor-kappa-B, a ubiquitous transcription factor in the initiation of diseases Clin Chem 1999 45 7 17 9895331
Neu E von Mikecz AH Hemmerich PH Peter HH Fricke M Deicher H Genth E Krawinkel U Autoantibodies against eukaryotic protein L7 in patients suffering from systemic lupus erythematosus and progressive systemic sclerosis: frequency and correlation with clinical, serological and genetic parameters. The SLE Study Group Clin Exp Immun 1995 100 198 204 7743655
Gravallese EM Bone destruction in arthritis Ann Rheum Dis 2002 61 Suppl 2 ii84 ii86 12379632
Carter RA O'Donnell K Sachthep S Cicuttini F Boyd AW Wicks IP Characterization of a human synovial cell antigen: VCAM-1 and inflammatory arthritis Immunol Cell Biol 2001 79 419 428 11564149 10.1046/j.1440-1711.2001.01019.x
Gonzalez-Alvaro I Ortiz AM Garcia-Vicuna R Balsa A Pascual-Salcedo D Laffon A Increased serum levels of interleukin-15 in rheumatoid arthritis with long-term disease Clin Exp Rheumatol 2003 21 639 642 14611115
Kataoka K Handa H Nishizawa M Induction of cellular antioxidative stress genes through heterodimeric transcription factor Nrf2/small Maf by antirheumatic gold(I) compounds J Biol Chem 2001 276 34074 34081 11429414 10.1074/jbc.M105383200
Leng L Metz CN Fang Y Xu J Donnelly S Baugh J Delohery T Chen Y Mitchell RA Bucala R MIF signal transduction initiated by binding to CD74 J Exp Med 2003 197 1467 1476 12782713 10.1084/jem.20030286
Superti-Furga A Neumann L Riebel T Eich G Steinmann B Spranger J Kunze J Recessively inherited multiple epiphyseal dysplasia with normal stature, club foot, and double layered patella caused by a DTDST mutation J Med Genet 1999 36 621 624 10465113
Hastbacka J de la Chapelle A Mahtani MM Clines G Reeve-Daly MP Daly M Hamilton BA Kusumi K Trivedi B Weaver A The diastrophic dysplasia gene encodes a novel sulfate transporter: positional cloning by fine-structure linkage disequilibrium mapping Cell 1994 78 1073 1087 7923357 10.1016/0092-8674(94)90281-X
Hastbacka J Superti-Furga A Wilcox WR Rimoin DL Cohn DH Lander ES Atelosteogenesis type II is caused by mutations in the diastrophic dysplasia sulfate-transporter gene (DTDST): evidence for a phenotypic series involving three chondrodysplasias Am J Hum Genet 1996 58 255 262 8571951
Superti-Furga A Hastbacka J Wilcox WR Cohn DH van der Harten HJ Rossi A Blau N Rimoin DL Steinmann B Lander ES Achondrogenesis type IB is caused by mutations in the diastrophic dysplasia sulphate transporter gene Nat Genet 1996 12 100 102 8528239 10.1038/ng0196-100
Neeck G Kluter A Dotzlaw H Eggert M Involvement of the glucocorticoid receptor in the pathogenesis of rheumatoid arthritis Ann NY Acad Sci 2002 966 491 495 12114309
Horneff G Sack U Kalden JR Emmrich F Burmester GR Reduction of monocyte-macrophage activation markers upon anti-CD4 treatment: decreased levels of IL-1, IL-6, neopterin and soluble CD14 in patients with rheumatoid arthritis Clin Exp Immunol 1993 91 207 213 7679058
Yu S Nakashima N Xu BH Matsuda T Izumihara A Sunahara N Nakamura T Tsukano M Matsuyama T Pathological significance of elevated soluble CD14 production in rheumatoid arthritis: in the presence of soluble CD14, lipopolysaccharides at low concentrations activate RA synovial fibroblasts Rheumatol Int 1998 17 237 243 9592864 10.1007/s002960050041
Olofsson P Holmberg J Tordsson J Lu S Akerstrom B Holmdahl R Positional identification of Ncf1 as a gene that regulates arthritis severity in rats Nat Genet 2003 33 25 32 12461526 10.1038/ng1058
van Driel MA Cuelenaere K Kemmeren PP Leunissen JA Brunner HG A new web-based data mining tool for the identification of candidate genes for human genetic disorders Eur J Hum Genet 2003 11 57 63 12529706 10.1038/sj.ejhg.5200918
Turner FS Clutterbuck DR Semple CA POCUS: mining genomic sequence annotation to predict disease genes Genome Biol 2003 4 R75 14611661 10.1186/gb-2003-4-11-r75
Perez-Iratxeta C Bork P Andrade MA Association of genes to genetically inherited diseases using data mining Nat Genet 2002 31 316 319 12006977
| 15899035 | PMC1174944 | CC BY | 2021-01-04 16:02:36 | no | Arthritis Res Ther. 2005 Feb 18; 7(3):R485-R492 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1700 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17011589903810.1186/ar1701Research ArticleANKH variants associated with ankylosing spondylitis: gender differences Tsui Hing Wo [email protected] Robert D [email protected] Andrew D [email protected] John D [email protected] Florence WL [email protected] Toronto Western Research Institute, Toronto, Ontario, Canada2 University of Toronto, Toronto, Ontario, Canada3 The Hospital for Sick Children, Toronto, Ontario, Canada4 The University of Texas-Houston Health Science Center, Houston, Texas, USA2005 25 2 2005 7 3 R513 R525 27 10 2004 16 12 2004 21 1 2005 24 1 2005 Copyright © 2005 Tsui et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The ank (progressive ankylosis) mutant mouse, which has a nonsense mutation in exon 12 of the inorganic pyrophosphate regulator gene (ank), exhibits aberrant joint ankylosis similar to human ankylosing spondylitis (AS). We previously performed family-based association analyses of 124 Caucasian AS families and showed that novel genetic markers in the 5' flanking region of ANKH (the human homolog of the murine ank gene) are modestly associated with AS. The objective of the present study was to conduct a more extensive evaluation of ANKH variants that are significantly associated with AS and to determine whether the association is gender specific. We genotyped 201 multiplex AS families with nine ANKH intragenetic and two flanking microsatellite markers, and performed family-based association analyses. We showed that ANKH variants located in two different regions of the ANKH gene were associated with AS. Results of haplotype analyses indicated that, after Bonferroni correction, the haplotype combination of rs26307 [C] and rs27356 [C] is significantly associated with AS in men (recessive/dominant model; P = 0.004), and the haplotype combination of rs28006 [C] and rs25957 [C] is significantly associated with AS in women (recessive/dominant model; P = 0.004). A test of interaction identified rs26307 (i.e. the region that was associated in men with AS) as showing a difference in the strength of the association by gender. The region associated with AS in women only showed significance in the test of interaction among the subset of families with affected individuals of both genders. These findings support the concept that ANKH plays a role in genetic susceptibility to AS and reveals a gender–genotype specificity in this interaction.
==== Body
Introduction
Ankylosing spondylitis (AS) is a disorder that results in chronic joint and entheseal inflammation, and ankylosis of axial and peripheral joints. It affects approximately 0.1–0.8% of Caucasians [1]. The disease usually begins in young adulthood and can be associated with chronic pain and significant disability. AS is strongly associated with HLA-B27 [2], but analyses of recurrence risk among family members [3] suggest that at least three other genetic loci in addition to HLA-B27 are required to confer full susceptibility to AS. However, genome-wide linkage studies have detected very few strongly linked non-major histocompatibility complex (MHC) loci [4-6], implying that non-MHC susceptibility loci have small effects and/or that heterogeneous sets of loci combine with HLA-B27 to confer susceptibility to AS. This complexity highlights the strategic advantage of testing predetermined candidate genes. In addition, although several chromosomal regions showed potential linkage in several genome-wide linkage studies conducted in AS families [5,6], the identities of the predisposing genes in these regions remain largely unknown.
Normal osteogenesis depends critically on maintaining the physiological level of inorganic pyrophosphate (PPi). Abnormal PPi levels can be associated with aberrant bone formation. PPi export from the cell is regulated by the ANK protein [7], and mutant mice (ank/ank), which have a premature stop codon in the 3' end of the ank gene, develop severe ankylosis. As a first step in testing the hypothesis that specific polymorphisms in the ANKH gene might contribute to AS susceptibility, we previously reported the identification of two novel polymorphic sites, one in the 5' noncoding region (ANKH-OR) and the other in the promoter region (ANKH-TR), of ANKH [8]. These two marker alleles are in complete linkage disequilibrium (LD). Our results from a linkage analysis of 124 North American AS families [8] indicated that AS is genetically linked to ANKH, and the locus-specific sibling recurrence risk of ANKH to AS susceptibility (λS) is 1.9 (λS for HLA-B27 is 5.2). Our family-based association analysis on the same families [8] showed that AS is modestly associated with ANKH-OR allele 1 (additive model: P = 0.03). Because of insufficient numbers of informative families, our results did not allow us to distinguish between different modes of inheritance. In addition, our analyses were focused on the 5' end of the gene, using only two markers. For these reasons, we have now carried out fine mapping of the complete ANKH region, including not only the AS families used in the previous study but also an additional 77 multiplex AS families (a total of 201 multiplex AS families).
The prevalence of AS is 2.5 times higher in men than in women [9]. Extensive fusion of the spine is a phenotype of the mouse model ank. There has been a clinical impression that radiographic severity (e.g. the bamboo spine) may be relatively less common in affected women than in men [10-13]. It has also been observed that long-term outcome in AS is worse in men than in women [14,15], but the basis for this difference in severity of clinical expression remains unclear. It is unlikely that the major genetic factors that account for these differences are X-linked because there is no linkage of AS susceptibility with X-chromosome markers [16]. Gender also has a significant impact on heritability in AS. AS has a higher prevalence in the offspring of women than men with AS, and sons of men with AS are 2.5 times more likely than daughters to inherit the disease [17,18]. It remains unclear whether there is gender heterogeneity in non-MHC loci that confer susceptibility to AS. In the present study, we asked whether there is any gender difference in the association of ANKH with AS in multiplex families.
Materials and methods
Ankylosing spondylitis families
The study group comprised 201 Caucasian AS families (a total of 226 nuclear families; Tables 1 and 2). This group was recruited from the Toronto Western Spondylitis Clinic (23 families) and from other sites in the North American Spondylitis Consortium (178 families). All patients met modified New York criteria for the diagnosis of AS [19], which include radiographic evidence of sacroiliitis. Of the affected and unaffected individuals, 60% and 47% were men, respectively. The ages of the individuals ranged from 8 to 75 years. The study was approved by the University Health Network Research Ethics Board and the Committee for the Protection of Human Subjects at the University of Texas Health Science Center-Houston.
Genotyping
DNA from the affected and unaffected family members was prepared from peripheral blood lymphocytes using standard techniques.
Microsatellite markers
Genotyping was performed using three microsatellite markers flanking ANKH on chromosome 5p: D5S1953, D5S1991 and D5S1954. Polymerase chain reaction fragments were run on native polyacrylamide gel, stained with ethidium bromide and visualized using an imager (Bio-Rad, Hercules, CA, USA).
Single nucleotide polymorphisms
Genotyping was performed using seven intronic single nucleotide polymorphisms (SNPs; rs26307 [C/T], rs27356 [C/T], 3088132 [G/C], rs153929 [A/G], rs258215 [A/T], rs28006 [C/T] and rs25957 [C/G]). Optimized allelic discrimination assays for SNPs were purchased from Applied Biosystems (Foster City, CA, USA). The plates were read on an ABI PRISM 7900 sequence detection system (Applied Biosystems).
Statistical analysis
Error checking
To minimize data errors, extensive error checking procedures were used. For microsatellite markers, allele assignment was checked manually for all genotypes by two independent individuals. Size data were converted into discrete allele numbers; samples not following Mendelian patterns of inheritance were identified using Pedmanager (available online at ), and these samples were subjected to repeat genotyping.
Family-based association analyses
The transmission disequilibrium test (TDT) was used to test for transmission of specific alleles from heterozygous parents to affected offspring [20]. We computed the test statistics using the empirical variance option of family-based association testing (FBAT) software, version 1.5.5 (available online at ) [21]. This option is used when testing for associations in an area of known linkage (the null hypothesis assumes no association but linkage) with multiple affected siblings in a family or when multiple nuclear families in a pedigree are considered. This program uses data from nuclear families, sibships, pedigrees or any combination, and provides unbiased tests with or without founder genotypes. Biallelic tests were performed using additive, dominant/recessive genetic models. Haplotype analyses were carried out using the haplotype-based association testing (HBAT) empirical variance (-e) option in the FBAT program. For Bonferroni correction, because eight tests (four haplotypes and two models) were carried out in the HBAT-e analyses, P < 0.00625 (0.05/8) was considered statistically significant.
For analysis of affected men/women, the FBAT command 'setafftrait' was used. The unaffected siblings and parents from the families were coded as unknown (0) phenotype, the affected men were coded as 2, and the affected women as 1. FBAT-e analyses using the setafftrait 1 0 0 command were used to test specifically for affected men, and analyses using the setafftrait 0 -1 0 command were used to test specifically for affected women. To test for differences between family-based association for affected men and women, the setafftrait 1 -1 0 command was used.
TDT was used to estimate the frequency of transmission to the affected men or women of the haplotypes of interest. Findings in one affected individual, randomly selected from each of the multiplex families, were used in the calculations.
Results
Association between specific ANKH variants and ankylosing spondilitis
The ANKH gene encodes for ANKH transcripts with different lengths at the 3' untranslated region. The longer transcript (3928 bp; AB046801) is derived from 12 exons, whereas the shorter transcript (2426 bp; AK001799, which contains the last 1721 bp of this transcript) is derived from 13 exons. We fine-mapped the ANKH gene using 11 markers (Fig. 1): three microsatellite markers (D5S1954, D5S1991 and D5S1953), one 5' untranslated region variant (ANKH-OR), and seven intronic SNPs (rs25957 and rs28006 in intron 1, rs258215 in intron 2, rs153929 in intron 7, 3088132 and rs27356 in intron 8, and rs26307 in intron 12).
As an extension to our previous study [8], we included a total of 201 multiplex AS families in a family-based association analysis (77 additional multiplex AS families were included, in addition to the 124 AS families considered in the first study). All of the families were genotyped with 11 markers in the ANKH region (D5S1953, rs26307, rs27356, 3088132, rs153929, rs258215, rs28006, rs25957, ANKH-OR, D5S1991 and D5S1954). FBAT analyses showed two regions in the ANKH gene where associations between ANKH variants and AS were detected. Using both additive and recessive models, rs27356 [C] was significantly associated with AS (additive model: Z score = 2.54, P = 0.011; recessive model: Z score = 2.32, P = 0.020). However, depending on the model used for the analysis, two different ANKH markers were also associated with AS. Using an additive model, an intron 1 SNP, namely rs25957 [C], was associated with AS (Z score = 2.02, P = 0.043). Using a dominant model, ANKH-OR allele 1 was associated with AS (Z score = 2.20, P = 0.027). The results are summarized in Table 3. However, these markers are located in different haplotype or LD blocks (see below), implying that there is more than one susceptibility locus in the ANKH gene.
Thus, our analyses of 201 multiplex AS families showed that ANKH variants found in two different regions of the ANKH gene are modestly associated with AS. Our working hypothesis was that there are two subsets of AS patients, each with a different predisposing polymorphism in the ANKH locus. Because ANKH has been shown to be an androgen responsive gene [22-24], we considered whether there are gender differences between family-based associations of ANKH variants to AS.
Men with ankylosing spondilitis differ from affected women for association with different ANKH variants
Radiographic features of AS vary between men and women, with extensive spinal ankylosis being relatively infrequent in women with AS [10]. Table 2 summarizes gender information for the affected individuals in the 201 North American multiplex AS families. There were 94 families with both affected men and women in each family, 74 families with affected men only, and 33 families with affected women only. In this cohort of North American multiplex AS families, men have a significantly earlier age at diagnosis than that for women (mean [± standard deviation] age of diagnosis for affected men = 28 ± 11 years [n = 213]; mean age of diagnosis for affected women = 30 ± 11 years [n = 149]; including family as an independent variable [using SAS PROC GLM; SAS Institute Inc., Cary, NC, USA]: F = 5.10, P = 0.025; Fig. 2). In addition, analysis of age at AS diagnosis in affected men did not reveal a normal distribution; rather the distribution was skewed toward an earlier onset.
In view of these gender differences, we re-analyzed our genotyping results along gender lines in two separate FBAT analyses using the setafftrait command. FBAT analysis of transmission of alleles to affected women showed that both rs25957 [G] and rs28006 [T] were associated with AS (additive model and biallelic test: rs25957 [G], Z score = 2.82, P = 0.004; rs28006 [T], Z score = 2.82, P = 0.004; Table 4). These results indicate that only ANKH variants at the 5' end, and not those at the 3' end, of ANKH are associated with AS in affected women. This also suggested that ANKH variants at the 3' end of the gene might be associated with AS only in affected men.
To test this hypothesis, we analyzed transmission of alleles to affected men. FBAT analysis of transmission of alleles to affected men using the setafftrait command showed that two neighbouring ANKH variants at the 3' end of the gene, namely rs26307 [C] and rs27356 [C] (16 kb apart), were associated with AS in affected men as was predicted (additive model: rs26307 [C], Z score = 2.06, P = 0.039; rs27356 [C], Z score = 2.63, P = 0.008; recessive model: rs26307 [C], Z score = 2.51, P = 0.012; rs27356 [C], Z score = 2.99, P = 0.002; Table 5).
Identification of ANKH haplotypes that are associated with ankylosinig spondylitis
Where the aetiological variant is not typed, haplotype-based analysis is more powerful for association studies in which there is significant LD in the region of interest. We took advantage of the data from the HapMap project (12 October 2004 release 12; [25]). The markers we used for genotyping are located in four different haplotype blocks (block 1: rs26307, rs27356; block 2: 3088132 and rs153929; block 3: rs28006 and rs25957; block 4: ANKH-OR and D5S1991).
We carried out haplotype analyses based on this information, using the HBAT empirical variance option in the FBAT program, and the results are summarized in Table 6. For HBAT analyses considering all 226 AS nuclear families, in each of three different haplotype blocks (blocks 1, 2 and 4) there was one haplotype with a significant P value, suggesting that there is heterogeneity in this locus. When HBAT analyses were carried out specifically for affected women, a haplotype with a significant P value was found in haplotype block 3 located at the 5' end of the gene. When HBAT analyses were conducted specifically for affected men, one haplotype with a significant P value was present in block 1, which is located at the 3' end of the gene. These results are consistent with those from single-marker tests in the FBAT analyses. Furthermore, after Bonferroni correction for the number of haplotypes and models (n = 8), the haplotype combination of rs26307 [C] and rs27356 [C] remained significantly associated with AS in men (recessive/dominant model: P = 0.004), and the haplotype combination of rs28006 [C] and rs25957 [C] was significantly associated with AS in women (recessive/dominant model: P = 0.004).
A direct test for differences between family-based association with affected men and women
In order to conclude that there are gender differences in ANKH variants associated with AS, one must show significant heterogeneity between affected men and women. For this purpose, we used the setafftrait 1 -1 0 command to conduct the FBAT-e analyses. We coded unaffected siblings and parents from the families as unknown phenotype (0), affected men as phenotype 2, and affected women as phenotype 1. The setafftrait 1 -1 0 command converted affect status to trait 1 (affected men), -1 (affected women) and 0 (unaffected siblings and parents), and the results are summarized in Table 7. The only marker with a significant P value was rs26307 [C] (dominant/recessive model: P = 0.03), suggesting that this marker was significantly associated with AS only in affected men.
In view of this finding, we considered whether there is a subset of AS multiplex families in which ANKH variants were significantly associated with AS only in affected women. As summarized in Table 2, there were two types of families in our cohort of multiplex AS families: families with affected individuals of both genders; and families with only one gender of affected individuals (either affected men or affected women).
To assess whether there was significant heterogeneity between affected men and women in the families of the first family type (with affected men and women in each family), we used the setafftrait 1 -1 0 command to conduct the FBAT-e analyses. The results are summarized in Table 8. Two markers (rs28006 [T] and rs25957 [G]) exhibited significant P values (additive model: P = 0.004 for rs28006 and P = 0.017 for rs25957), suggesting that these two markers were associated with AS only in affected women in the subset of AS families with affected individuals of both genders.
We also conducted FBAT-e analysis using setafftrait command 1 -1 0 in families with only one gender of affected individuals (data not shown). However, there were few informative families (<20 families from which we could track the transmission of alleles), and so the results might not be reliable.
Selective transmission of haplotypes of interest to the affected men/women
In order to estimate the magnitude of the effect, we calculated the frequency at which the haplotypes of interest were transmitted to the affected men or women using TDT. For the haplotype rs28006 [C] rs25957 [C], the frequency of transmission was 74% (17/23) to affected women and 40% (12/30) to affected men. Thus, the 'odds ratio' for increased risk is 1.85 (0.74/0.4). More dramatic proportions were seen in the subset of families with affected individuals of both genders. In these families, this haplotype was transmitted to affected women 79% of the time (15/19) but to affected men only 27% of the time (3/11). In this case, the 'odds ratio' for increased risk approaches 3.0 (0.79/0.27 = 2.92).
For the haplotype rs26307 [C] rs27356 [C], the frequency of transmission was 70% (21/30) to affected men and 43% (13/30) to affected women (an 'odds ratio' for increased risk of 1.75). In the subset of families with only affected men, 94% (16/17) of the time this haplotype was transmitted to affected men. There were too few informative families with only affected women with this variant (n = 6), and so we do not have a reliable assessment of the frequency at which this haplotype was transmitted to affected women in this subset for comparison.
Discussion
In this study of the association of ANKH genetic markers with AS, including 201 AS multiplex families, we found that ANKH variants located in two different regions of the ANKH gene were associated with AS. A more striking finding was that the genetic association for men and women with AS differed. In men, AS was associated with genetic markers at the 3' end of the ANKH gene, whereas in women AS appeared to be associated with genetic markers at the 5' end of the ANKH gene. As expected, when the genders of AS patients were analyzed separately, we observed more than one SNP in each region (within the same haplotype block) showing significant association with AS. Haplotype analyses appeared to confirm the results of the single-marker tests (FBAT analyses), indicating that the predisposing polymorphism(s) for men with AS probably lies at the 3' end of the ANKH gene, whereas those for affected women are probably at the 5' end of the gene. After Bonferroni correction for the number of haplotypes and models, the haplotype combination of rs26307 [C] and rs27356 [C] was significantly associated with men with AS; and the haplotype combination of rs28006 [C] and rs25957 [C] was significantly associated with women with AS. However, in both cases, the significance level was modest. We attribute this to the fact that we have not identified the aetiological variants in the men/women with AS. Despite the modest P values (which are a function of sample size), the calculated 'odds ratios' for increased risk (which provide estimates of the magnitude of the effect) were close to 2 for the transmission of rs26307 [C] rs27356 [C] to affected men, and close to 3 for the transmission of rs28006 [C] rs25957 [C] to affected women in the subset of families with affected individuals of both genders.
A test of interaction identified the region that was associated in men with AS (rs26307) as showing a difference in the strength of the association by gender. The region associated with AS in women only showed significance of the test of interaction among the subset of families with affected individuals of both genders. Our current efforts are to identify and analyze more common SNPs in these two regions, ultimately finding the predisposing polymorphisms in men and women.
The rationale for studying multiplex AS families is to enhance the chances of identifying the genes involved. There are very few studies that directly compare familial versus sporadic AS. In one study [26], familial versus sporadic Dutch AS patients exhibited no difference in age at disease onset, age at diagnosis, or prevalence of peripheral arthritis and acute anterior uveitis. In another study, familial AS disease was significantly milder than sporadic disease, as assessed by spinal mobility score, Arthritis Impact Measurement Scales (AIMS) overall impact score, AIMS physical activity score, AIMS social function score and AIMS pain score [27]. Thus, findings from multiplex families might not be directly applicable to individuals affected with sporadic AS. Most studies assessing the impact gender has on age at AS onset or diagnosis have been conducted without addressing whether the individuals had familial or sporadic disease [28-30]; these studies showed that the age at disease onset is similar between genders. However, in our cohort of AS multiplex families, men had a significantly earlier age at diagnosis compared with that for women (for men 28 ± 11 years [n = 213]; and for women 30 ± 11 years [n = 149]). Because these are AS multiplex families, it is unlikely that there is a bias leading physicians to delay diagnosis in affected women. The misconception that AS is exclusively a male disease may yet be a confounding factor. In the subset of families with affected individuals of both genders, men have an even earlier age at diagnosis (27.8 ± 11 years [n = 94]) compared with that for women (32.6 ± 11 years [n = 101]), whereas both men and women have similar ages of diagnosis in the subsets of families with affected individuals of only one gender (for men 28.6 ± 12 years [n = 130]; for women 29.9 ± 12 years [n = 53]). This finding suggests that there is heterogeneity even in multiplex AS families.
The ANKH variants that were significantly associated with AS are located in introns 1, 8 and 12. It is likely that the predisposing polymorphisms affect gender-specific regulation of ANKH expression. Very little is known regarding the molecular mechanisms that underlie the regulation of ANKH expression. One study [31] reported that ANKH is a growth factor responsive gene. Three recent reports [22-24] showed that ANKH is an androgen responsive gene. In androgen-treated prostate cancer cell lines, the abundance of ANKH transcripts was sixfold higher than in the untreated cells. In the ANKH promoter, there is a sequence at position -1015 (AGAACAcacTtTcCT) with 83% match to an androgen response element (ARE) consensus sequence [22]. It remains unclear whether this ARE-like motif is functional. In view of the locations of the ANKH variants associated with AS, it remains unclear whether this ARE-like motif at the promoter region can directly contribute to the regulation of ANKH expression by the predisposing polymorphisms. It is also unknown whether there is a different mode of ANKH regulation in women.
A report recently concluded that ANKH did not significantly contribute to susceptibility or specific disease expression in AS patients from the UK [32]. In that report, a case–control study was conducted using five ANKH SNPs within the coding region and flanking splice sites and three known promoter variants. There was no association between these polymorphisms and AS or the clinical pattern of the disease. In addition, using 185 affected sib-pair AS families, no linkage between ANKH and AS was observed. However, the exact linkage results were not shown. Using multipoint exclusion mapping of the ANKH region, the presence of a gene contributing more than 10% of the recurrence risk to AS (λS = 1.4) was excluded. Using λS of 1.4 as the cutoff may exclude genes with modest effects. In that report, the LD between markers was not shown. In situations where the aetiological variant is not typed, haplotype-based analysis may be a more powerful analytical method when there is significant LD.
The basis for the discrepancy between the UK results [32] and ours is not entirely clear, but there are several possible explanations. First, the UK group focused on analyzing exonic variants, variants near splice junctions and in the promoter region. Their analysis did not include any ANKH variants in the 3' region, where we detected association with AS in men. Second, although the UK group included a gender breakdown of their patients (63.5% men and 36.5% women), the analysis did not include a breakdown of AS patients by gender, and variants with modest gender-specific effects might have been missed. Third, it is possible that there are some intrinsic differences between the two populations (UK versus North American). Genome-wide linkage scans performed in the two groups revealed some similar susceptibility regions, such as on chromosomes 6p (the MHC), 5q and 10q [5,6]. However, the linkage identified on chromosome 11q23 in the North American Spondylitis Consortium families was not seen in the UK study. In addition, the linkage identified on chromosome 2q in the UK study was not seen in the North American Spondylitis Consortium study. The intrinsic differences could reflect clinical differences in the patient population recruited, or they could be due to population-specific mechanisms of genetic susceptibility. Finally, because both groups analyzed about 200 AS families, there might not be sufficient power to detect genes with 'small effects' consistently, leading to discrepancies between results.
In our cohort of North American multiplex AS families, the age at diagnosis was significantly younger in men than in women. However, FBAT analyses using the offset option (-o; an option which works for both quantitative and qualitative traits) did not show any significant association of age at diagnosis in the men or women with AS, even in subsets of families, using the ANKH markers (data not shown), suggesting that ANKH variants are responsible for disease susceptibility. Our finding of gender-specific polymorphisms in the ANKH gene conferring differential susceptibility to AS might shed light on the biological basis of these clinical observations.
In view of the difficulty in locating susceptibility loci with modest effects in recent genome-wide linkage studies conducted in AS families, it will be of interest to assess whether gender subsetting in the analyses of genome-wide linkage studies might yield further insights into the genetic basis of rheumatic diseases, many of which have a strong gender predilection.
Conclusion
Taken together, our findings showed that, after Bonferroni correction, two intronic markers at the 3' end of the ANKH gene were significantly associated with AS only in affected men, and two intronic markers at the 5' end of the ANKH gene were significantly associated with AS only in affected women. This may partly account for the gender difference in the prevalence of AS.
Abbreviations
AIMS = Arthritis Impact Measurement Scales; ARE = androgen response element; AS = ankylosing spondylitis; bp = base pairs; FBAT = family-based association testing; HBAT = haplotype-based association testing; LD = linkage disequilibrium; MHC = major histocompatibility complex; PPi = inorganic pyrophosphate; SNP = single nucleotide polymorphism; TDT = transmission disequilibrium test.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
HWT conducted all of the genotyping and analyzed the data. RDI conceived the study, provided some of the patients' blood/cells for extracting DNA and reviewed the manuscript. ADP designed the study, supervised the statistical analyses and revised the manuscript. JDR coordinated the recruitment of individuals from AS families, provided most of the DNA samples and reviewed the manuscript. FWLT conceived, designed and coordinated the study, analyzed and interpreted the data, performed statistical analyses, and drafted and revised the manuscript. All authors read and approved the final manuscript.
Acknowledgements
We thank Dr Cathy Barr for making the ABI sequence detection system available, Karen Wigg for reading the plates, and Dr Celia Greenwood for review of the manuscript and helpful suggestions on the statistical analyses. We also thank the Ontario Spondylitis Association and the Spondylitis Association of America for their assistance in recruiting the families included in this study.
This work was supported by grants from the Canadian Institutes of Health Research, the Arthritis Center of Excellence, Genome Canada, and by the NIH (National Institute of Arthritis and Musculoskeletal and Skin Diseases grant R01-AR-46208 to Dr Reveille). Dr Paterson is a Canada Research Chair.
Figures and Tables
Figure 1 Locations and spacings of genetic markers used for genotyping. D5S1991 and ANKH-OR are located at the 5' flanking region of ANKH. All seven single nucleotide polymorphisms used are located in the introns of ANKH.
Figure 2 Age of diagnosis for (a) men and (b) women in the North American multiplex ankylosing spondilitis families.
Figure 3 A summary of family-based association analyses (FBAT; additive model and biallelic tests). Each column in the charts represents the -log FBAT P value for each marker located at a distance relative to one another. D5S1953 is positioned at 0 distance. Chart ALL shows analysis of 226 ankylosing spondylitis (AS) nuclear families with 894 persons. Chart Women shows analysis of 127 AS families with 184 affected women. Chart Men shows analysis of 168 AS families with 282 affected men.
Table 1 Characteristics of 226 nuclear families included in the family-based association studies
Number of affected siblings Number of unaffected siblings Number of unaffected parents Number of affected parents Number of nuclear families
2 0 0 0 71
2 0 2 0 33
2 0 1 0 14
2 0 1 1 5
2 1 0 0 4
2 2 0 0 3
2 1 0 1 3
2 1 2 0 2
2 2 2 0 1
3 0 0 0 9
3 0 2 0 6
3 0 1 0 4
3 0 1 1 2
3 1 0 0 1
3 1 1 0 1
4 0 0 0 1
4 0 2 0 2
4 0 1 0 1
4 1 2 0 1
1 0 2 0 25
1 0 1 1 8
1 0 1 0 2
1 0 0 1 7
1 1 0 0 3
1 1 0 1 3
1 1 2 0 1
1 1 1 1 1
1 2 0 0 1
1 2 0 1 1
1 2 2 0 1
1 3 0 0 1
1 3 2 0 1
0 2 0 0 1
0 2 1 1 2
0 1 0 1 3
Table 2 Gender information for affected individuals in the 201 ankylosing spondylitis families
Number of affected men/women in a family Number of families
Families with both affected men and women 94
1/1 60
2/1 17
3/1 1
1/2 8
2/2 3
3/2 1
1/3 2
2/3 2
Families with only affected men 74
1/0 5
2/0 54
3/0 12
4/0 2
5/0 1
Families with only affected women 33
0/1 2
0/2 26
0/3 4
0/4 1
Table 3 FBAT-e analyses conducted in 226 ankylosing spondylitis nuclear families (201 pedigrees, 894 persons)
Marker Allele Allele frequency Number of informative families Z score P
Additive model: biallelic test
D5S1953 2 0.45 54 0.57 0.569
rs26307 C 0.81 35 1.73 0.084
rs27356 C 0.80 39 2.54 0.011*
3088132 G 0.79 28 1.28 0.198
rs153929 A 0.76 48 1.61 0.107
rs258215 A 0.59 43 1.58 0.113
rs28006 C 0.74 34 1.62 0.104
rs25957 C 0.76 36 2.02 0.043*
ANKH-OR 1 0.47 59 1.62 0.105
D5S1991 2 0.48 58 1.33 0.181
D5S1954 1 0.64 54 0.65 0.515
Recessive model: biallelic test
D5S1953 2 0.43 48 0.45 0.683
rs26307 C 0.80 40 1.49 0.137
rs27356 C 0.79 44 2.32 0.020*
3088132 G 0.79 31 1.19 0.233
rs153929 A 0.76 50 1.86 0.062
rs258215 A 0.58 32 1.55 0.120
rs28006 C 0.74 37 1.46 0.142
rs25957 C 0.76 38 1.65 0.098
ANKH-OR 2 0.52 37 -2.20 0.027*
D5S1991 1 0.52 45 -1.96 0.050*
D5S1954 1 0.64 54 0.65 0.517
*Statistically significant findings. FBAT, family-based association testing.
Table 4 FBAT-e analyses using setafftrait 0 -1 0, testing specifically for affected women
Marker Allele Allele frequency Number of informative families Z score P
Additive model: biallelic test
D5S1953 1 0.56 47 0.47 0.635
rs26307 T 0.19 22 0.32 0.751
rs27356 T 0.20 24 1.03 0.302
3088132 C 0.21 20 0.37 0.712
rs153929 G 0.24 46 1.31 0.191
rs258215 T 0.41 30 1.54 0.123
rs28006 T 0.26 31 2.82 0.004*
rs25957 G 0.23 23 2.82 0.004*
ANKH-OR 2 0.52 59 0.94 0.347
D5S1991 1 0.52 53 1.10 0.270
D5S1954 2 0.36 45 1.38 0.168
Recessive model: biallelic test
D5S1953 2 0.44 39 -1.21 0.227
rs26307 C 0.81 26 0.52 0.606
rs27356 T 0.20 10 1.81 0.069
3088132 G 0.79 22 0.08 0.930
rs153929 A 0.76 44 -1.39 0.162
rs258215 A 0.58 21 -1.76 0.077
rs28006 C 0.74 28 -2.49 0.012*
rs25957 C 0.76 23 -2.25 0.024*
ANKH-OR 2 0.52 34 1.14 0.254
D5S1991 1 0.51 33 1.52 0.127
D5S1954 1 0.64 45 -1.54 0.122
*Statistically significant findings. FBAT, family-based association testing.
Table 5 FBAT-e analyses using setafftrait 1 0 0, testing specifically for affected men
Marker Allele Allele frequency Number of informative families Z score P
Additive model: biallelic test
D5S1953 2 0.44 47 0.25 0.805
rs26307 C 0.80 22 2.06 0.039*
rs27356 C 0.79 25 2.63 0.008*
3088132 G 0.79 18 1.23 0.216
rs153929 A 0.76 36 0.49 0.619
rs258215 A 0.58 32 0.81 0.419
rs28006 T 0.26 37 0.15 0.877
rs25957 C 0.77 33 0.76 0.446
ANKH-OR 1 0.48 64 1.04 0.297
D5S1991 2 0.49 64 0.39 0.696
D5S1954 2 0.36 49 0.33 0.736
Recessive model: biallelic test
D5S1953 1 0.56 40 -0.83 0.406
rs26307 C 0.81 25 2.51 0.012*
rs27356 C 0.80 28 2.99 0.002*
3088132 G 0.79 22 1.44 0.149
rs153929 A 0.76 39 0.71 0.477
rs258215 A 0.59 23 0.60 0.543
rs28006 T 0.26 16 0.33 0.739
rs25957 C 0.76 29 0.64 0.518
ANKH-OR 2 0.52 44 -1.26 0.205
D5S1991 1 0.51 46 -0.71 0.472
D5S1954 1 0.64 43 -0.65 0.510
*Statistically significant findings. FBAT, family-based association testing.
Table 6 HBAT-e analyses using ANKH markers in four haplotype blocks defined in HapMap
Markers in the haplotype block AS nuclear families (n = 226) Testing specifically for affected women Testing specifically for affected men
Additive model
rs26307, rs27356 [C,C]; 0.78; 39; 0.04 NS [C,C]; 0.79; 24; 0.014
3088132, rs153929 NS NS NS
rs28006, rs25957 NS [T,G]; 0.25; 19; 0.007 NS
ANKH-OR, D5S1991 [1,2]; 0.43; 57; 0.013 NS NS
Recessive/dominant model
rs26307, rs27356 [C,C]; 0.78; 40; 0.02 NS [C,C]; 0.79; 25; 0.004*
3088132, rs153929 [G,A]; 0.70; 32; 0.02 NS NS
rs28006, rs25957 NS [C,C]; 0.71; 18; 0.004* NS
ANKH-OR, D5S1991 NS NS NS
Data are expressed as [allele]; allele frequency; number of informative families; P value.
*Significant P value after Bonferroni correction. (Because eight tests [four haplotypes and two models] were carried out in the haplotype-based association testing [HBAT]-e analyses, P < 0.00625 [0.05/8] is considered statistically significant.) AS, ankylosing spondylitis; NS, not significant.
Table 7 FBAT-e analyses considering 226 ankylosing spondylitis nuclear families (201 pedigrees, 894 persons): summary of results using setafftrait 1 -1 0
Marker Allele Allele frequency Number of informative families Z score P
Additive model: biallelic test
D5S1953 1 0.56 63 0.15 0.879
rs26307 C 0.81 35 1.29 0.195
rs27356 C 0.80 39 1.12 0.261
3088132 G 0.79 29 0.69 0.489
rs153929 G 0.24 52 0.61 0.541
rs258215 T 0.41 41 0.46 0.638
rs28006 T 0.26 44 1.69 0.090
rs25957 G 0.23 39 0.90 0.366
ANKH-OR 1 0.48 80 0.20 0.841
D5S1991 1 0.52 80 0.29 0.769
D5S1954 2 0.36 63 1.21 0.227
Recessive model: biallelic test
D5S1953 1 0.56 46 -0.89 0.370
rs26307 C 0.81 37 2.17 0.030*
rs27356 C 0.80 41 1.91 0.055
3088132 G 0.79 29 1.17 0.240
rs153929 G 0.24 18 0.55 0.582
rs258215 A 0.59 29 -0.75 0.453
rs28006 C 0.74 40 -1.13 0.186
rs25957 C 0.77 36 -0.66 0.507
ANKH-OR 2 0.52 50 -0.34 0.733
D5S1991 2 0.49 52 -0.28 0.779
D5S1954 1 0.64 61 -1.53 0.126
*Statistically significant findings. FBAT, family-based association testing.
Table 8 FBAT-e analyses considering 108 ankylosing spondylitis nuclear families (94 pedigrees, 425 persons) in which both affected men and women are present in each family: summary of the results using setafftrait 1 -1 0
Marker Allele Allele frequency Number of informative families Z score P
Additive model: biallelic test
D5S1953 1 0.56 39 0.77 0.437
rs26307 C 0.90 10 0.50 0.617
rs27356 C 0.77 17 0.15 0.875
3088132 C 0.21 13 0.63 0.527
rs153929 G 0.23 32 0.81 0.418
rs258215 T 0.49 21 1.62 0.104
rs28006 T 0.29 27 2.81 0.004*
rs25957 G 0.32 22 2.37 0.017*
ANKH-OR 1 0.48 51 0.25 0.801
D5S1991 2 0.50 48 0.13 0.891
D5S1954 2 0.33 31 1.13 0.254
Recessive model: biallelic test
D5S1953 2 0.44 31 -1.34 0.177
rs26307 C 0.79 19 1.00 0.314
rs27356 C 0.80 20 1.17 0.239
3088132 G 0.79 14 -0.26 0.788
rs153929 G 0.23 10 1.06 0.288
rs258215 A 0.51 12 -2.49 0.012*
rs28006 C 0.71 20 -2.344 0.019*
rs25957 C 0.67 15 -1.97 0.048*
ANKH-OR 1 0.48 27 0.83 0.406
D5S1991 2 0.49 29 0.98 0.322
D5S1954 1 0.66 31 -1.67 0.093
*Statistically significant findings. FBAT, family-based association testing.
==== Refs
Calin A Maddison PJ, Isenberg DA, Woo P, Glass DN Ankylosing spondylitis Oxford Textbook of Rheumatology 1998 2 Oxford: Oxford University Press 1058 1070
Braun J Bollow M Remlinger G Eggens U Rudwaleit M Distler A Sieper J Prevalence of spondylarthropathies in HLA-B27 positive and negative blood donors Arthritis Rheum 1998 41 58 67 9433870 10.1002/1529-0131(199801)41:1<58::AID-ART8>3.0.CO;2-G
Brown MA Laval SH Brophy S Calin A Recurrence risk modeling of the genetic susceptibility to ankylosing spondylitis Ann Rheum Dis 2000 59 883 886 11053066 10.1136/ard.59.11.883
Brown MA Pile KD Kennedy LG Campbell D Andrew L March R Shatford JL Weeks DE Calin A Wordsworth BP A genome-wide screen for susceptibility loci in ankylosing spondylitis Arthritis Rheum 1998 41 588 595 9550467 10.1002/1529-0131(199804)41:4<588::AID-ART5>3.0.CO;2-0
Laval SH Timms A Edwards L Bradbury L Brophy S Milicic A Rubin L Siminovitch KA Weeks DE Calin A Whole-genome screening in ankylosing spondylitis: evidence of non-MHC genetic-susceptibility loci Am J Hum Genet 2001 68 918 926 11231900 10.1086/319509
Zhang G Luo J Bruckel J Weisman MA Schumacher HR Kahn MA Inman RD Mahowald M Maksymowych WP Martin TM Genetic studies in familial ankylosing spondylitis susceptibility Arthritis Rheum 2004 50 2246 2254 15248224 10.1002/art.20308
Ho AM Johnson MD Kingsley DM Role of the mouse ank gene in control of tissue calcification and arthritis Science 2000 289 265 270 10894769 10.1126/science.289.5477.265
Tsui FWL Tsui HW Cheng EY Stone M Payne U Reveille JD Paterson AD Inman RD Novel genetic markers in the 5'-flanking region of ANKH are associated with ankylosing spondylitis Arthritis Rheum 2003 48 791 797 12632434 10.1002/art.10844
Calin A Ankylosing spondylitis Clin Rheum Dis 1985 11 41 60 3158467
Eustace S Coughlan RJ McCarthy C Ankylosing spondylitis. A comparison of clinical and radiographic features in men and women Ir Med J 1993 86 120 122 8360039
Gran JT Husby G Hordvik M Stormer J Romberg-Andersen O Radiological changes in men and women with ankylosing spondylitis Ann Rheum Dis 1984 43 570 575 6332580
Will R Edmunds L Elswood J Calin A Is there sexual inequality in ankylosing spondylitis? A study of 498 women and 1202 men J Rheumatol 1990 17 1649 1652 2084239
Kidd B Mullee M Frank A Cawley M Disease expression of ankylosing spondylitis in males and females J Rheumatol 1988 15 1407 1409 3199401
Guillemin F Briancon S Pourel J Gaucher A Longterm disability and prolonged sick leave as outcome measures in ankylosing spondylitis Arthritis Rheum 1990 33 1001 1006 2369416
Doran MF Brophy S MacKay K Taylor G Calin A Predictors of longterm outcome in ankylosing spondylitis J Rheum 2003 30 316 320 12563688
Hoyle E Laval SH Calin A Wordsworth BP Brown MA The X-chromosome and susceptibility to ankylosing spondylitis Arthritis Rheum 2000 43 1353 1355 10857794 10.1002/1529-0131(200006)43:6<1353::AID-ANR19>3.0.CO;2-B
Calin A Brophy S Blake D Impact of sex on inheritance of ankylosing spondylitis: a cohort study Lancet 1999 354 1687 1690 10568571 10.1016/S0140-6736(99)03219-5
Brophy S Taylor G Blake D Calin A The interrelationship between sex, susceptibility factors, and outcome in ankylosing spondylitis and its associated disorders including inflammatory bowel disease, psoriasis and iritis J Rheumatol 2003 30 2054 2058 12966615
Van der Linden S Valkenburg H Cats A Evaluation of diagnostic criteria for ankylosing spondylitis: a proposal for modification of the New York criteria Arthritis Rheum 1984 27 361 368 6231933
Spielman RS McGinnis RE Ewens WJ Transmission test for linkage disequilibrium: the insulin gene region and insulin-dependent diabetes mellitus (IDDM) Am J Hum Genet 1993 52 506 516 8447318
Lake SL Blacker D Laird NM Family-based tests of association in the presence of linkage Am J Hum Genet 1993 52 506 516 8447318
Nelson PS Clegg N Arnold H Ferguson C Bonham M White J Hood L Lin B The program of androgen-responsive genes in neoplastic prostate epithelium Proc Natl Acad Sci USA 2002 99 11890 11895 12185249 10.1073/pnas.182376299
DePrimo SE Diehn M Belson JB Reiter RE Matese J Fero M Tibshirani R Brown PO Brooks JD Transcriptional programs activated by exposure of human prostate cancer cells to androgen Genome Biol 2002 3 32.1 32.12 10.1186/gb-2002-3-7-research0032
Segawa T Nau ME Xu LL Chilukuri RN Makaren M Zhang W Petrovics G Sesterhenn IA McLeod DG Moul JW Androgen-induced expression of endoplasmic reticulum (ER) stress response genes in prostate cancer cells Oncogene 2002 21 8749 8758 12483528 10.1038/sj.onc.1205992
The international HapMap Consortium The international HapMap project Nature 2003 426 789 795 14685227 10.1038/nature02168
Paardt M Dijkmans B Giltay E vander Horst-Bruinsma I Dutch patients with familial and sporadic ankylosing spondylitis do not differ in disease phenotype J Rheumatol 2002 29 2583 2584 12465156
Calin A Kennedy LG Edmunds L Will R Familial versus sporadic ankylosing spondylitis. Two different diseases? Arthritis Rheum 1993 36 676 681 8489546
Van der Linden SM Valkenburg HA de Jongh BM Cats A The risk of developing ankylosing spondylitis in HLA-B27 positive individuals Arthritis Rheum 1984 27 241 249 6608352
Gran JT Ostensen M Husby G A clinical comparison between males and females with ankylosing spondylitis J Rheumatol 1985 12 126 129 3872366
Feldtkeller E Khan MA van der Heijde Désirée van der Linden S Braun J Age at disease onset and diagnosis delay in HLA-B27 negative vs. positive patients with ankylosing spondylitis Rheumatol Int 2003 23 61 66 12634937
Guo Y Hsu DKW Feng SY Richards CM Winkles JA Polypeptide growth factors and phorbol ester induce progressive ankylosis (ank) gene expression in murine and human fibroblasts J Cell Biochem 2001 84 27 38 11746513 10.1002/jcb.1263
Timms AE Zhang Y Bradbury L Wordsworth BP Brown MA Investigation of the role of ANKH in ankylosing spondylitis Arthritis Rheum 2003 48 2898 2902 14558096 10.1002/art.11258
| 15899038 | PMC1174945 | CC BY | 2021-01-04 16:02:36 | no | Arthritis Res Ther. 2005 Feb 25; 7(3):R513-R525 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1701 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17021589903710.1186/ar1702Research ArticleHistone deacetylase inhibitors modulate metalloproteinase gene expression in chondrocytes and block cartilage resorption Young David A [email protected] Rachel L [email protected] Caroline J [email protected] Debra [email protected] Lara [email protected] Dylan R [email protected] Timothy E [email protected] Ian M [email protected] School of Biological Sciences, University of East Anglia, Norwich, UK2 Department of Rheumatology, University of Newcastle-upon-Tyne, Newcastle-upon-Tyne, UK3 Department of Rheumatology, University of Newcastle-upon-Tyne, Newcastle-upon-Tyne, UK2005 22 2 2005 7 3 R503 R512 11 11 2004 23 12 2004 7 1 2005 25 1 2005 Copyright © 2005 Young et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Cartilage destruction in the arthritides is thought to be mediated by two main enzyme families: the matrix metalloproteinases (MMPs) are responsible for cartilage collagen breakdown, and enzymes from the ADAMTS (a disintegrin and metalloproteinase domain with thrombospondin motifs) family mediate cartilage aggrecan loss. Many genes subject to transcriptional control are regulated, at least in part, by modifications to chromatin, including acetylation of histones. The aim of this study was to examine the impact of histone deacetylase (HDAC) inhibitors on the expression of metalloproteinase genes in chondrocytes and to explore the potential of these inhibitors as chondroprotective agents. The effects of HDAC inhibitors on cartilage degradation were assessed using a bovine nasal cartilage explant assay. The expression and activity of metalloproteinases was measured using real-time RT-PCR, western blot, gelatin zymography, and collagenase activity assays using both SW1353 chondrosarcoma cells and primary human chondrocytes. The HDAC inhibitors trichostatin A and sodium butyrate potently inhibit cartilage degradation in an explant assay. These compounds decrease the level of collagenolytic enzymes in explant-conditioned culture medium and also the activation of these enzymes. In cell culture, these effects are explained by the ability of HDAC inhibitors to block the induction of key MMPs (e.g. MMP-1 and MMP-13) by proinflammatory cytokines at both the mRNA and protein levels. The induction of aggrecan-degrading enzymes (e.g. ADAMTS4, ADAMTS5, and ADAMTS9) is also inhibited at the mRNA level. HDAC inhibitors may therefore be novel chondroprotective therapeutic agents in arthritis by virtue of their ability to inhibit the expression of destructive metalloproteinases by chondrocytes.
==== Body
Introduction
Articular cartilage is made up of two main extracellular-matrix (ECM) macromolecules, namely, type II collagen and aggrecan (a large, aggregating proteoglycan) [1,2]. The type II collagen scaffold endows the cartilage with its tensile strength, while the aggrecan, by virtue of its high negative charge, draws water into the tissue, swelling against the collagen network, and enabling the tissue to resist compression. Quantitatively more minor components (e.g. types IX, XI, and VI collagens; biglycan; decorin; cartilage oligomeric matrix protein; etc.) also have important roles in controlling matrix structure and organisation [2].
Normal cartilage ECM is in a state of dynamic equilibrium, with a balance between synthesis and degradation. For the degradative process, the major players are metalloproteinases that degrade the ECM, and their inhibitors. Pathological cartilage destruction can therefore be viewed as a disruption of this balance, favouring proteolysis.
The matrix metalloproteinases (MMPs) are a family of 23 enzymes in man that facilitate turnover and breakdown of the ECM in both physiology and pathology. The MMP family contains the only mammalian proteinases that can specifically degrade the collagen triple helix at neutral pH. These include the 'classical' collagenases – MMP-1, -8, and -13 – and also MMP-2 and MMP-14 (which cleave the triple helix with less catalytic efficiency). The enzyme(s) responsible for cartilage collagen cleavage in the arthritides remains open to debate [3].
A second group of metalloproteinases, the ADAMTS (a disintegrin and metalloproteinase domain with thrombospondin motifs) family, consists of 19 members, including the so-called 'aggrecanases', currently ADAMTS-1, -4, -5, -8, -9, and -15 [4-7]. Current data support the hypothesis that aggrecanases are active early in the disease process, with later increases in MMP activity (several MMPs can also degrade aggrecan), but the exact enzyme(s) responsible for cartilage aggrecan destruction at any stage in arthritis is unclear [3,8,9].
A family of four specific inhibitors, the tissue inhibitors of metalloproteinases (TIMPs), has been described. TIMPs are endogenous inhibitors of MMPs and potentially of ADAMTSs [10]. The ability of TIMP-1 to -4 to inhibit active MMPs is largely promiscuous, though a number of functional differences have been uncovered. TIMP-3 appears to be the most potent inhibitor of ADAMTSs, for example, with a subnanomolar Ki against ADAMTS-4 [3].
Metalloproteinase activity is regulated at multiple levels, including gene transcription. However, the role of chromatin modification, and in particular acetylation, is little researched in the metalloproteinase arena. The packaging of eukaryotic DNA into chromatin plays an important role in regulating gene expression. The DNA is wound round a histone octamer consisting of two molecules each of histones H2A, H2B, H3, and H4, to form a nucleosome [11]. This unit is repeated at intervals of approximately 200 base pairs, with histone H1 associating with the intervening DNA. Nucleosomes are generally repressive to transcription, hindering access of the transcriptional apparatus [11]. However, two major mechanisms modulate chromatin structure to allow transcriptional activity: ATP-dependent nucleosome remodellers such as the Swi/Snf complex [12,13]; and the enzymatic modification of histones, via acetylation, methylation, and phosphorylation [14-16].
Acetylation by histone acetyltransferases occurs on specific lysine residues on the N-terminal tails of histones H3 and H4. This neutralisation of positive charge leads to a loosening of the histone:DNA structure, allowing access of the transcriptional machinery; furthermore, the acetyl groups may associate with and recruit factors containing bromodomains [11]. Many transcriptional activators or coactivators have (or recruit) histone acetyltransferase activity, giving a mechanism whereby acetylation can be targeted at specific gene promoters [15,16]. Conversely, histone deacetylases (HDACs) have also been characterised. Hypoacetylation of histones associates with transcriptional silence, and several transcriptional repressors and corepressors have been identified that have (or recruit) HDAC activity [17-19]. Nonhistone substrates of histone acetyltransferases have also been described, for example, p53, E2F, nuclear factor κB, Sp3, and c-Jun [20,21].
There are two families of HDACs, the NAD+-dependent, so-called SIR2 family (sometimes called class III HDACs), and the classical HDAC family. The classical HDACs can be grouped into three classes (I, II, and IV) based on phylogeny [22]. Class I HDACs (HDAC1, 2, 3, and 8) are related to yeast RPD3, and class II HDACs (HDAC4, 5, 6, 7, 9, and 10) are more closely related to yeast HDA1 [17]. HDAC11 alone represents class IV, and HDAC11-related proteins have been described in all eukaryotic organisms other than fungi [22]. Trichostatin A (TSA) and sodium butyrate (NaBy), are HDAC inhibitors [23,24] with a broad spectrum of activity against class I and II HDACs, but not the SIR2 family. Addition of these reagents to cells should therefore block histone deacetylation and result in increased acetylation of histones on susceptible genes. The prediction would be that this would lead to an increase in gene expression, and this is largely borne out experimentally. However, there are many instances of HDAC inhibitors acting as repressors of gene expression [25-29].
HDAC inhibitors have potent antiproliferative and pro-apoptotic activities in cancer cells and this has led to the development of specific inhibitors for cancer chemotherapy. Such compounds are currently in both preclinical development and clinical trials [30].
Two recent reports demonstrate that HDAC inhibitors modulate gene expression in synovial cells [31]. In an animal model of rheumatoid arthritis (adjuvant arthritis), the expression of tumour necrosis factor α (TNF-α) was inhibited and this led to a reduction in synovial hyperplasia and joint swelling with maintenance of joint integrity [31]. Similar results were obtained in an autoantibody-mediated murine model [32]. However, no study has looked at the effect of these inhibitors on cartilage. Here, we show for the first time that HDAC inhibitors repress the expression of several members of the metalloproteinase family in chondrocytes and block cartilage destruction in an ex vivo model system. Hence, inhibition of HDAC activity offers a potential therapeutic strategy to prevent cartilage destruction in the arthritides.
Materials and methods
Cell culture
SW1353 human chondrosarcoma cells (ATCC, USA) were routinely cultured in DMEM (Invitrogen, Paisley, UK) containing 10% fetal bovine serum (Invitrogen), 2 mM glutamine, 100 IU/ml penicillin, and 100 μg/ml streptomycin. Serum-free conditions used identical medium without fetal bovine serum. For assays, cells were grown to confluency, then starved of serum for 24 hours before the addition of IL-1α (R&D Systems) (5 ng/ml) and oncostatin M (OSM) (R&D Systems) (10 ng/ml) in the absence or presence of HDAC inhibitors (TSA, 50 to 500 ng/ml, and NaBy, 1 to 10 mM) (Calbiochem, Nottingham, UK). Experiments were performed in six-well plates (Nunc, Fisher Scientific, Loughborough, UK) with all conditions in duplicate or triplicate. To obtain primary human chondrocytes, fresh human articular cartilage samples (from patients undergoing hip or knee replacement surgery at the Norfolk and Norwich University Hospital) were digested overnight in DMEM containing 2 mg/ml of collagenase Type 1A (Sigma, Poole, UK). The resulting cells were washed with PBS, resuspended in DMEM containing 10% FCS and antibiotics as above, and then plated at 1 × 106 cells in 75-cm2 flasks. At confluence, cells were passaged and replated at 1:2 dilution.
RNA isolation and synthesis of cDNA
RNA was isolated from monolayer cultures using Trizol reagent (Invitrogen). cDNA was synthesised from 1 μg of total RNA using Superscript II reverse transcriptase (Invitrogen) and random hexamers in a total volume of 20 μl according to the manufacturer's instructions. cDNA was stored at -20°C until use in downstream PCR.
RT-PCR
For quantitative real-time PCR, sequences and validation for MMP and TIMP primers and probes are as previously described [33] and so are ADAMTS primers and probes [34]. In order to control against amplification of genomic DNA, primers were placed within different exons close to an intron/exon boundary with the probe spanning two neighbouring exons where possible. BLAST searches for all the primer and probe sequences were also conducted to ensure gene specificity. The 18 S ribosomal RNA gene was used as an endogenous control to normalise for differences in the amount of total RNA present in each sample; 18 S rRNA primers and probe were purchased from Applied Biosystems (Warrington, UK).
Relative quantification of genes was performed using the ABI Prism 7700 sequence detection system (Applied Biosystems) in accordance with the manufacturer's protocol. PCR reactions contained 5 ng of reverse transcribed RNA (1 ng for 18 S analyses), 50% TaqMan 2X Master Mix (Applied Biosystems), 100 nM of each primer, and 200 nM of probe in a total volume of 25 μl. Conditions for the PCR reaction were 2 min at 50°C, 10 min at 95°C, and then 40 cycles each consisting of 15 s at 95°C and 1 min at 60°C.
Conventional RT-PCR for collagen and aggrecan expression was as previously described [35].
Cartilage degradation assay
Bovine nasal cartilage was cultured as previously described [36]. Briefly, discs (approximately 2 mm in diameter by 1 to 2 mm thick) were punched from bovine nasal septum cartilage; three discs per well in a 24-well plate were incubated overnight in control, serum-free medium (DMEM containing 25 mM HEPES, 2 mM glutamine, 100 μg/ml streptomycin, 100 IU/ml penicillin, 2.5 μg/ml gentamicin, and 40 u/ml nystatin). Fresh control medium with or without test reagents (each condition in quadruplicate) was then added (day 0). Cartilage was incubated until day 7 and supernates were harvested and replaced with fresh medium containing the same test reagents as day 0. On day 14, supernates were harvested and the remaining cartilage was digested with papain. The viability of cartilage explants was assessed by measurement of lactate dehydrogenase (LDH) in the conditioned medium (CytoTox 96 assay, Promega, Southampton, UK). Hydroxyproline release was assayed as a measure of collagen degradation [37], and glycosaminoglycan release was assayed as a measure of proteoglycan degradation [36]. Collagenase activity was determined by the 3H-acetylated collagen diffuse fibril assay using a 96-well plate modification [38] and a standard curve and appropriate sample dilutions; one unit of collagenase activity degraded 1 μg of collagen per minute at 37°C. APMA (4-aminophenylmercuric acetate) was used to activate procollagenases [38]. Statistical analysis was performed using Student's t-test.
Gelatin zymography
Samples were electrophoresed under nonreducing conditions by SDS–PAGE in 10% polyacrylamide gels copolymerised with 1% gelatin. Gels were washed vigorously twice for 15 min in 2.5% Triton X-100 to remove SDS, then incubated overnight in 50 mM Tris/HCl, pH7.5, 5 mM CaCl2 at 37°C. Gels were then stained with Coomassie brilliant blue. Parallel gels were incubated in buffers containing either 5 mM EDTA or 2 mM 1,10-phenanthroline to show that lysis of gelatin was due to metalloproteinase activity.
Western blotting
Samples of conditioned culture medium were precipitated with an equal volume of ice-cold 10% w/v trichloroacetic acid. Precipitates were resuspened in loading buffer and electrophoresed under reducing conditions by SDS–PAGE in 10% polyacrylamide gels. Proteins were then transferred to an Immobilon P membrane (Millipore, Watford, UK) and probed with either rabbit anti-(human MMP-1), [39], sheep anti-(human MMP-1) [40], or sheep anti-(human MMP-13) [40].
Results
Histone deacetylase inhibitors block cartilage resorption and concurrently decrease collagen- and gelatin-degrading proteolytic activities
The combination of IL-1α and OSM has previously been shown to reproducibly and potently induce cartilage proteoglycan and collagen proteolysis both in vitro and in vivo [41,42]. The addition of TSA or NaBy to bovine nasal cartilage explant culture stimulated to resorb with IL-1α and OSM caused a dose-dependent (50 to 500 ng/ml TSA, 1 to 10 mM NaBy) inhibition of both proteoglycan and collagen release (at day 7 and 14 respectively) (Fig. 1). TSA is reported to have an IC50 (median inhibitory concentration) in the nanomolar range (50 ng/ml = 165 nM), but this does vary depending upon the HDAC and assay used (e.g. [43]); NaBy is reported to have an IC50 in themillimolar range. The need for TSA, a hydroxamate, to penetrate the highly negatively charged cartilage matrix will also raise the effective IC50 in the cartilage explant assay. The time points of medium collection, days 7 and 14, are those at which release of, respectively, proteoglycan and collagen are reproducibly close to 100%, since proteoglycan release is an earlier event in cartilage degradation. At these time points, proteoglycan release showed less sensitivity to HDAC inhibitors than collagen release, behaviour that may reflect its more rapid kinetics. Indeed, in a preliminary experiment at day 3, where IL-1α/OSM-induced proteoglycan release was approximately 50%, an increased sensitivity to HDAC inhibitors was seen (data not shown). Lactate dehydrogenase release, used as a measure of toxicity, was no greater in the presence of TSA or NaBy (at any concentration) than in the comparator control cultures (i.e. either without any addition or treated with IL-1α/OSM) at either day 7 or day 14; furthermore, no dose-dependent effects of TSA or NaBy on the release of lactate dehydrogenase were observed (data not shown).
Figure 2a shows collagenase activity at day 14 in the absence or presence of TSA, assayed in the conditioned medium from the explant assay discussed above. Treatment with IL-1α and OSM increased collagenase activity in the medium, and all collagenases were in the active form (since the addition of the procollagenase activator APMA did not lead to an increase in activity). The additional presence of TSA at the lowest dose (50 ng/ml) decreased the level of active collagenase, whereas total collagenase was unchanged; that is, the percentage of collagenase that was active was decreased (since the addition of APMA led to increased activity in the assay, demonstrating the presence of procollagenases). With increasing dose, TSA decreased the level of both active and total collagenase; that is, the total amount of collagenase in the medium was decreased and the percentage of this enzyme(s) that was activated also decreased. Similar results were obtained using NaBy (data not shown).
Figure 2b shows a gelatin zymogram of the day-14 medium cartilage-explant-conditioned medium in the absence or presence of TSA. Unstimulated explants produce a low constitutive level of gelatinolytic activity, which was probably due to proMMP-2. The addition of IL-1α and OSM induced three major gelatinolytic activities, which ran as poorly resolved doublets (all activities shown were blocked by metalloproteinase inhibitors EDTA and 1,10-phenanthroline, and were therefore due to the action of metalloproteinases; see Materials and methods). The largest of these probably equates to bovine MMP-9 (both pro- and active); there was an induction and activation of MMP-2 and an induction of a lower-molecular-weight activity that may represent collagenases MMP-1 and MMP-13, but could include other MMPs, many of which have at least some activity against gelatin. Both of the collagenases, and particularly MMP-13, have gelatinolytic activity [44,45], and this would therefore be in agreement with the induction of collagenase activity shown in Fig. 2. TSA at the lowest dose (50 ng/ml) caused a marked reduction in the lowest-molecular-weight activity, while increasing doses reduced the activities of all the gelatinolytic enzymes to background levels.
Histone deacetylase inhibitors modulate MMP gene expression
Using the SW1353 chondrosarcoma cell line as a model in which to look at the regulation of metalloproteinase and TIMP gene expression, we profiled the expression of all MMPs, ADAMTSs, and TIMPs in cells stimulated with IL-1α and OSM in the absence or presence of HDAC inhibitors at the doses used for the cartilage explant experiments above. Figure 3a shows typical responses for genes induced by IL-1α and OSM. A number of genes – MMP1, MMP3, MMP7, MMP8, MMP10, MMP12, MMP13, ADAMTS4, and ADAMTS9 – were robustly induced by the combination of IL-1α and OSM (though both ADAMTS4 and ADAMTS5 were expressed only at low levels in this cell line, with ADAMTS5 showing a weak induction with IL-1α and OSM). Of these induced genes, including ADAMTS5, all but ADAMTS4 showed repression by both TSA and NaBy. ADAMTS4, while strongly induced by IL-1α and OSM, was not repressed by either HDAC inhibitor in this cell line. The expression of a number of genes (MMP2, MMP9, MMP16, and MMP19; ADAMTS1, ADAMTS2, ADAMTS7, ADAMTS12, ADAMTS13, and ADAMTS20; TIMP3) was unaffected by the HDAC inhibitors, whereas the expression of several genes was induced by HDAC inhibitors alone (MMP17, MMP23, MMP28; ADAMTS15 and ADAMTS17; TIMP2). The varying response to HDAC inhibitors across the gene families also affirms that the compounds are not simply showing a nonspecific toxicity.
In order to verify that the effects of HDAC inhibitors were not specific to the SW1353 cell line, we undertook a similar experiment on a subset of genes, using primary articular chondrocytes isolated from both knee and hip joint (i.e. from two different donors). MMP1 and MMP13, the two major specific collagenases, were strongly induced by IL-1α and OSM, and this induction was repressed by both TSA (500 ng/ml) and NaBy (10 mM) (Fig. 3b). MMP8 was expressed at much lower levels in these cells but followed the same pattern of responses (data not shown). The IL-1α and OSM induction of MMP3 gene expression was only poorly repressed by HDAC inhibitors in the primary chondrocytes (data not shown). In these cells, ADAMTS4, ADAMTS5, and ADAMTS9 were all induced by IL-1α and OSM and the induction was repressed by HDAC inhibitors (Fig. 3c). These primary chondrocytes, although grown in monolayer culture, still express type II collagen and aggrecan at the passage at which this experiment was performed.
Histone deacetylase inhibitors repress MMP protein expression and activity
In order to ascertain if changes at the level of steady-state mRNA are mirrored at the protein level, we performed western blots on the conditioned medium of SW1353 cells at a 24-hour time point. Both MMP-1 and MMP-13 proteins were potently induced by treatment with IL-1α and OSM and this induction was repressed by both TSA and NaBy in the same manner as the mRNA (Fig. 4). Two different anti-MMP-1 antibodies (one raised in rabbits [39] and one raised in sheep [40]) cross-react with a protein of slightly lower Mr than MMP-1 in the SW1353-conditioned medium. The identity of this protein is unknown, but its expression has been previously documented [40] (though misinterpreted as that of active MMP-1), it is unaltered by the stimuli used, and it is not present in conditioned medium from primary chondrocytes (data not shown).
Gelatin zymography showed some induction of MMP-9, as well as multiple bands at around the Mr of the collagenases that are induced by IL-1α/OSM and repressed by the additional presence of HDAC inhibitors in this system.
Discussion
HDAC inhibitors are currently being developed as cancer therapeutics, largely by virtue of their impact upon the cell cycle and apoptosis [30] in transformed cells. However, it is clear that such compounds have pleiotropic effects on gene expression. Conceptually, the action of HDAC inhibitors leading to an increase in histone acetylation should induce expression of susceptible genes, but in fact, many instances of a repression of gene expression have been reported [25-29]. In yeast, the ability of TSA to down-regulate some genes very rapidly (within 15 min of exposure) suggests that HDACs may function as direct transcriptional activators in some instances [25].
The combination of IL-1 and oncostatin M potently induces both cartilage aggrecan and collagen degradation in vitro and in vivo [41,42] and these factors induce the expression of a number of metalloproteinase genes in chondrocyte cell lines [46]. The addition of either of two chemically distinct HDAC inhibitors to cartilage explant cultures blocks IL-1/OSM-induced cartilage catabolism with a decrease in collagenolytic activity in the conditioned culture medium. TSA and NaBy themselves do not directly inhibit collagenase activity, and it therefore seemed likely that they were altering expression of genes encoding the metalloproteinases or their inhibitors. Using SW1353 chondrosarcoma cells, which are known to respond to IL-1/OSM [40], and primary chondrocytes, real-time RT-PCR gene profiling showed that the expression of a number of MMP and ADAMTS genes was robustly induced by IL-1/OSM and repressed by HDAC inhibitors. In SW1353 cells, MMP2 is not induced by IL-1/OSM nor altered by HDAC inhibitors; MMP9 is weakly induced by IL-1/OSM and this induction is repressed by HDAC inhibitors. In primary chondrocytes, MMP2 expression is induced approximately twofold to fourfold by IL-1/OSM, but is not then repressed by HDAC inhibitors. This is in marked contrast to the zymography data from cartilage explants and suggests a role for cell–matrix interactions in mediating the effects of IL-1/OSM on these gelatinase genes.
Previous studies have shown that TSA represses MMP2 expression in mouse 3T3 fibroblasts but not in human HT1080 fibrosarcoma cells [47,48], showing that the effects of HDAC inhibitors on MMP expression are specific to cell type and potentially to species. In primary chondrocytes, the effects of HDAC inhibitors on the collagenases (MMP1, MMP8, MMP13) mirrored that seen in the SW1353 cell line; however, MMP3, though strongly induced by IL-1/OSM in primary chondrocytes, was not significantly repressed by HDAC inhibitors. The ability of HDAC inhibitors to repress MMP expression at the mRNA level is reiterated at the protein level, as we have shown for MMP-1 and MMP-13. In primary chondrocytes, the aggrecanases ADAMTS4, ADAMTS5, and ADAMTS9 were also strongly induced by IL-1/OSM and were repressed by HDAC inhibitors. In SW1353 cells, both ADAMTS4 and ADAMTS5 genes are expressed at a very low level and are therefore difficult to measure; ADAMTS9, however, is expressed robustly, is induced by IL-1/OSM, and is repressed by HDAC inhibitors, with a pattern similar to that shown for MMP1 and MMP13 in Fig. 3a. This repression of aggrecanase gene expression is consistent with the ability of HDAC inhibitors to inhibit cartilage glycosaminoglycan release as shown in Fig. 1.
HDAC inhibitors appear to affect not only the expression of collagenolytic and gelatinolytic MMPs, but also their activation (Fig. 2). It is known that activation of procollagenases is a key control point in cartilage resorption and this can be mediated by cascades within the MMP family (e.g. MMP-3 can activate procollagenases) or via the action of other enzyme families (e.g. plasmin, a serine proteinase) [49,50]. Therefore, it is likely that HDAC inhibitors repress the expression of one or more key procollagenase activators in cartilage; study of, for example, plasminogen or plasminogen activator expression might be informative.
Since almost all metalloproteinase genes that are robustly induced by IL-1/OSM are then repressed by the further addition of HDAC inhibitors, a likely explanation is the ability of HDAC inhibitors to interfere with IL-1/OSM signalling. Since these cytokines are proinflammatory mediators, action via nuclear factor κB is one possibility; however, the literature shows that TSA actually potentiates signalling through this pathway [51,52]. OSM, an IL-6 family cytokine, signals through the STAT pathway; recent reports show that HDAC activity plays an essential role in at least STAT1 signalling, and that TSA can therefore abrogate STAT1-induced gene expression [53,54]. We (TEC and co-workers) have previously reported that at least STAT3 signalling indirectly mediates the ability of IL-1/OSM to induce MMP1 gene expression [55]. Dissecting the pathways that mediate the impact of HDAC inhibitors on induction of metalloproteinase genes by IL-1/OSM will therefore be one focus of our future work.
Two previous reports using the rodent models of rheumatoid arthritis (rat adjuvant arthritis and murine autoantibody-mediated arthritis) showed that HDAC inhibitors (TSA, phenylbutyrate, or FK228) block proliferation of cultured synovial fibroblasts with accompanying up-regulation of cell cycle inhibitors (p16INK4 and p21Cip1) [31,32]. In vivo, this was mirrored with inhibition of synovial hyperplasia and pannus formation, leading to abrogation of cartilage destruction in the models. Interestingly, the HDAC inhibitors also repressed expression of TNF-α and/or IL-1 in synovial tissue. These reports suggest that HDAC inhibitors may represent a new class of compounds for treatment of rheumatoid arthritis [31,32]. Anti-inflammatory properties of another HDAC inhibitor, suberoylanilide hydroxamic acid (SAHA), have also been demonstrated in vitro and in vivo, via the suppression of proinflammatory cytokines such as TNF-α and IL-1 [56,57].
All of these papers suggest a major effect of HDAC inhibitors in repressing the production of proinflammatory cytokines. Our current data show for the first time that HDAC inhibitors can also function as potent repressors of key metalloproteinase expression in cartilage and chondrocytes and thus block cartilage breakdown. This suggests they may have wider therapeutic use outside of just the inflammatory arthritides, as chondroprotective agents. It should be underlined that where the action of HDAC inhibitors is to repress the induced expression of MMP or ADAMTS genes, this expression is pushed back to control levels but not to zero. This may be important in any therapeutic use of HDAC inhibitors, since normal connective tissue turnover may therefore be unimpaired (though it should also be noted that some MMP and ADAMTS genes are induced by HDAC inhibitors). Our future work will identify the HDAC(s) that have an effect on metalloproteinase expression and identify the mechanism by which this occurs. This has the potential to allow the design and use of compounds specific for one HDAC (or a small number of HDACs), which may be crucial in avoiding toxicity in vivo.
Conclusion
HDAC inhibitors can repress the expression and activity of matrix-degrading proteinases in chondrocytes and cartilage. These compounds, in preclinical development as chemotherapeutic agents, also have strong potential as chondroprotective agents.
Abbreviations
ADAMTS = a disintegrin and metalloproteinase domain (ADAM) with thrombospondin motifs; APMA = 4-aminophenylmercuric acetate; DMEM = Dulbecco's modified Eagle's medium; ECM = extracellular matrix; FCS = fetal calf serum; HDAC = histone deacetylase; IC50 = median inhibitory concentration; IL = interleukin; MMP = matrix metalloproteinase; NaBy = sodium butyrate; OSM = oncostatin M; PCR = polymerase chain reaction; RT = reverse transcriptase; TIMP = tissue inhibitor of metalloproteinases; TNF = tumour necrosis factor; TSA = trichostatin A.
Competing interests
IMC, DAY, and TEC have filed a patent relating to the contents of this manuscript.
Authors' contributions
DAY helped conceive the study, designed and carried out cell experiments, carried out some of the real-time RT-PCR experiments, and helped to draft the manuscript. RLL carried out cartilage resorption assays and assessed toxicity of the HDAC inhibitors in this system. CJP designed real-time PCR primer probe sets and carried out some of the real-time RT-PCR. DJ carried out proteinase activity assays associated with the study. LK carried out real-time RT-PCR associated with the cell experiments. DRE designed and validated the real-time PCR methodology and helped to draft the manuscript. TEC designed the cartilage resorption assays and helped to draft the manuscript. IMC helped conceive, design, and coordinate the study, carried out some cell experiments, and helped to draft the manuscript. All authors read and approved the final manuscript.
Acknowledgements
DAY was funded by the Dunhill Medical Trust. LK is supported by an Industrial CASE studentship from BBSRC (Biotechnology and Biological Sciences Research Council) and AstraZeneca.
Figures and Tables
Figure 1 Histone deacetylase inhibitors block cartilage glycosaminoglycan and collagen loss induced by IL-1α/OSM. Bovine nasal cartilage discs were cultured in the presence or absence of I/O (a combination of IL-1α and oncostatin M (OSM)) and a histone deacetylase inhibitor, either (a) I/O (1 ng/ml IL-1α, 10 ng/ml OSM) with trichostatin A (TSA) or (b) I/O (0.2 ng/ml IL-1α, 2 ng/ml OSM) with sodium butyrate (NaBy). Cartilage was incubated until day 7 and supernates were harvested and replaced with fresh reagents until day 14. Glycosaminoglycan (GAG) release is shown as at day 7 and was assayed using the dimethylmethylene blue method. Collagen release is shown as at day 14 and was measured using an assay for hydroxyproline. Viability was assessed by measuring lactate dehydrogenase in the conditioned medium. Assays were performed at least twice using quadruplicate samples; means ± standard deviations are represented. *P < 0.05, **P < 0.01, ***P < 0.001.
Figure 2 HDAC inhibitors decrease collagenolytic and gelatinolytic activity from bovine nasal explants and block collagenase activation. Conditioned media from cartilage assays (day 14) as in Fig. 1a were assayed (a) for collagenase activity in the presence or absence of 0.67 mM APMA, an activator of procollagenases (means ± standard errors of the mean), and (b) for gelatinase activity, using gelatin zymography. APMA, aminophenylmercuric acetate; HDAC, histone deacetylase; I/O, a combination of IL-1α and OSM; MMP, matrix metalloproteinase; OSM, oncostatin M; TSA, trichostatin A.
Figure 3 Histone deacetylase inhibitors abrogate IL-1α/OSM-induced expression of key metalloproteinase genes. Cells were starved of serum for 24 hours before stimulation with I/O, a combination of IL-1α (5 ng/ml) and oncostatin M (OSM) (10 ng/ml) for 6 hours in the absence or presence of either trichostatin A (TSA) ((a) as shown or (b, c) 500 ng/ml) or NaBy ((a) as shown; (b, c) 10 mM). Total RNA was isolated and subjected to quantitative RT-PCR for expression of the genes (a, b) MMP1 and MMP13 and (c) ADAMTS4, ADAMTS5, and ADAMTS9. Data were normalised to the 18 S rRNA housekeeping gene; means and ranges are plotted. Absolute numbers are primer/probe-set-dependent and so cannot be compared between genes. (a) SW1353 chondrosarcoma cells; (b, c) primary human chondrocytes; (b) (inset) expression of COL2A1 and aggrecan by conventional RT-PCR. I/O, O/I, a combination of IL-1α and OSM; NaBy, sodium butyrate.
Figure 4 Histone deacetylase inhibitors repress matrix metalloproteinase (MMP) protein expression and activity. Cells were starved of serum for 24 hours before stimulation with I/O, a combination of IL-1α (5 ng/ml) and oncostatin M (10 ng/ml), for 24 hours in the absence or presence of trichostatin A (TSA) (500 ng/ml) or sodium butyrate (NaBy) (10 mM). Conditioned media were subjected to western blot analysis using a rabbit anti-(human MMP-1) antibody or a sheep anti-(human MMP-13) antibody or gelatin zymography, as described in Materials and methods.
==== Refs
Eyre D Collagen of articular cartilage Arthritis Res 2002 4 30 35 11879535 10.1186/ar380
Roughley PJ Articular cartilage and changes in arthritis: noncollagenous proteins and proteoglycans in the extracellular matrix of cartilage Arthritis Res 2001 3 342 347 11714388 10.1186/ar326
Clark IM Parker AE Metalloproteinases: their role in arthritis and potential as therapeutic targets Expert Opin Ther Targets 2003 7 19 34 12556200 10.1517/eott.7.1.19.21171
Vankemmelbeke MN Jones GC Fowles C Ilic MZ Handley CJ Day AJ Knight CG Mort JS Buttle DJ Selective inhibition of ADAMTS-1, -4 and -5 by catechin gallate esters Eur J Biochem 2003 270 2394 2403 12755694 10.1046/j.1432-1033.2003.03607.x
Somerville RP Longpre JM Jungers KA Engle JM Ross M Evanko S Wight TN Leduc R Apte SS Characterization of ADAMTS-9 and ADAMTS-20 as a distinct ADAMTS subfamily related to Caenorhabditis elegans GON-1 J Biol Chem 2003 278 9503 9513 12514189 10.1074/jbc.M211009200
Yamanouchi PC Patent WO0134785 2001
Collins-Racie LA Flannery CR Zeng W Corcoran C Annis-Freeman B Agostino MJ Arai M DiBlasio-Smith E Dorner AJ Georgiadis KE ADAMTS-8 exhibits aggrecanase activity and is expressed in human articular cartilage Matrix Biol 2004 23 219 230 15296936 10.1016/j.matbio.2004.05.004
Lark MW Bayne EK Flanagan J Harper CF Hoerrner LA Hutchinson NI Singer IIDonatelli SA Weidner JR Aggrecan degradation in human cartilage. Evidence for both matrix metalloproteinase and aggrecanase activity in normal, osteoarthritic, and rheumatoid joints J Clin Invest 1997 100 93 106 9202061
Porter S Clark IM Kevorkian L Edwards DR The ADAMTS metalloproteinases Biochem J 2005 386 15 27 15554875 10.1042/BJ20040424
Baker AH Edwards DR Murphy G Metalloproteinase inhibitors: biological actions and therapeutic opportunities J Cell Sci 2002 115 3719 3727 12235282 10.1242/jcs.00063
Wolffe AP Guschin D Review: chromatin structural features and targets that regulate transcription J Struct Biol 2000 129 102 122 10806063 10.1006/jsbi.2000.4217
Narlikar GJ Fan HY Kingston RE Cooperation between complexes that regulate chromatin structure and transcription Cell 2002 108 475 487 11909519 10.1016/S0092-8674(02)00654-2
Sudarsanam P Winston F The Swi/Snf family nucleosome-remodeling complexes and transcriptional control Trends Genet 2000 16 345 351 10904263 10.1016/S0168-9525(00)02060-6
Berger SL Histone modifications in transcriptional regulation Curr Opin Genet Dev 2002 12 142 148 11893486 10.1016/S0959-437X(02)00279-4
Clayton AL Rose S Barratt MJ Mahadevan LC Phosphoacetylation of histone H3 on c-fos- and c-jun-associated nucleosomes upon gene activation Embo J 2000 19 3714 3726 10899125 10.1093/emboj/19.14.3714
Kouzarides T Acetylation: a regulatory modification to rival phosphorylation? Embo J 2000 19 1176 1179 10716917 10.1093/emboj/19.6.1176
de Ruijter AJ van Gennip AH Caron HN Kemp S van Kuilenburg AB Histone deacetylases (HDACs): characterization of the classical HDAC family Biochem J 2003 370 737 749 12429021 10.1042/BJ20021321
Gao L Cueto MA Asselbergs F Atadja P Cloning and functional characterization of HDAC11, a novel member of the human histone deacetylase family J Biol Chem 2002 277 25748 25755 11948178 10.1074/jbc.M111871200
Ng HH Bird A Histone deacetylases: silencers for hire Trends Biochem Sci 2000 25 121 126 10694882 10.1016/S0968-0004(00)01551-6
Braun H Koop R Ertmer A Nacht S Suske G Transcription factor Sp3 is regulated by acetylation Nucleic Acids Res 2001 29 4994 5000 11812829 10.1093/nar/29.24.4994
Vries RG Prudenziati M Zwartjes C Verlaan M Kalkhoven E Zantema A A specific lysine in c-Jun is required for transcriptional repression by E1A and is acetylated by p300 Embo J 2001 20 6095 6103 11689449 10.1093/emboj/20.21.6095
Gregoretti IV Lee YM Goodson HV Molecular evolution of the histone deacetylase family: functional implications of phylogenetic analysis J Mol Biol 2004 338 17 31 15050820 10.1016/j.jmb.2004.02.006
Yoshida M Kijima M Akita M Beppu T Potent and specific inhibition of mammalian histone deacetylase both in vivo and in vitro by trichostatin A J Biol Chem 1990 265 17174 17179 2211619
Kruh J Effects of sodium butyrate, a new pharmacological agent, on cells in culture Mol Cell Biochem 1982 42 65 82 6174854
Bernstein BE Tong JK Schreiber SL Genomewide studies of histone deacetylase function in yeast Proc Natl Acad Sci USA 2000 97 13708 13713 11095743 10.1073/pnas.250477697
Mulholland NM Soeth E Smith CL Inhibition of MMTV transcription by HDAC inhibitors occurs independent of changes in chromatin remodeling and increased histone acetylation Oncogene 2003 22 4807 4818 12894222 10.1038/sj.onc.1206722
Nair AR Boersma LJ Schiltz L Chaudhry MA Muschel RJ Chaudry A Paradoxical effects of trichostatin A: inhibition of NF-Y-associated histone acetyltransferase activity, phosphorylation of hGCN5 and downregulation of cyclin A and B1 mRNA Cancer Lett 2001 166 55 64 11295287 10.1016/S0304-3835(01)00418-9
Pujuguet P Radisky D Levy D Lacza C Bissell MJ Trichostatin A inhibits beta-casein expression in mammary epithelial cells J Cell Biochem 2001 83 660 670 11746508 10.1002/jcb.1260
Saunders N Dicker A Popa C Jones S Dahler A Histone deacetylase inhibitors as potential anti-skin cancer agents Cancer Res 1999 59 399 404 9927053
Johnstone RW Histone-deacetylase inhibitors: novel drugs for the treatment of cancer Nat Rev Drug Discov 2002 1 287 299 12120280 10.1038/nrd772
Chung YL Lee MY Wang AJ Yao LF A therapeutic strategy uses histone deacetylase inhibitors to modulate the expression of genes involved in the pathogenesis of rheumatoid arthritis Mol Ther 2003 8 707 717 14599803 10.1016/S1525-0016(03)00235-1
Nishida K Komiyama T Miyazawa S Shen ZN Furumatsu T Doi H Yoshida A Yamana J Yamamura M Ninomiya Y Histone deacetylase inhibitor suppression of autoantibody-mediated arthritis in mice via regulation of p16INK4a and p21(WAF1/Cip1) expression Arthritis Rheum 2004 50 3365 3376 15476220 10.1002/art.20709
Nuttall RK Pennington CJ Taplin J Wheal A Yong VW Forsyth PA Edwards DR Elevated membrane-type matrix metalloproteinases in gliomas revealed by profiling proteases and inhibitors in human cancer cells Mol Cancer Res 2003 1 333 345 12651907
Porter S Scott SD Sassoon EM Williams MR Jones JL Girling AC Ball RY Edwards DR Dysregulated expression of adamalysin-thrombospondin genes in human breast carcinoma Clin Cancer Res 2004 10 2429 2440 15073121
Robbins JR Thomas B Tan L Choy B Arbiser JL Berenbaum F Goldring MB Immortalized human adult articular chondrocytes maintain cartilage-specific phenotype and responses to interleukin-1beta Arthritis Rheum 2000 43 2189 2201 11037878 10.1002/1529-0131(200010)43:10<2189::AID-ANR6>3.0.CO;2-S
Billington CJ Cartilage proteoglycan release assay Methods Mol Biol 2001 151 451 456 11217320
Shingleton WD Ellis AJ Rowan AD Cawston TE Retinoic acid combines with interleukin-1 to promote the degradation of collagen from bovine nasal cartilage: matrix metalloproteinases-1 and -13 are involved in cartilage collagen breakdown J Cell Biochem 2000 79 519 531 10996843 10.1002/1097-4644(20001215)79:4<519::AID-JCB10>3.0.CO;2-U
Koshy PJ Rowan AD Life PF Cawston TE 96-well plate assays for measuring collagenase activity using (3)H-acetylated collagen Anal Biochem 1999 275 202 207 10552905 10.1006/abio.1999.4310
Clark IM Wright JK Cawston TE Hazleman BL Polyclonal antibodies against human fibroblast collagenase and the design of an enzyme-linked immunosorbent assay to measure TIMP-collagenase complex Matrix 1992 12 108 115 1318493
Cowell S Knauper V Stewart ML D'Ortho MP Stanton H Hembry RM Lopez-Otin C Reynolds JJ Murphy G Induction of matrix metalloproteinase activation cascades based on membrane-type 1 matrix metalloproteinase: associated activation of gelatinase A, gelatinase B and collagenase 3 Biochem J 1998 331 Pt 2 453 458 9531484
Cawston TE Curry VA Summers CA Clark IM Riley GP Life PF Spaull JR Goldring MB Koshy PJ Rowan AD The role of oncostatin M in animal and human connective tissue collagen turnover and its localization within the rheumatoid joint Arthritis Rheum 1998 41 1760 1771 9778217 10.1002/1529-0131(199810)41:10<1760::AID-ART8>3.0.CO;2-M
Rowan AD Hui W Cawston TE Richards CD Adenoviral gene transfer of interleukin-1 in combination with oncostatin M induces significant joint damage in a murine model Am J Pathol 2003 162 1975 1984 12759253
Furumai R Komatsu Y Nishino N Khochbin S Yoshida M Horinouchi S Potent histone deacetylase inhibitors built from trichostatin A and cyclic tetrapeptide antibiotics including trapoxin Proc Natl Acad Sci USA 2001 98 87 92 11134513 10.1073/pnas.011405598
Knauper V Lopez-Otin C Smith B Knight G Murphy G Biochemical characterization of human collagenase-3 J Biol Chem 1996 271 1544 1550 8576151 10.1074/jbc.271.3.1544
Clark IM Cawston TE Fragments of human fibroblast collagenase. Purification and characterization Biochem J 1989 263 201 206 2557822
Koshy PJ Lundy CJ Rowan AD Porter S Edwards DR Hogan A Clark IM Cawston TE The modulation of matrix metalloproteinase and ADAM gene expression in human chondrocytes by interleukin-1 and oncostatin M: a time-course study using real-time quantitative reverse transcription-polymerase chain reaction Arthritis Rheum 2002 46 961 967 11953973 10.1002/art.10212
Ailenberg M Silverman M Differential effects of trichostatin A on gelatinase A expression in 3T3 fibroblasts and HT-1080 fibrosarcoma cells: implications for use of TSA in cancer therapy Biochem Biophys Res Commun 2003 302 181 185 12604329 10.1016/S0006-291X(03)00150-5
Ailenberg M Silverman M Trichostatin A-histone deacetylase inhibitor with clinical therapeutic potential-is also a selective and potent inhibitor of gelatinase A expression Biochem Biophys Res Commun 2002 298 110 115 12379227 10.1016/S0006-291X(02)02420-8
Milner JM Elliott SF Cawston TE Activation of procollagenases is a key control point in cartilage collagen degradation: interaction of serine and metalloproteinase pathways Arthritis Rheum 2001 44 2084 2096 11592371 10.1002/1529-0131(200109)44:9<2084::AID-ART359>3.0.CO;2-R
Milner JM Rowan AD Elliott SF Cawston TE Inhibition of furin-like enzymes blocks interleukin-1alpha/oncostatin M-stimulated cartilage degradation Arthritis Rheum 2003 48 1057 1066 12687549 10.1002/art.10873
Ashburner BP Westerheide SD Baldwin AS Jr The p65 (RelA) subunit of NF-kappaB interacts with the histone deacetylase (HDAC) corepressors HDAC1 and HDAC2 to negatively regulate gene expression Mol Cell Biol 2001 21 7065 7077 11564889 10.1128/MCB.21.20.7065-7077.2001
Chen L Fischle W Verdin E Greene WC Duration of nuclear NF-kappaB action regulated by reversible acetylation Science 2001 293 1653 1657 11533489 10.1126/science.1062374
Nusinzon I Horvath CM Interferon-stimulated transcription and innate antiviral immunity require deacetylase activity and histone deacetylase 1 Proc Natl Acad Sci USA 2003 100 14742 14747 14645718 10.1073/pnas.2433987100
Klampfer L Huang J Swaby LA Augenlicht L Requirement of histone deactylase activity for signaling by STAT1 J Biol Chem 2004 279 30358 30368 15123634 10.1074/jbc.M401359200
Catterall JB Carrere S Koshy PJ Degnan BA Shingleton WD Brinckerhoff CE Rutter J Cawston TE Rowan AD Synergistic induction of matrix metalloproteinase 1 by interleukin-1alpha and oncostatin M in human chondrocytes involves signal transducer and activator of transcription and activator protein 1 transcription factors via a novel mechanism Arthritis Rheum 2001 44 2296 2310 11665970 10.1002/1529-0131(200110)44:10<2296::AID-ART392>3.0.CO;2-9
Leoni F Zaliani A Bertolini G Porro G Pagani P Pozzi P Dona G Fossati G Sozzani S Azam T The antitumor histone deacetylase inhibitor suberoylanilide hydroxamic acid exhibits antiinflammatory properties via suppression of cytokines Proc Natl Acad Sci USA 2002 99 2995 3000 11867742 10.1073/pnas.052702999
Reddy P Maeda Y Hotary K Liu C Reznikov LL Dinarello CA Ferrara JL Histone deacetylase inhibitor suberoylanilide hydroxamic acid reduces acute graft-versus-host disease and preserves graft-versus-leukemia effect Proc Natl Acad Sci USA 2004 101 3921 3926 15001702 10.1073/pnas.0400380101
| 15899037 | PMC1174946 | CC BY | 2021-01-04 16:02:37 | no | Arthritis Res Ther. 2005 Feb 22; 7(3):R503-R512 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1702 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17041589905110.1186/ar1704Research ArticleTolerability and adverse events in clinical trials of celecoxib in osteoarthritis and rheumatoid arthritis: systematic review and meta-analysis of information from company clinical trial reports Moore R Andrew [email protected] Sheena [email protected] Geoffrey T [email protected] Henry J [email protected] Pain Research and Nuffield Department of Anaesthetics, University of Oxford, Oxford Radcliffe NHS Trust, Oxford, UK2 Department of Outcomes Research and Evidence-based Medicine, Pfizer Ltd, Walton Oaks, Surrey, UK2005 24 3 2005 7 3 R644 R665 24 11 2004 4 1 2005 21 1 2005 28 1 2005 Copyright © 2005 Moore et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The objective was to improve understanding of adverse events occurring with celecoxib in the treatment of osteoarthritis and rheumatoid arthritis. Data were extracted from company clinical trial reports of randomised trials of celecoxib in osteoarthritis or rheumatoid arthritis lasting 2 weeks or more. Outcomes were discontinuations (all cause, lack of efficacy, adverse event, gastrointestinal adverse event), endoscopically detected ulcers, gastrointestinal or cardio-renal events, and major changes in haematological parameters. The main comparisons were celecoxib (all doses) versus placebo, paracetamol (acetaminophen) 4,000 mg daily, rofecoxib 25 mg daily, or nonsteroidal anti-inflammatory drugs (NSAIDs) (naproxen, diclofenac, ibuprofen, and loxoprofen). For NSAIDs, celecoxib was compared both at all doses and at licensed doses (200 to 400 mg daily). Thirty-one trials included 39,605 randomised patients. Most patients had osteoarthritis and were women of average age 60 years or above. Most trials lasted 12 weeks or more. Doses of celecoxib were 50 to 800 mg/day. Compared with placebo, celecoxib had fewer discontinuations for any cause or for lack of efficacy, fewer serious adverse events, and less nausea. It had more patients with dyspepsia, diarrhoea, oedema, more adverse events that were gastrointestinal or treatment related, and more patients experiencing an adverse event. There were no differences for hypertension, gastrointestinal tolerability, or discontinuations for adverse events. Compared with paracetamol, celecoxib had fewer discontinuations for any cause, for lack of efficacy, or diarrhoea, but no other differences. Compared with rofecoxib, celecoxib had fewer patients with abdominal pain and oedema, but no other differences. Compared with NSAIDs, celecoxib had fewer symptomatic ulcers and bleeds, endoscopically detected ulcers, and discontinuations for adverse events or gastrointestinal adverse events. Fewer patients had any, or a gastrointestinal, or a treatment-related adverse event, or vomiting, abdominal pain, dyspepsia, or reduced haemoglobin or haematocrit. Discontinuations for lack of efficacy were higher. No differences were found for all-cause discontinuations, serious adverse events, hypertension, diarrhoea, nausea, oedema, myocardial infarction, cardiac failure, or raised creatinine. Company clinical trial reports present much more information than published papers. Adverse event information is clearly presented in company clinical trial reports, which are an ideal source of information for systematic review and meta-analysis.
==== Body
Introduction
Arthritis is a common, progressive condition, which is associated with considerable pain and inflammation, and has a strong impact on quality of life. It is the major reason for hip or knee replacements [1].
It is more prevalent in women than men, and in older people. One community-based study [2] conducted in Scotland showed that 25% of patients had arthritis by age 65. Of these, a quarter had pain that was highly disabling and at least moderately limiting. A further quarter had pain that was more severe. In a UK general practice survey of patients' perspectives in osteoarthritis [3], a quarter of responders reported some dissatisfaction with their treatment and another quarter stated that their pain control was poor. High levels of negative impact were associated with inability to walk, bathe, dress, or sleep, with 40% of patients saying that these activities were often or always affected. A quarter of patients used over-the-counter medicines, mainly paracetamol or ibuprofen, in addition to those prescribed by their doctor. Half of responders were over age 65, and two-thirds were women.
Drug treatment is ideally effective, safe, and well tolerated. NSAIDs have provided the mainstay of pain therapy, particularly in the early stages of disease, but are often associated with clinically relevant adverse events.
Common events such as nausea or dizziness, often considered minor, can have an impact on people's lives and reduce compliance with prescribed dose. Patients with arthritis avoid adverse events, choosing less effective medicine with less likelihood of adverse events over more effective medicine with more adverse events [4]. Only 20% of patients with arthritis prescribed NSAIDs will be taking the same drug after one year [5], adverse events being a major reason for discontinuation.
Serious adverse events occur infrequently, but the consequence to the individual may be considerable. With conventional NSAIDs, there is the risk of major harm through gastrointestinal ulceration, perforation, and bleeding. These events consume considerable resources through cost of hospitalisation and treatment, or through coprescription of gastroprotective agents to minimise the risk of major harm [6].
Cox-2-selective inhibitors (coxibs) are an alternative to NSAIDs, developed to give better gastrointestinal safety and tolerability. For evaluation of the adverse-event profiles of coxibs, outcomes of interest include endoscopically detected ulcers and erosions, and symptomatic ulcers, which may progress to bleeding ulcers, and can even cause death [7]. Renal failure [8,9] and heart failure [10,11] also occur with NSAIDs or coxibs. Other adverse event outcomes that are useful to know include those describing discontinuation (early withdrawal from the trial), particularly discontinuation because of adverse events or lack of efficacy.
This systematic review and meta-analysis of celecoxib in osteoarthritis and rheumatoid arthritis was conducted using information from company clinical trial reports, supplied by Pfizer Ltd, of completed randomised, double-blind trials from the celecoxib clinical trials programme. The objectives were to examine tolerability, minor and major adverse events, and endoscopically detected ulceration associated with celecoxib in arthritis.
Materials and methods
Randomised, double-blind, controlled trials, of 2 weeks' duration or longer with any dose of celecoxib and any comparator, in osteoarthritis or rheumatoid arthritis, were supplied as company clinical trial reports by Pfizer Ltd. Open-label extension studies were not included. A declaration was signed by Pfizer that all completed (by December 2003) trials of relevance from the celecoxib clinical trial programme had been made available. A protocol for the review and analysis, including definitions of outcomes, was agreed beforehand.
Financial support was provided by Pfizer Ltd, with the provision that all relevant trial reports completed by December 2003 were made available, and that the authors were free to publish their findings whatever the outcome of the review. Other funding was from Pain Research funds of the Oxford Pain Relief Trust. No funding source had any role in deciding what to publish, when to publish, or where to publish it.
Trials
Thirty-one Phase II, III, and IV clinical trial reports of celecoxib in osteoarthritis or rheumatoid arthritis were provided for evaluation. All compared celecoxib in various dosing regimens with placebo, paracetamol (acetaminophen) 4,000 mg/day, rofecoxib 25 mg/day, or an NSAID commonly used in the treatment of arthritis. Comparator NSAIDs were given at the maximum licensed dose; these were naproxen 1,000 mg, ibuprofen 2,400 mg, diclofenac 100 to 150 mg, and loxoprofen 180 mg daily. Details of the included trials are in Table 1.
Trial inclusion and exclusion criteria
Patients were adults who had a clinical diagnosis of osteoarthritis or rheumatoid arthritis that was symptomatic, usually of 3 months' duration or longer, and required long-term treatment with anti-inflammatory drugs or other analgesics for the control of pain. Further details of inclusion and exclusion criteria for both osteoarthritis and rheumatoid arthritis can be found in Additional file 1.
Trial methods
Eligible patients typically entered a pretreatment period of up to 14 days, during which baseline observations were conducted. Nonstudy NSAIDs and other analgesics were discontinued, with the exception of aspirin (up to 325 mg daily) and paracetamol (up to 2 g per day for a maximum of 3 days but not within 48 hours of arthritis assessments), which were permitted for reasons other than control of arthritis pain. Other drugs specifically excluded were antibiotics for Helicobacter pylori eradication, metronidazole, anticoagulants, lithium, and anti-ulcer drugs including proton pump inhibitors, H2 antagonists, antacids, sucralfate, and misoprostol.
Patients were randomised under double-blind conditions to receive oral celecoxib, paracetamol, rofecoxib, an NSAID, or placebo. Several studies had both an active and a placebo comparator, and several compared different fixed dose regimens of celecoxib. Table 1 shows the study treatments, dosing, and number and baseline characteristics of patients for the individual trials. All trials conformed to good clinical practice guidelines.
Information collected on adverse events
In all studies, information was collected on patients who experienced any adverse event, serious adverse events, adverse events relating to body systems, and discontinuations. Information was collected on the occurrence of endoscopically detected ulcers and erosions from those trials in which all patients were scheduled to have endoscopy before and at various times during treatment. Definitions used in the trials were those of the World Health Organization (Adverse Reaction Terminology). The definitions used in this review are in Additional file 2.
Meta-analysis
Outcomes chosen for the meta-analysis
Outcomes chosen related to adverse events and tolerability. These included discontinuation (all-cause, lack of efficacy, adverse event, and gastrointestinal adverse event), patients with any adverse event, patients with any treatment-related adverse event, and patients with any serious adverse event.
For gastrointestinal adverse events, we included an overall measure of gastrointestinal tolerability as well as individual gastrointestinal adverse events of nausea, vomiting, abdominal pain, dyspepsia, diarrhoea, and ulcers or bleeds. Treatment-emergent ulcers and bleeds were analysed together because of their important sequelae. Endoscopically detected ulcers were taken from reports in which all patients in the trial had endoscopy with the specific intent of measuring endoscopic lesions, and where this was a prime outcome in the trial. They were additionally analysed according to the concomitant use of low-dose aspirin.
Specific cardio-renal adverse events included cardiac failure, hypertension, raised creatinine, and oedema at any body site. Analysis of oedema by body site, or hypertension by subcategory, was not carried out, as event numbers were too low for practicable analysis.
Trial quality and validity
Three authors independently read each clinical trial report and scored the reports for reporting quality and validity. Disagreements were discussed and consensus achieved. Trials were scored for quality using a three-item, 1- to 5-point scale [12], and at least two points, one each for randomisation and double blinding, were required for inclusion. Trials were scored for validity using an eight-item, 16-point scale [13]; there was no minimum requirement for inclusion in the systematic review.
Analysis
Guidelines for quality of reporting of meta-analyses were followed where appropriate [14].
The prior intention was to pool data where there was clinical homogeneity, with similarity in terms of patients, dose, duration, outcomes, and comparators. It was recognised, however, that this could lead to a large number of comparisons, with small numbers of events, where random chance could dominate effects of treatment on adverse events [15].
The main issues were the comparator treatments in trials and the dose of celecoxib. Pooling of data was therefore restricted to comparison between celecoxib and placebo, paracetamol, rofecoxib, and NSAIDs, because each comparator had a different mechanism of action from any other. In addition, analysis of celecoxib against all active comparators combined was carried out. For active comparisons, most of the information was likely to reside in those between celecoxib and NSAIDs, and we chose to perform two analyses: comparisons of all doses of celecoxib with all doses of NSAIDs, and between licensed daily doses of celecoxib and licensed doses of NSAIDs. NSAIDs were used at licensed doses, usually at maximum daily dose, and rofecoxib was used at 25 mg daily.
Information for osteoarthritis and rheumatoid arthritis was combined because the number of patients in trials with rheumatoid arthritis was small. Though there are differences between the conditions, notably age of onset, there are no clear reasons why treatment-emergent adverse events should differ between conditions. Analysis of celecoxib dose, and of duration of studies, was restricted to discontinuations due to lack of efficacy or to adverse events, where there were more than 20 events, and where the outcome had direct clinical relevance.
Analysis of data could potentially be performed in two ways. The simplest method would be to combine the absolute proportions of patients experiencing an adverse event, using the intention-to-treat population (randomised, at least one dose of drug) as the denominator. This method has a potential disadvantage of not taking into account different durations of studies, and possible different exposures between treatments because of different withdrawal rates. An alternative method would be to calculate adverse events as the rate of events occurring per year of exposure, theoretically taking both different durations and differential exposure into account.
This second method was impractical for several reasons. Trial reports generally did not have information to allow calculation of median duration of use. For instance, they reported neither average days of use nor individual days of use, so that an average could not be calculated. The reports generally had information on compliance, and generally there was no significant difference between celecoxib and its comparators. The two largest trials, with over half the patients, gave patient years of exposure in the trial reports, and these were identical for celecoxib and NSAID. In a separate analysis of cardiovascular events in celecoxib trials, which included 30,000 of the 40,000 patients in this review, there were negligible differences between treatment durations [16].
Outcomes were pooled in an intention-to-treat (number of patients randomised and receiving at least one dose of trial drug) analysis. Homogeneity tests and funnel plots, though commonly used in meta-analysis, were not used here because they have been found to be unreliable [17-19]. Instead clinical homogeneity was examined graphically [20]. Relative benefit (or risk) and number-needed-to-treat (or harm) were calculated with 95% confidence intervals. Relative risk was calculated using a fixed effects model [21], with no statistically significant difference between treatments assumed when the 95% confidence intervals included unity. We added 0.5 to celecoxib and comparator arms of trials in which at least one arm had no events. Number-needed-to-treat (or harm) was calculated by the method of Cook and Sackett [22], using the pooled number of observations.
Adverse outcomes were described in terms of harm or prevention of harm, as follows. When significantly fewer adverse events occurred with celecoxib than with a control substance (placebo or active), we used the term 'the number-needed-to-treat to prevent one event' (NNTp). When significantly more adverse events occurred with celecoxib than with an active comparator (paracetamol, rofecoxib, NSAID) we used the term 'number-needed-to-treat to harm one patient' (NNH).
Results
Trials
Clinical reports of 31 randomised trials – 21 in osteoarthritis, 4 in rheumatoid arthritis, and 6 in mixed osteoarthritis or rheumatoid arthritis – were provided for the analysis. Full company study reports for 23 trials contained 180,000 pages. These were comprehensive documents including detailed methods and results sections, tables, and figures. Appendices provided descriptions of the outcome measurement tools used, individual patient outcomes, compliance, case report forms, detailed statistical analyses, and protocol amendments. Full clinical trial reports were not available for eight trials, but extensive clinical trial summaries were provided. Information was extracted directly from the clinical trial reports or summaries.
All trials scored the maximum of five points for quality (Table 1), since they clearly described withdrawals in addition to the methods of randomisation and double blinding. All studies also scored the maximum of 16 points on the validity scale.
The 31 trials had 39,605 patients who were randomised and received at least one dose of study medication (intention-to-treat population). Of these, 25,903 had osteoarthritis, 3,232 had rheumatoid arthritis, and 10,470 were in trials including patients with both conditions. Sixteen of 21 trials in osteoarthritis (8,947 patients) lasted 2 to 6 weeks (13 lasted six weeks), and five (16,956 patients) lasted 12 weeks. One of the four trials (327 patients) in rheumatoid arthritis lasted 6 weeks, the other three (2,905 patients) lasted 12 or 24 weeks. Five trials in both osteoarthritis and rheumatoid arthritis (2,502 patients) lasted 12 weeks, and the other (7,968 patients) lasted 52 weeks (though the mean duration of exposure in all three treatment groups was about 7 months; 0.54 to 0.58 years). Most of the observations (77%) were therefore in trials of 12 weeks or longer.
Doses of celecoxib were 50 to 800 mg daily, mostly as twice-daily dosing. In trials of 2 to 6 weeks, 88% of the doses were 200 mg daily. In trials of 12 weeks' duration, 46% of doses were 200 mg and 46% were of 400 mg daily. In trials of 24 weeks or longer, 92% of doses were of 800 mg daily. Longer-lasting trials used higher doses of celecoxib. In comparisons with placebo, 88% of 6,857 patients taking celecoxib had doses in the licensed range of 200 to 400 mg daily. In comparisons with paracetamol and rofecoxib, the celecoxib dose was 200 mg daily. Analysis of licensed doses of celecoxib (200 to 400 mg daily) and NSAIDs not only avoided higher (800 mg) doses, but also the 52-week study that used 800 mg of celecoxib.
Patients and adverse events
Details of the patients included in the trials are in Table 1. In most trials, the majority of patients were women whose average age was 60 years or above (range 17 to 96 years). The relevant medical history, notably about NSAID intolerance or gastrointestinal symptoms after use of NSAIDs and about use of prophylactic low-dose aspirin, was usually reported. Three trials (002, 149, 181) specifically recruited patients with stable, treated hypertension in addition to arthritis. Patients were predominantly Caucasian, but several studies specifically recruited only Asian participants, or those of mixed Asian, Afro-Caribbean, or Hispanic descent.
The adverse event outcomes measured in each trial are detailed in Additional file 3. All of the adverse events were those reported by trial investigators, and none was reported after independent, blinded adjudication.
Adverse events were measured by recording treatment-emergent events, clinical laboratory test results, or changes from baseline in vital signs found by physical examination. At each follow-up visit, patients were asked if they had experienced any symptoms not associated with their arthritis. Patients and study personnel were blinded to the identification of medication throughout the study, and if randomisation blind was broken, the patient was removed from the study.
Discontinuation
Details of discontinuations are shown in Table 2. All-cause and lack-of-efficacy discontinuations were less frequent with celecoxib than with placebo or paracetamol. Adverse-event and gastrointestinal-adverse-event discontinuation (Fig. 1) was less frequent with celecoxib than with NSAIDs (licensed dose or any dose) or any active comparator. All-cause discontinuations were also less frequent with any dose of celebcoxib compared with NSAID or any active comparator. Licensed doses of celebcoxib were not significantly different. Celecoxib did not differ from rofecoxib. The NNTp to prevent discontinuation due to lack of efficacy was 9 (8 to 11) compared with placebo, and 27 (14 to 390) compared with paracetamol. Licensed doses of celecoxib had an NNTp of 74 (47 to 180) for discontinuations due to an adverse event, and an NNTp of 58 (42 to 98) for discontinuations due to a gastrointestinal adverse event, compared with NSAIDs.
Proportions discontinuing because of lack of efficacy or adverse events varied according to drug, dose, and duration. Regarding duration, for instance, discontinuation because of gastrointestinal adverse events was higher for NSAIDs than celecoxib in the one 52-week trial and in trials of shorter duration (Fig. 1).
The details for all 39,605 patients in all trials are shown in Table 3. Discontinuation because of lack of efficacy was high with placebo, 18% over 2 to 6 weeks and 46% by 12 weeks. Effective treatment with licensed doses of celecoxib or NSAIDs reduced discontinuations due to lack of efficacy, with evidence of a dose-response for celecoxib over the range of 100 to 400 mg daily.
There was considerable variation between individual trials regarding discontinuations due to lack of efficacy at 12 weeks, for celecoxib and naproxen. The variability seemed unrelated to condition, and no sensible reason presented itself.
Discontinuations due to adverse events were low with placebo (6% at 12 weeks), little different with celecoxib, and somewhat higher with NSAIDs (Tables 2 and 3). In trials of 24 weeks or longer, discontinuations due to adverse events with 800 mg celecoxib, 100/150 mg diclofenac, and 2,400 mg ibuprofen were between 22% and 26%.
Any adverse event
The proportion of patients reporting any adverse event was of the order of 50% (Table 4). Patients taking celecoxib reported adverse events more frequently than those taking placebo (NNH 15; 11 to 21), and less frequently than with NSAIDs (NNTp 18; 14 to 23 for licensed doses) or any active comparator. There was no difference between celecoxib and either paracetamol or rofecoxib.
Treatment-related adverse events
About one-third of all reported adverse events were considered to be treatment related (Table 4). There was no difference between celecoxib and paracetamol or rofecoxib. More patients taking celecoxib than placebo had a treatment-related adverse event (NNH 71; 39 to 450). Fewer patients experienced a treatment-related adverse event with celecoxib than with NSAID (NNTp 24; 19 to 31 for licensed doses) or any active comparator.
Serious adverse events
The proportion of patients with a serious adverse event was low, averaging 1 to 3% (Table 4). Fewer patients taking celecoxib than placebo had serious adverse events (NNTp 280; 120 to 790). There was no difference in serious adverse event rates for celecoxib compared with paracetamol, rofecoxib, NSAID (Fig. 2), or any active comparator (Table 4). Serious adverse events occurred more often, at 6%, in the single 52-week trial than in trials of shorter duration (Fig. 2), but not more often than with NSAID.
Any gastrointestinal adverse event
The proportion of patients reporting any gastrointestinal adverse event was of the order of 25% (Table 4). More patients taking celecoxib than placebo reported a gastrointestinal adverse event (NNH 14; 12 to 19). There was no difference between celecoxib and either paracetamol or rofecoxib. Celecoxib had fewer patients reporting any gastrointestinal adverse event than either NSAID (NNTp 12; 10 to 13 for licensed doses) or any active comparator.
Gastrointestinal tolerability
Gastrointestinal tolerability (the proportion of patients having moderate or severe nausea, dyspepsia, or abdominal pain) was about 5% with celecoxib (Table 5). There was no difference between celecoxib and placebo, paracetamol, or rofecoxib. Celecoxib had less gastrointestinal intolerance than NSAIDs (NNTp 28; 24 to 36 for licensed doses of celecoxib) or any active comparator.
Nausea
The proportion of patients reporting nausea was about 3% with celecoxib (Table 5a). Nausea was significantly lower with celecoxib than placebo (NNTp 155; 71 to 840), and for celecoxib at any dose compared with NSAID or any active comparator. There was no difference between celecoxib and paracetamol, or rofecoxib, or between licensed doses of celecoxib and NSAIDs.
Vomiting
The proportion of patients experiencing vomiting was about 1% with celecoxib (Table 5a). There was no difference between celecoxib and placebo, paracetamol, or rofecoxib. Celecoxib at both licensed dose and any dose had fewer patients with vomiting than NSAID (NNTp 173; 115 to 350 for licensed doses) or any active comparator.
Abdominal pain
The proportion of patients reporting abdominal pain was about 5% with celecoxib (Table 5b). There was no difference between celecoxib and placebo, or paracetamol. Celecoxib (any dose) produced less abdominal pain than rofecoxib 25 mg (NNTp 67; 35 to 920). Celecoxib at both licensed dose and any dose had fewer patients reporting abdominal pain than NSAID (NNTp 41; 32 to 57 for licensed doses) or any active comparator.
Dyspepsia
The proportion of patients reporting dyspepsia was about 7% with celecoxib (Table 5b). Celecoxib (any dose) produced more dyspepsia than placebo (NNH 46; 32 to 84). There was no difference between celecoxib and paracetamol, or rofecoxib. Celecoxib at both licensed and any dose had fewer patients reporting dyspepsia than NSAID (NNTp 61; 43 to 100 for licensed doses) or any active comparator.
Diarrhoea
The proportion of patients experiencing diarrhoea was about 6% with celecoxib (Table 5b). Celecoxib (any dose) produced more diarrhoea than placebo (NNH 53; 37 to 97). Celecoxib (any dose) produced less diarrhoea than paracetamol 4,000 mg (NNTp 41; 22 to 450). There was no difference between celecoxib and rofecoxib, or between celecoxib (at the licensed dose or any dose) and NSAID, or any active comparator.
Clinical ulcers and bleeds
Clinical ulcers and bleeds in the company clinical trial reports were as reported by investigators, and were not subjected to independent, blinded adjudication in trials where this was not a primary outcome. The proportion of patients having a clinical ulcer or bleed was under 0.5% with celecoxib (Table 5b). No analysis was possible for clinical ulcers and bleeds for the comparisons between celecoxib and placebo, paracetamol, and rofecoxib, as there were only three events, no events, and one event, respectively. Celecoxib at both the licensed dose and any dose had fewer patients with clinical ulcers and bleeds than NSAID (NNTp 250; 170 to 450 for licensed doses) or any active comparator.
Myocardial infarction
Myocardial infarction in the company clinical trial reports was as reported by investigators, and was not subjected to independent, blinded adjudication. The numbers of reported myocardial infarctions in each arm of each trial are given in Additional file 3.
The proportion of patients having a myocardial infarction was under 0.3% with celecoxib (Table 6). No analysis was possible for myocardial infarction for the comparisons between celecoxib and placebo, paracetamol, and rofecoxib, as there were only 10 events, no events, and 1 event, respectively. Proportions for celecoxib at both the licensed dose and any dose were not significantly different from NSAID, any active comparator, any active comparator excluding rofecoxib, or any comparator, including both rofecoxib and placebo.
The numbers of events were small, with fewer than 60 cases of myocardial infarction in the largest comparison. Most trials had either no cases of myocardial infarction, or a single case in one of the treatment arms. No analysis demonstrated a statistical difference between celecoxib and any comparator (Table 6). For the comparison of all celecoxib doses with all comparators except rofecoxib, the number of events was 39/20,933 (0.19%) for celecoxib and 18/15,383 (0.12%) for comparators. For the comparison of licensed doses of celecoxib with NSAID, the number of events was 20/13,509 (0.15%) for celecoxib 200 to 400 mg daily and 3/8,309 (0.04%) for NSAID.
Forty-four cases of myocardial infarction occurred in the two largest trials (096 and 102), with 21,162 patients. Their planned duration was 12 and 52 weeks, and they had a combined actual duration of about 4.5 months. Here 29/12,787 (0.23%) of patients taking celecoxib (200 to 800 mg) suffered a myocardial infarction, compared with 15/8,375 (0.18%) on NSAID. The relative risk was 1.7 (0.88 to 3.2). Calculating the NNH gave a figure of 2,100 with a 95% confidence interval of 588 patients harmed to 1,337 patients where harm was prevented.
Cardiac failure
The proportion of patients with cardiac failure was under 0.2% with celecoxib (Table 6). No analysis was possible for the comparisons between celecoxib and placebo, paracetamol, and rofecoxib, as there were only 5 events, no events, and 10 events, respectively. Proportions for celecoxib at both the licensed dose and any dose were not significantly different from NSAID or any active comparator.
Raised creatinine
For the incidence of creatinine raised to 1.3 times the upper limit of normal or more, data were available only for the comparisons between celecoxib and placebo, celecoxib at licensed doses and NSAID, and celecoxib compared with any active comparator. There were no significant differences (Table 6). The proportion of any patient having raised creatinine was up to 1% with celecoxib.
Hypertension and aggravated hypertension
This outcome combined a new diagnosis of hypertension with aggravated hypertension in patients with an existing diagnosis of hypertension, but in whom changed or additional treatment was needed for control of hypertension. The proportion of any patient having hypertension or aggravated hypertension was 1 to 2% with celecoxib (Table 6). There was no significant difference between celecoxib and any comparator, placebo, rofecoxib, or NSAIDs. For paracetamol there were only four events.
Oedema at any site
Oedema was reported in various ways in the trials, occasionally just as oedema, sometimes broken down by body site. The proportion of patients with oedema was usually about 3% (Table 6), but it was much higher at 23 to 38% in two trials (149, 181) in patients with osteoarthritis and treated hypertension, with oedema as a predefined end point. Proportions were 5 to 10% in another trial (002) in patients with osteoarthritis, diabetes, and hypertension, also with oedema as a predefined end point.
Celecoxib was associated with significantly more oedema than placebo (NNH 79; 54 to 145). Celecoxib was no different from paracetamol. Celecoxib (200 mg daily) had significantly less oedema than rofecoxib (25 mg daily), with an NNTp of 14 (10 to 25). Celecoxib at licensed doses or at any dose was no different from NSAID for oedema (Fig. 3), but was significantly better than any active comparator (NNTp 62; 48 to 87).
Haemoglobin fall of 20 g/L or more
This parameter was not reported in studies comparing celecoxib with paracetamol or rofecoxib. The incidence of a haemoglobin fall of 20 g/L or more was about 1% with celecoxib (Table 7). There was no difference between celecoxib and placebo. Celecoxib at both the licensed dose and any dose had a lower incidence than NSAID (NNTp 92; 66 to 150 for licensed doses) or any active comparator.
Haematocrit fall of 5% or more
This parameter was not reported in studies comparing celecoxib with paracetamol. The incidence of a haematocrit fall of 5% or more was about 10% with celecoxib (Table 7). There was no difference between celecoxib and placebo or rofecoxib. Celecoxib at both the licensed dose and any dose had a lower incidence than NSAID (NNTp 18; 14 to 25 for licensed doses) or any active comparator.
Endoscopically detected ulcers
Seven trials were designed to ascertain the presence of endoscopically detectable ulcers of 3 mm or more; in these, celecoxib was compared with placebo and/or NSAID (Additional file 4). Six reported at 12 weeks, and one at 24 weeks. Five trials also reported results according to the use of low-dose aspirin of 325 mg or less daily. These results are shown in Table 8 and Fig. 4, analysed across all patients and according to aspirin use. In no comparison was there any significant difference between celecoxib and placebo. For both celecoxib and NSAID, there was the same 6% absolute increase in endoscopically detected ulcers with aspirin use. Celecoxib, at both the licensed dose and any dose, always produced more endoscopically detected ulcers than NSAID. The NNTp was the same at 7 to 8 both with and without concomitant aspirin use.
Deaths
There were 28 deaths during the trials or within 28 days of stopping medication, of which 21 were cardio/cerebrovascular, 1 was of unknown cause, and 6 were due to other causes. We included the unknown with the cardiovascular deaths for analysis. The incidence with celecoxib was 0.01% (1/6,844) compared with 0.03% (1/3,060) with placebo, and 0.01% (1/13,904) with licensed doses of celecoxib compared with 0.07% (6/8,704) with NSAIDs. When all doses of celecoxib were analysed, the incidence was 0.03% (6/18,325), compared with 0.11% (14/12,685) with NSAIDs and 0.10% with all active comparators.
Discussion
There have been a number of systematic reviews of published papers of coxibs in arthritis, and several have examined specific adverse events. Serious upper gastrointestinal events in phase II and III studies were reported for rofecoxib [23] and celecoxib [24]. Others have looked at renal [25] or cardiac adverse events [26]. Cochrane reviews of cyclooxygenase inhibitors in rheumatoid arthritis have limited information to date on efficacy and safety of rofecoxib [27], and only five trials with 5,400 patients taking celecoxib [28]. Two previous systematic reviews of coxibs used company clinical trial reports. Deeks and co-workers [29] examined 15,000 patients in nine of the earlier trials of celecoxib, and Edwards and co-workers [30] examined some 5,700 patients in nine trials of valdecoxib.
Reviews looking at adverse events generally [29,30] have analysed adverse events by combining the absolute proportions of patients experiencing an adverse event, using the intention-to-treat population (randomised, at least one dose of drug) as the denominator. Those examining particular, rare adverse events (gastrointestinal bleeding, cardiovascular events) have tended to use exposure correction, together with independent blinded adjudication of the event [16,25,26].
This systematic review greatly increases the quantity and quality of information available on adverse events with celecoxib in arthritis. We had data from 31 trials, with almost 40,000 patients. The individual trials all scored the maximum on two systems for scoring reporting quality and validity in pain trials. Use of similar methods for collecting and reporting adverse events ensured data of uniform nature and quality.
The average age in the trials was about 60 years (Table 1), but there was a wide range (17 to 96 years). Several studies recruited special groups, for instance, patients with diabetes or hypertension, or patients who were solely Asian, or of mixed Asian, Afro-Caribbean, or Hispanic descent. Most trials documented relevant medical history, such as previous NSAID use or intolerance, or use of prophylactic low-dose aspirin. While non-Caucasians were under-represented, and many patients with significant comorbidities were excluded from the trials, this population is probably as representative as possible in clinical trials.
This gives credibility to the review in terms of size, quality, and validity, allowing us to make sense of all but the most rare adverse event. At the same time, there are limitations.
Multiple comparisons could be made, including condition treated, duration of study, comparator drug, and dose. Ideally all these would be tested by sensitivity analysis. We limited our analyses to comparator and dose to avoid excessive subdivision and proliferation of statistical testing, which can lead to spurious statistical significance [31]. Analysis by condition or duration was avoided because few patients (8%) were in trials with rheumatoid arthritis only, and few observations (23%) were made in trials lasting less than 12 weeks. Instead we concentrated on analysis by comparator, where there was the possibility of major differences based on large amounts of high-quality experimental evidence, and on dose. Most doses were in the licensed range, but for completeness we chose to perform analyses of celecoxib versus NSAID by all doses, and those within the licensed range.
Generally, trial reports indicate that World Health Organization Adverse Reaction Terminology criteria were used to define adverse events, but these are not immediately accessible. For any particular treatment-emergent adverse event, we have had to assume that the same criteria were used consistently both within and between trials. Although adequate clinical trial monitoring makes this highly probable, we have no positive evidence that this was the case. Definitions are problematical for reporting adverse events [32,33].
The statistical direction of the results for each adverse event outcome and each comparison is shown in Fig. 5.
In comparison with placebo (10,000 patients), celecoxib had fewer all-cause and lack-of-efficacy discontinuations, but more adverse events. Lower discontinuations result from greater efficacy, but an active drug at an effective therapeutic dose is likely to produce some adverse events. Importantly, there was no difference in gastrointestinal tolerability or endoscopically detected ulceration.
Only two trials (1,056 patients) compared celecoxib (200 mg/day) with paracetamol 4,000 mg/day. There were fewer all-cause and lack-of-efficacy discontinuations with celecoxib, and almost identical adverse event profiles, indicating better efficacy with no excess harm. It is worth noting that recent large randomised comparisons of paracetamol with placebo over 12 weeks have failed to show any better efficacy for paracetamol than placebo [34].
Five trials (2,671 patients) compared celecoxib (200 mg/day) with rofecoxib (25 mg/day). Celecoxib had less abdominal pain and oedema. Rofecoxib is another cyclooxygenase-2 selective inhibitor, and similarity between their adverse event profiles is to be expected.
In the comparisons with NSAIDs, the better adverse event profile of celecoxib was marked, both at licensed doses (23,000 patients) and any dose (31,000 patients). There were more discontinuations for lack of efficacy with celecoxib at licensed doses than with NSAIDs, balanced by fewer adverse-event discontinuations or gastrointestinal-adverse-event discontinuations. There were fewer adverse events overall, treatment-related adverse events, combined and individual gastrointestinal adverse events, with the exception of diarrhoea, but including gastrointestinal tolerability, and endoscopically detected ulcers. There were also possible benefits relating to loss of blood in the lower gastrointestinal tract, with fewer patients having falls in haemoglobin or haematocrit. These results again are expected, and are similar to results for celecoxib, valdecoxib, and rofecoxib in recent analyses and a trial [35-37].
Cyclooxygenase-2 selective inhibitors are known to produce fewer upper gastrointestinal ulcers and bleeds [38-42], and less gastrointestinal upset [43], than NSAIDs. The results here confirm this for celecoxib. For gastrointestinal tolerability (moderate or severe nausea, dyspepsia, or abdominal pain), one patient fewer would suffer for every 28 treated with celecoxib than with NSAID. One in 17 would not have a haematocrit fall of 5% or more.
The lack of difference between celecoxib and NSAIDs with regard to cardio-renal adverse events is not unexpected. There are no known benefits for cyclooxygenase-2 selective inhibitors over nonspecific inhibitors relating to cardiac or renal function, and the known associations between NSAID use and renal failure [8] and heart failure [10,11] are likely to apply to cyclooxygenase-2 selective inhibitors.
Endoscopically detected ulcers were affected both by whether celecoxib or NSAID was used, and by whether or not prophylactic low-dose aspirin was used (Table 8). The number-needed-to-treat to prevent one endoscopically detected ulcer was about 7, with or without aspirin. The protective effect of celecoxib was the same whether aspirin was present or not, and use of aspirin increased endoscopically detected ulcers by the same absolute incidence of 6%. This was nearly identical to results found in a systematic review of studies of valdecoxib in arthritis [30], but different comparisons make it difficult to know whether rofecoxib is different [37] (Fig. 6). The much lower incidence of endoscopically detected ulceration with celecoxib compared with NSAID reflected a similar result for rofecoxib [44,45], though the rofecoxib studies had no patients using aspirin. What is clear is that celecoxib plus low-dose aspirin produces no more endoscopically detected ulcers than NSAID without aspirin, and fewer than NSAID plus aspirin.
On maximum-dose NSAID, or celecoxib, or paracetamol, up to 30% of patients withdrew from treatment. The main reasons were lack of efficacy or adverse events. Withdrawals increased with duration of study, as would be expected (Table 2). They were also influenced by drug and dose (Table 2), though small numbers of events hindered comparisons. The tendency for fewer withdrawals with celecoxib than NSAID mirrors what has been found in clinical practice [46], though not in clinical trials of valdecoxib [30], based on many fewer patients than in this review. Overall medical costs of cyclooxygenase-2 selective inhibitors are not different from those of NSAIDs [46,47], because higher acquisition costs of cyclooxygenase-2 selective inhibitors appear to be balanced by higher costs of treating or preventing adverse events with NSAIDs.
Even with as large a data set as here, some rare but serious adverse events occur in so few people that it is difficult to determine whether apparent differences (significant or nonsignificant) between treatments are real or meaningful. Examples are cardiac failure, myocardial infarction, and death, with total maximum numbers of 55, 59, and 28 respectively. The incidence of these events was of the order of 0.3 per 1,000 patients to 2 per 1,000 patients. Cardiac failure and death with celecoxib were lower than with NSAIDs (but not significantly), while myocardial infarction rates were higher (but not significantly). Incidence may be additionally affected by exposure bias, different exposure with different treatments (Table 2). Analysis correcting for exposure bias may then be more appropriate [16], even though there appears to be little exposure bias between celecoxib and NSAIDs in arthritis trials.
Where adverse events are rare, even large numbers of patients in randomised, controlled trials will accumulate few events. If such trials are of relatively short duration, then there is even less opportunity to accumulate these rare events. In the 31 trials in this review, the longest duration of exposure was an average of about 7 months, and most were less than 3 months. The consequence is a residual uncertainty, as here for attributing higher risk of myocardial infarction with celecoxib than with other non-coxib comparators. Limitations in number of events and duration of constituent trials means that any possible relationship between celecoxib and myocardial infarction cannot be completely dispelled by these data alone, despite lack of a statistically significant difference.
Conclusion
This review of a large number of randomised trials and patients provides more accurate estimates of frequency and more confidence in the adverse event pattern. These are likely to be the minimum expected in clinical practice, where the population may be sicker, or take more medications, than in clinical trials.
Using company clinical trial reports removes some of the problems of selective reporting in published papers due to strict word limitations. Here the company clinical trial reports and extensive trial summaries provided about five pages of information per patient. While efficacy in published studies is poorly presented [48], it is available in clinical reports [49]. Information about adverse events is even more poorly presented in published papers [50], but it is clearly presented in company clinical trial reports. Company clinical trial reports represent an ideal source of information for systematic review and meta-analysis.
Abbreviations
NNH = number-needed-to-harm; NNT = number-needed-to-treat; NNTp = number-needed-to-treat to prevent one event; NSAID = nonsteroidal anti-inflammatory drug.
Competing interests
RAM and HJM have received lecture fees from pharmaceutical companies. The authors have received research support from charities and government sources at various times. GM is an employee of Pfizer. This work was supported by an unrestricted educational grant from Pfizer Ltd. The terms of the financial support from Pfizer included freedom for authors to reach their own conclusions, and an absolute right to publish the results of their research, irrespective of any conclusions reached. Pfizer did have the right to view the final manuscript before publication, and did so. No author other than GM has any direct stock holding in any pharmaceutical company.
Authors' contributions
RAM was involved with planning the study, data extraction, analysis, and preparing the manuscript; SD with data extraction, analysis, and writing; GTM with planning, data extraction, analysis, and writing the manuscript; HJM with planning, analysis, and writing. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
A Word file containing a table giving trial inclusion and exclusion criteria for arthritis.
Click here for file
Additional File 2
A Word file containing a table giving definitions of adverse events.
Click here for file
Additional File 3
An Adobe Acrobat file containing a table showing details of treatment, discontinuations, and adverse events.
Click here for file
Additional File 4
An Adobe Acrobat file containing a table showing detailed endoscopic outcomes.
Click here for file
Figures and Tables
Figure 1 Scatter plot of trials comparing celecoxib with NSAID for discontinuations due to gastrointestinal adverse events. Celecoxib at any dose is represented. The red symbol represents the longest trial, at 52 weeks. GI, gastrointestinal; NSAID, nonsteroidal anti-inflammatory drug.
Figure 2 Scatter plot of trials comparing any dose of celecoxib with NSAID for serious adverse events. The red symbol represents the longest trial, at 52 weeks. AE, adverse events; NSAID, nonsteroidal anti-inflammatory drug.
Figure 3 Scatter plot of trials comparing any dose of celecoxib with NSAID for oedema. The red symbol represents the longest trial, at 52 weeks. NSAID, nonsteroidal anti-inflammatory drug.
Figure 4 Endoscopically identified ulcers in patients taking celecoxib and NSAID, with and without prophylactic low-dose aspirin. NSAID, nonsteroidal anti-inflammatory drug.
Figure 5 Summary of main comparisons of treatment of arthritis with celecoxib or other drugs and placebo. AE, adverse event; GI, gastrointestinal; L, licensed doses; LoE, lack of efficacy; NSAID, nonsteroidal anti-inflammatory drug.
Figure 6 Endoscopically identified ulcers after treatment of arthritis with placebo, low-dose aspirin, NSAID, coxibs, and combinations. A, aspirin; N, NSAID; NSAID, nonsteroidal anti-inflammatory drug; C, coxib.
Table 1 Included studies of tolerability, adverse events, and endoscopically detected ulceration associated with celecoxib in arthritis
Drug, dose, number randomised
Study Details of participants Relevant medical history Celecoxib Placebo Other Duration (weeks) Efficacy outcomes Safety outcomes Total in trial (ITT)
Osteoarthritis
C-002 OA Hip/Knee (ACR) requiring daily NSAID therapy, FCC 1–3
Stable hypertension, type 2 diabetes
Age 62 (range 40–89) years 61% female ≥ 75% Caucasian Data not provided 1 × 200 mg/day, n = 36 No placebo Rofecoxib 1 × 25 mg/day, n = 132
Naproxen 2 × 500 mg/day, n = 128 12 WOMAC
Patient's assessment of arthritis pain
VAS
Patient's global assessment of arthritis
Patient's satisfaction Withdrawal due to lack of efficacy Withdrawals Adverse events
Serious adverse events
Laboratory tests 396
C-003 OA Knee (ACR) with flare, requiring daily NSAID/analgesic, FCC 1–3, baseline pain 40 on 100 mm VAS.
Age 63 (range 39–90) years Duration of disease 8 (range 0.2–51) years 67% female ≥ 85% Caucasian Cardioprotective ASA 20% NSAID intolerance 4% GI ulcer 6% GI bleed 1% Renal insufficiency 1% 1 × 200 mg/day, n = 189 n = 96 Rofecoxib 1 × 25 mg/day, n = 190 6 Patient's assessment of arthritis pain WOMAC (total)
Patient's global assessment of arthritis pain
VAS
OASI
Physician's global assessment of arthritis Patient's assessment of satisfaction Withdrawals Adverse events
Serious adverse events
Laboratory tests 475
C-010 OA Hip/Knee (K-L 2–4), requiring chronic NSAID/analgesic, initial pain 40–90 on 100 mm VAS
Age 63 (range 38–91) years Duration of disease 9 (0.1–54) years 62% female Cardioprotective ASA 20% GI-related NSAID intolerance 1% Gastroduodenal ulcer 8% GI bleed 0.6% Some type of GI history (unspecified) 48% 1 × 200 mg/day n = 181 n = 172 Paracetamol 4 × 1,000 mg/day, n = 171 6 WOMAC index MDHAQ
Patient's global rating of helpfulness Physician's global assessment of status SF-36 General clinical safety Withdrawals Adverse events
Serious adverse events
Laboratory tests 524
C-013 OA Knee (ACR) with flare, FCC 1–3 Mean age 62 (range 29–92) years
Duration of disease 10 (0.2–50) years
69% female
90% Caucasian Cardioprotective ASA permitted NSAID intolerance 13% Gastroduodenal ulcer 16% GI bleed 3% CVD 52% 2 × 40 mg/day, n = 73 2 × 100 mg/day, n = 75 2 × 200 mg/day, n = 73 n = 70 No active comparator 2 Physican's global assessment
Patient's global assessment
Patient's arthritis pain SF-36 Withdrawals Adverse events
Serious adverse events
Laboratory tests 291
C-020 OA knee/hip (ACR) with flare, FCC 1–3 Age 62 (range 21–89) years Duration of disease 9 (0.1–52) years
66% female
79% Caucasian Cardioprotective ASA permitted
NSAID
intolerance 7% Gastroduodenal ulcer 9%
GI bleed 2%
CVD 53% 2 × 50 mg/day, n = 218
2 × 100 mg/day, n = 217
2 × 200 mg/day, n = 222 n = 219 Naproxen 2 × 500 mg/day, n = 216 12 Patient's global assessment
Physician's global assessment
WOMAC Patient's assessment of pain Withdrawals Adverse events
Serious adverse events
Laboratory tests 1,092
C-021 OA Knee/Hip (ACR) with flare, FCC 1–3 No ulcer at baseline endoscopy Age 61 (range 22–89) years Duration of disease 9 (range 0.1–52)
years
54% female
83% Caucasian Cardioprotective ASA permitted.
NSAID
intolerance 10% Gastroduodenal ulcer 17%
GI bleed 2%
CVD 60% 2 × 50 mg/day, n = 258
2 × 100 mg/day, n = 239
2 × 200 mg/day, n = 237 n = 247 Naproxen 2 × 500 mg/day, n = 233 12 Patient's global assessment
Patient's assessment of pain Physician's global assessment
WOMAC Withdrawals Adverse events
Serious adverse events
Laboratory tests Endoscopic ulcers 1,214
C-042 Symptomatic OA Hip/Knee (ACR) ≥ 6 months, requiring NSAID, FCC 1–3
Age 63 (range 34–91) years Duration of disease 7 (0.5–48) years
72% female
94% Caucasian NSAID
intolerance 2% Gastroduodenal ulcer 3%
GI bleed 0.5%
CVD 45% 2 × 100 mg/day, n = 346 No placebo Diclofenac 2 × 50 mg/day, n = 341 6 Patient's global assessment
Patient's assessment of pain Physician's global assessment
SF-36 Withdrawals Adverse events
Serious adverse events
Laboratory tests 667
C-047 OA Knee (ACR) with flare, FCC 1–3
Age 63 (29–91) years Duration of disease 9 (0.5–60) years
72% female
84% Caucasian Cardioprotective ASA permitted
NSAID
intolerance 10%
Gastroduodenal ulcer 10%
GI bleed 4%
CVD 62% 2 × 25 mg/day, n = 100
2 × 100 mg/day, n = 101
2 × 400 mg/day, n = 99 n = 101 No active comparator 4 Patient's global assessment
Patient's assessment of pain Physician's global assessment
SF-36 WOMAC Withdrawals Adverse events
Serious adverse events
Laboratory tests 401
C-054 OA Hip (ACR) with flare, FCC 1–3
Age 62 (28–93) years
Duration of disease 7 (0.1–64) years
66% female
92% Caucasian Cardioprotective ASA permitted
NSAID
intolerance 13% Gastroduodenal ulcer 12%
GI bleed 2%
CVD 60% 2 × 50 mg/day, n = 216
2 × 100 mg/day, n = 207
2 × 200 mg/day, n = 213 n = 217 Naproxen 2 × 500 mg/day, n = 207 12 Patient's global assessment
Patient's assessment of pain Physician's global assessment
SF-36 WOMAC Withdrawals Adverse events
Serious adverse events
Laboratory tests 1,060
C-060 OA Knee (ACR) with flare, FCC 1–3
Age 63 (29–88) years
Duration of disease 9 (0.1–59) years
66% female
88% Caucasian Cardioprotective ASA permitted
NSAID
intolerance 4%
Gastroduodenal ulcer 6%
GI bleed 2%
CVD 58% 2 × 100 mg/day, n = 231
1 × 200 mg/day, n = 222 n = 231 No active comparator 6 Patient's global assessment
Patient's assessment of pain Physician's global assessment
SF-36
WOMAC Withdrawals Adverse events
Serious adverse events
Laboratory tests 684
C-087 OA Knee (ACR) with flare, FCC 1–3
Age 61 (18–89) years
Duration of disease 9 (0.1–60) years
70% female
86% Caucasian Cardioprotective ASA permitted
NSAID
intolerance 7%
Gastroduodenal ulcer 16%
GI bleed 2%
CVD 64% 2 × 100 mg/day, n = 241
1 × 200 mg/day, n = 231 n = 243 No active comparator 6 Patient's global assessment
Patient's assessment of pain Physician's global assessmen
WOMAC Withdrawals Adverse events
Serious adverse events Laboratory tests 715
C-096 OA
Knee/Hip/Hand ≥ 6 months (ACR) requiring daily
analgesic/ NSAID, FCC 1–3 Age 62 (range 21–96) years
76% female Duration of disease 7 (0.3–59) years Cardioprotective ASA use 7%
CVD 41%
Renal insufficiency 0.2%
Respiratory disease 5%
Diabetes 8% 2 × 100 mg/day, n = 4,393
2 × 200 mg/day, n = 4,407 No placebo Naproxen 2 × 500 mg/day, n = 905
Diclofenac 2 × 50 mg/day, n = 3,489 12 Patient's global rating of arthritis
Patient's assessment of pain (VAS)
WOMAC
Physician's global assessment of arthritis Withdrawals Adverse events
Serious adverse events
Laboratory tests 13,194
C-118 OA Knee (ACR) with flare, FCC 1–3
Age 61 (29–88) years
Duration of disease 8 (0.1–62) years
65% female
82% Caucasian Cardioprotective ASA permitted
NSAID intolerance 3% Gastroduodenal ulcer 8%
GI bleed 1%
CVD 66% 2 × 100 mg/day, n = 199 n = 200 Diclofenac 3 × 50 mg/day, n = 199 6 Patient's global assessment
Patient's assessment of pain Physician's global assessment
WOMAC Withdrawals Adverse events
Serious adverse events
Laboratory tests 598
C-149 OA
Hip/Knee/Hand (ACR) requiring NSAID, FCC 1–3 Stable treated hypertension Age 74 (range 64–95) years Duration of disease range 0.3–61 years 67% female Majority Caucasian Cardioprotective ASA 38%
NSAID
intolerance 3%
Gastroduodenal ulcer 10%
GI bleed 3%
Oedema 26%
CHF 5% 1 × 200 mg/day, n = 411 No placebo Rofecoxib 25 mg/day, n = 399 6 Oedema Aggravated hypertension Renal events Withdrawals Adverse events
Serious adverse events Laboratory tests 810
C-152 OA Knee (ACR) with flare, FCC 1–3, baseline pain 35 on 100 mm VAS
Age 62 (range 40–88) years Duration of disease 11 (range 0.5–s47) years
71% female
80% Caucasian Cardioprotective ASA permitted
NSAID
intolerance 4%
Gastroduodenal ulcer 9%
GI bleed 0.5% 1 × 200 mg/day, n = 63 n = 60 Rofecoxib 1 × 25 mg/day, n = 59 6 Patient's assessment of arthritis pain
OA VAS
scale
Patient's global assessment of arthritis
WOMAC Withdrawals Adverse events
Serious adverse events
Laboratory tests 182
C-181 OA
Hip/Knee/Hand (ACR) requiring daily NSAID, FCC 1–3
Stable treated hypertension Age 73 (range 65–96) years Duration of disease 12 (0–63) years
62% female
88% Caucasian Cardioprotective ASA permitted
NSAID
intolerance 2%
Gastroduodenal ulcer 8%
GI bleed 2%
Oedema 27%
CHF 3% 1 × 200 mg/day, n = 549 No placebo Rofecoxib 1 × 25 mg/day, n = 543 6 Blood pressure Oedema Weight Anti-hypertensive medication Withdrawals Adverse events
Serious adverse events
Laboratory tests 1,092
C-209 OA Knee with flare (ACR), requiring chronic NSAID, FCC 1–3, initial pain 40–90 on 100 mm VAS
Age 58 (range 45–83) years Duration of disease 5 (range 0.1–36) years
80% female Afro-American population Data not provided 1 × 200 mg/day, n = 125 n = 66 Naproxen 2 × 500 mg/day, n = 125 6 Patient's assessment of arthritis pain
Patient's global assessment Physician's global assessment
WOMAC Withdrawals Adverse events
Serious adverse events
Laboratory tests 316
C-210 OA Knee (ACR) with flare, FCC 1–3, requiring daily therapy, baseline pain 40–90 on 100 mm
VAS
Age 65 (range 42–90) years 68% female Duration of disease 5 (0.3–38) years
Asian American population 100% Asian descent Data not provided 1 × 200 mg/day, n = 145 n = 76 Naproxen 2 × 500 mg/day, n = 141 6 Patient's assessment of arthritis pain
Patient's global assessment
Physician's global assessment
Pain Satisfaction
WOMAC Withdrawals Adverse events
Serious adverse events
Laboratory tests 362
C-211 OA Knee (ACR) with flare, requiring daily NSAID, FCC 1–3, baseline pain 40–90 on 100 mm VAS
Age 60 (range 40–88) years Duration of disease 6 (range 0.1–36 yrs) years
67% female Hispanic population Data not provided 1 × 200 mg/day, n = 125 n = 61 Naproxen 2 × 500 mg/day, n = 129 6 Patient's assessment of arthritis pain
Patient's global assessment
Physician's global assessment
WOMAC
Patient's satisfaction Withdrawals Adverse events
Serious adverse events
Laboratory tests 315
C-216 OA Knee, symptomatic, requiring NSAID, initial pain 40 on 100 mm VAS
Age 63 (range 20–92) years Duration of disease 4 (range 0.1–37) years
66% female Asian population Cardioprotective ASA 3%
NSAID
intolerance 0.1%
GI bleed 0.2%
Gastroduodenal ulcer 6%
CVD 30% 2 × 100 mg/day, n = 382 n = 192 Loxoprofen 3 × 60 mg/day, n = 385 4 Final global improvement rating
Patient's assessment of arthritis pain Physician's and
patient's global assessment of arthritis
WOMAC Withdrawals Adverse events
Serious adverse events
Laboratory tests Global safety rating 959
C-249 OA Hip/Knee (K-L confirmed), baseline pain 40–90 on 100 mm
VAS
Age 63 (range 45–89) years Duration of disease 9 (range 0.1–50) years
66% female
≥ 80% Caucasian Cardioprotective ASA 21% GI-related
NSAID
intolerance 2%
Gastroduodenal ulcer 7%
GI bleed 0.7% 1 × 200 mg/day, n = 189 n = 182 Paracetamol 4 × 1,000 mg/day, n = 185 2 × 6 crossover WOMAC
MDHAQ
Investigator global assessment
Patient's assessments of helpfulness and arthritis
SF-36 Withdrawals Adverse events
Serious adverse events
Laboratory tests 556
Rheumatoid arthritis
C-012 Adult RA with flare (ACR) ≥ 6 months, requiring NSAID, FCC 1–3 Age 56 (range 21–86) years Duration of disease 11 (range 0.5–50) years
78% female
84% Caucasian Cardioprotective ASA permitted
NSAID
intolerance 9%
Gastroduodenal ulcer 3%
GI bleed 0.6%
CVD 43% 2 × 40 mg/day, n = 80
2 × 200 mg/day, n = 82
2 × 400 mg/day, n = 81 n = 84 No active comparator 4 Patient's global rating of arthritis
Arthritis pain, joint
tenderness, joint swelling Withdrawals Adverse events
Serious adverse events
Laboratory tests 327
C-022 RA with flare (ACR) requiring NSAID, FCC 1–3 No ulcer at baseline endoscopy Age 54 (range 20–90) years Duration of disease 10 (0.3–58) years
73% female
86% Caucasian Cardioprotective ASA permitted
NSAID
intolerance 10%
Gastroduodenal ulcer 15%
GI bleed 2%
CVD 44% 2 × 100 mg/day, n = 240
2 × 200 mg/day, n = 235
2 × 400 mg/day, n = 217 n = 231 Naproxen 2 × 500 mg/day, n = 225 12 Patient's global assessment of arthritis
Physician's global assessment of arthritic
condition No. of swollen joints
ACR-20
responder index No. of tender/painful joints Withdrawals Adverse events
Serious adverse events
Laboratory tests Endoscopic ulcers 1,148
C-023 RA (ACR) with flare requiring NSAID, FCC 1–3 Age 55 (range 21–84) years Duration of disease 10 (range 0.3–60) years
73% female
86% Caucasian Cardioprotective ASA permitted
NSAID
intolerance 10%
Gastroduodenal ulcer 8%
GI bleed 1%
CVD 44% 2 × 100 mg/day, n = 228
2 × 200 mg/day, n = 218
2 × 400 mg/day, n = 217 n = 221 Naproxen 2 × 500 mg/day, n = 218 12 Patient's global assessment of arthritis
Physician's global assessment of arthritic
condition No. of swollen joints
ACR -20 responder index No. of tender/painful joints Withdrawals Adverse events
Serious adverse events
Laboratory tests 1,102
C-041 Adult onset RA (ACR) ≥ 6 months, requiring NSAID, FCC 1–3 No ulcer at baseline endoscopy Age 55 (range 20–85) years Duration of disease10 (0.6–53) years
73% female
98% Caucasian Cardioprotective ASA not permitted
NSAID
intolerance 7%
Gastroduodenal ulcer 8%
GI bleed 0.7%
CVD 25% 2 × 200 mg/day, n = 326 No placebo Diclofenac (slow release) 2 × 75 mg/day, n = 329 24 Patient's global assessment
Physician's global assessment
Swollen joints Patient's assessment of arthritis pain
SF-36 Withdrawals Adverse events
Serious adverse events
Laboratory tests
Endoscopic ulcers (not all patients had endoscopy) 655
Osteoarthritis and rheumatoid arthritis
C-062 OA/RA ≥ 3 months, requiring NSAID, FCC 1–3 No ulcer at baseline endoscopy Duration of OA 10 (0.3–50) years, RA 10 (0.4–43) years Age 57 (range 22–86) years
67% female
83% Caucasian Cardioprotective ASA permitted
NSAID
intolerance 13%
Gastroduodenal ulcer 20%
GI bleed 4%
CVD 53% 2 × 200 mg/day, n = 269 No placebo Naproxen 2 × 500 mg/day, n = 267 12 Patient's global assessment
Physcian's global assessment
SF-36 Withdrawals Adverse events
Serious adverse events
Laboratory tests
Endoscopic ulcers 536
C-071 OA/RA ≥ 3 months, requiring NSAID, FCC 1–3 No ulcer at baseline Age 57 (22–87) years Duration of disease 10 (0.3–48) years
68% female
82% Caucasian Cardioprotective ASA permitted
NSAID
intolerance 7%
Gastroduodenal ulcer 12%
GI bleed 2%
CVD 42% 2 × 200 mg/day, n = 365 No placebo Diclofenac 2 × 75 mg/day, n = 387
Ibuprofen 3 × 800 mg/day, n = 345 12 Patient's global assessment
Physcian's global assessment
SF-36 Withdrawals Adverse events
Serious adverse events
Laboratory tests
Endoscopic ulcers 1,097
C-102 OA/RA, requiring NSAID >3 months Age 60 (range 18–90) years
69% female
88% Caucasian Cardioprotective ASA permitted
NSAID
intolerance 9%
Gastroduodenal ulcer 8%
GI bleed 2%
CVD 40% 2 × 400 mg/day, n = 3,987 No placebo Ibuprofen 3 × 800 mg/day, n = 1,985
Diclofenac 2 × 75 mg/day, n = 1,996 52 Patient's global assessment
Patient's assessment of arthritis pain
SF-36
SODA Withdrawals Adverse events
Serious adverse events
Laboratory tests
CSUGIEs 7,968
C-105 OA/RA
(documented clinical diagnosis for ≥ 3 months), requiring NSAID, FCC 1–3 Age 50 (range 17–78) years Duration of disease not given 84% female Asian population Cardioprotective ASA permitted
Gastroduodenal ulcer 0.5%
GI bleed 0.02%
CVD 1% 2 × 100 mg/day, n = 327 No placebo Diclofenac 2 × 50 mg/day, n = 330 12 Patient's global assessment
Physcian's global assessment
Patient's assessment of arthritis pain Withdrawals Adverse events
Serious adverse events
Laboratory tests
Endoscopic ulcers 657
C-106 OA/RA
(documented clinical diagnosis), requiring NSAID, FCC 1–3 Age 55 (range 18–80) years Duration of disease not given 17% female ≥ 99% Asian Cardioprotective ASA permitted Gastroduodenal ulcer 9% GI bleed 3% CVD 10% 2 × 100 mg/day, n = 63 No placebo Diclofenac 2 × 50 mg/day, n = 61 12 Patient's global assessment Physcian's global assessment Patient's assessment of arthritis pain Withdrawals Adverse events Serious adverse events Laboratory tests Endoscopic ulcers 124
C-107 OA/RA (documented clinical diagnosis ≥ 3 months) requiring NSAID, FCC 1–3 Age 53 (range 24–88) years Duration OA 4 (0.5–13) years, RA 6 (0.5–19) years 83% female ≥ 99% Asian Cardioprotective ASA permitted Gastroduodenal ulcer 10% GI bleed 3% CVD 14% 2 × 100 mg/day, n = 44 No placebo Diclofenac 2 × 50 mg/day, n = 44 12 Patient's global assessment Physcian's global assessment Patient's assessment of arthritis pain Withdrawals Adverse events Serious adverse events Laboratory tests Endoscopic ulcers 88
C-849 (Pooled 105, 106, 107) OA/RA 2 × 100 mg/day, n = 434 No placebo Diclofenac 2 × 50 mg/day, n = 435 12 Endoscopic ulcers (pooled 105, 106, 107) 880
All trials had a quality score of 5/5, and a validity score of 16/16. ACR, American College of Rheumatology; ASA, acetylsalicylic acid; CHF, chronic heart failure; CSUGIE, clinically significant upper gastrointestinal event; CVD, cardiovascular disease; FCC, functional capacity class; GI, gastrointestinal; ITT, intention to treat; K-L, Kellgren-Lawrence; MDHAQ, Multidimensional Health Assessment Questionnaire; NSAID, nonsteroidal anti-inflammatory drug; OA, osteoarthritis; OASI, OA severity index; QS, quality score; RA, rheumatoid arthritis; SODA, sequential occupational dexterity index; VAS, visual analogue scale; VS, validity score; WOMAC, Western Ontario and McMaster Universities Osteoarthritis Index.
Table 2 Analysis of discontinuations by comparator, in studies of adverse events associated with celecoxib in arthritis
Number of Incidence of events (%)
Outcome and comparisons Celecoxib daily dose Comparator and daily dose Trials Patients Celecoxib Comparator Relative riska (95% CI) NNTpb or NNHc (95% CI)
All-cause discontinuation
Celecoxib v placebo Any Placebo 19 9,919 28 40 0.64 (0.61–0.68)a 8.4 (7–10)b
Celecoxib v paracetamol Any Paracetamol 4,000 mg 2 1,056 17 25 0.69 (0.54–0.88)a 13 (8–35)b
Celecoxib v rofecoxib Any Rofecoxib 25 mg 5 2,671 14 14 1.0 (0.8–1.2)
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 19 22,616 23 23 0.96 (0.91–1.01)
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 20 31,711 31 34 0.96 (0.93–0.99)a 28 (22–40)b
Celecoxib (any dose) v any active Any Any active comparator 26 35,302 29 32 0.95 (0.92–0.98)a 36 (27–57)b
Lack-of-efficacy discontinuation
Celecoxib v placebo Any Placebo 19 9,914 17 28 0.53 (0.49–0.57)a 9.0 (8–11)b
Celecoxib v paracetamol Any Paracetamol 4,000 mg 2 1,056 7.2 11 0.66 (0.45–0.97)a 27 (14–390)b
Celecoxib v rofecoxib Any Rofecoxib 25 mg 5 2,671 2.2 1.5 1.5 (0.84–2.6)
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 19 22,613 8.0 6.3 1.1 (1.02–1.23)a 58 (42–97)c
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 20 31,708 11.3 10.4 1.02 (0.96–1.1)
Celecoxib (any dose) v any active Any Any active comparator 26 35,299 10.6 9.6 1.0 (0.95–1.1)
Adverse-event discontinuation
Celecoxib v placebo Any Placebo 19 9,914 6.6 5.5 1.2 (0.97–1.4)
Celecoxib v paracetamol Any Paracetamol 4,000 mg 2 1,056 4.3 5.4 0.81 (0.47–1.4)
Celecoxib v rofecoxib Any Rofecoxib 25 mg 5 2,662 6.2 6.8 0.91 (0.68–1.2)
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 19 22,613 8.5 9.9 0.84 (0.77–0.92)a 74 (47–180)b
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 20 31,708 11.4 14.6 0.86 (0.81–0.91)a 31 (25–41)b
Celecoxib (any dose) v any active Any Any active comparator 26 35,299 10.9 13.5 0.87 (0.82–0.92)a 38 (30–51)b
Gastrointestinal-adverse-event discontinuation
Celecoxib v placebo Any Placebo 11 5,933 2.5 2.0 1.2 (0.8–1.7)
Celecoxib v paracetamol Any Paracetamol 4,000 mg 2 726 1.6 2.6 0.6 (0.2–1.6)
Celecoxib v rofecoxib Any Rofecoxib 25 mg 5 2,671 2.2 2.9 0.7 (0.5–1.2)
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 11 18,639 4.8 6.5 0.7 (0.6–0.8)a 58 (42–98)b
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 12 27,299 6.4 9.6 0.75 (0.7–0.8)a 31 (26–40)b
Celecoxib (any dose) v any active Any Any active comparator 18 30,560 6 8.7 0.75 (0.7–0.8)a 37 (30–48)b
aRelative risk: bold indicates statistically significant difference. bNNTp (number-needed-to-treat to prevent one event) is indicated by bold. cNNH (number-needed-to-treat to harm one patient). CI, confidence interval; NSAID, nonsteroidal anti-inflammatory drug.
Table 3 Discontinuations of treatment in arthritis because of lack of efficacy or adverse events
Lack-of-efficacy discontinuations Adverse-event discontinuations
Duration (weeks) Treatment Dose (mg/day) Number of events Total number Discontinuations, % (95% CI) Number of events Total number Discontinuations, % (95% CI)
2–6 Placebo 339 1,925 17.6 (15.8–19.4) 97 1,925 5.0 (4.0–6.0)
Celecoxib <100 42 253 16.6 (12.1–21.1) 8 253 3.2 (1.0–5.4)
Celecoxib 100 No data No data
Celecoxib 200 203 4,190 4.8 (4.2–5.4) 223 4,190 5.3 (4.7–5.9)
Celecoxib 400 12 155 7.7 (3.6–11.8) 5 155 3.2 (0.5–5.9)
Celecoxib 800 15 180 8.3 (4.2–12.4) 14 180 7.8 (3.9–11.7)
Paracetamol 4,000 55 502 11.0 (8.3–13.7) 27 502 5.4 (3.4–7.4)
Rofecoxib 25 19 1,191 1.6 (0.8–2.4) 77 1,191 6.5 (5.1–7.9)
Naproxen 1,000 5 395 1.3 (0.1–2.5) 31 395 7.8 (5.3–10.3)
Diclofenac 100/150 13 540 2.4 (1.0–3.8) 51 540 9.4 (6.9–11.9)
12 Placebo 521 1,135 45.9 (43.0–48.8) 70 1,135 6.2 (4.8–7.6)
Celecoxib 100 145 692 21 (18.1–23.9) 52 692 7.5 (5.5–9.5)
Celecoxib 200 571 6,094 9.4 (8.6–10.2) 488 6,094 8.0 (7.4–8.6)
Celecoxib 400 492 6,166 8.0 (7.4–8.6) 590 6,166 9.6 (8.8–10.4)
Celecoxib 800 128 435 29.4 (25.1–33.7) 28 435 6.4 (4.0–8.8)
Paracetamol 4,000 No data No data
Rofecoxib 25 1 132 0.8 (0.0–2.4) 13 132 9.8 (4.7–14.9)
Naproxen 1,000 374 2,399 15.6 (14.2–17.0) 316 2,399 13.2 (11.8–14.6)
Diclofenac 100/150 120 4,311 2.8 (2.2–3.4) 338 4,311 7.8 (7.0–8.6)
Ibuprofen 2,400 14 345 4.1 (1.9–6.3) 37 345 10.7 (7.4–14)
24+ Placebo No data No data
Celecoxib 100 No data No data
Celecoxib 200 No data No data
Celecoxib 400 26 326 8.0 (5.1–10.9) 34 326 10.4 (7.1–13.7)
Celecoxib 800 691 3,987 17.3 (16.1–18.5) 892 3,987 22.4 (21–23.8)
Paracetamol 4,000 No data No data
Rofecoxib 25 No data No data
Naproxen 1,000 No data No data
Diclofenac 100/150 331 2,325 14.2 (12.8–15.6) 593 2,325 25.5 (23.7–27.3)
Ibuprofen 2,400 456 1,985 23.0 (21.2–24.8) 456 1,985 23.0 (21.2–24.8)
CI, confidence interval.
Table 4 Analysis of arthritis patients according to gastrointestinal adverse events
Number of Incidence of events (%)
Outcome and comparisons Celecoxib daily dose Comparator and daily dose Trials Patients Celecoxib Comparator Relative riska (95% CI) NNTpb or NNHc (95% CI)
Patient with any adverse event
Celecoxib v placebo Any Placebo 19 9,919 55 48 1.08 (1.04–1.13)a 15 (11–21)c
Celecoxib v paracetamol Any Paracetamol 4,000 mg 2 1,056 32 32 1.0 (0.84–1.2)
Celecoxib v rofecoxib Any Rofecoxib 25 mg 3 769 48 49 0.97 (0.84–1.1)
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 19 22,615 45 50 0.92 (0.89–0.95)a 18 (14–23)b
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 20 31,711 53 60 0.96 (0.94–0.98)a 15 (13–18)b
Celecoxib (any dose) v any active Any Any active comparator 24 33,400 53 59 0.96 (0.94–0.98)a 17 (14–21)b
Patient with any treatment-related adverse event
Celecoxib v placebo Any Placebo 19 9,919 9.5 8.1 1.22 (1.06–1.40)a 71 (39–450)c
Celecoxib v paracetamol Any Paracetamol 4,000 mg 3 1,056 9.0 8.8 1.04 (0.71–1.5)
Celecoxib v rofecoxib Any Rofecoxib 25 mg 4 1,579 6.6 9.0 0.74 (0.53–1.04)
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 19 22,615 13.0 17.3 0.77 (0.72–0.82)a 24 (19–31)b
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 19 23,743 12.7 17.3 0.77 (0.72–0.82)a 22 (18–27)b
Celecoxib (any dose) v any active Any Any active comparator 24 26,242 12.3 16.2 0.78 (0.73–0.83)a 26 (21–33)b
Patient with any serious adverse event
Celecoxib v placebo Any Placebo 19 9,919 1.0 1.4 0.67 (0.46–0.98)a 280 (120–790)b
Celecoxib v paracetamol Any Paracetamol 4,000 mg 2 1,056 0.5 0.6 0.76 (0.14–4.1)
Celecoxib v rofecoxib Any Rofecoxib 25 mg 5 2,671 2.3 2.1 1.1 (0.68–1.9)
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 19 22,612 2.5 2.6 0.91 (0.77–1.08)
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 20 31,708 3.3 3.6 1.02 (0.91–1.15)
Celecoxib (any dose) v any active Any Any active comparator 26 35,299 3.2 3.4 1.02 (0.91–1.15)
Patient with any gastrointestinal adverse event
Celecoxib v placebo Any Placebo 17 9,512 26.0 19.0 1.2 (1.1–1.4)a 14 (12–19)c
Celecoxib v paracetamol Any Paracetamol 4,000 mg 2 1,056 12.0 11.0 1.1 (0.8–1.6)
Celecoxib v rofecoxib Any Rofecoxib 25 mg 5 2,671 16.0 18.0 0.87 (0.74–1.03)
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 18 30,043 26.0 34.0 0.84 (0.81–0.87)a 12 (10–13)b
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 18 31,171 26.0 34.0 0.84 (0.81–0.87)a 12 (10–13)b
Celecoxib (any dose) v any active Any Any active comparator 24 34,762 26.0 32.0 0.85 (0.82–0.88)a 14 (12–16)b
aRelative risk: bold indicates statistically significant difference. bNNTp (number-needed-to-treat to prevent one event) is indicated by bold. cNNH (number-needed-to-treat to harm one patient). CI, confidence interval; NSAID, nonsteroidal anti-inflammatory drug.
Table 5a Gastrointestinal adverse events reported in studies of arthritis patients (part 1)
Number of Incidence of events (%)
Outcome and comparisons Celecoxib daily dose Comparator and daily dose Trials Patients Celecoxib Comparator Relative riska (95% CI) NNTpb or NNHc (95% CI)
Gastrointestinal tolerability
Celecoxib v placebo Any Placebo 19 9,919 5.3 4.6 1.0 (0.82–1.2)
Celecoxib v paracetamol Any Paracetamol 4,000 mg 2 1,056 2.0 2.0 1.0 (0.43–2.4)
Celecoxib v rofecoxib Any Rofecoxib 25 mg 5 2,671 3.2 4.4 0.72 (0.49–1.06)
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 19 22,615 5.4 8.9 0.62 (0.56–0.68) 28 (24–36)b
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 19 23,743 5.5 8.9 0.61 (0.55–0.67) 29 (24–36)b
Celecoxib (any dose) v any active Any Any active comparator 25 27,334 5.2 8.0 0.63 (0.57–0.69) 35 (29–45)b
Nausea
Celecoxib v placebo Any Placebo 17 9,510 2.7 3.4 0.76 (0.60–0.97) 155 (71–840)b
Celecoxib v paracetamol Any Paracetamol 4,000 mg 2 1,056 2.9 1.8 1.6 (0.73–3.7)
Celecoxib v rofecoxib Any Rofecoxib 25 mg 4 1,579 1.8 2.8 0.62 (0.32–1.2)
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 17 22,072 2.7 3.3 0.87 (0.74–1.02)
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 18 31,168 3.8 5.6 0.80 (0.72–0.89) 56 (44–77)b
Celecoxib (any dose) v any active Any Any active comparator 23 33,667 3.7 5.3 0.81 (0.73–0.90) 63 (49–88)b
Vomiting
Celecoxib v placebo Any Placebo 15 9,030 1.1 0.7 1.4 (0.86–2.4)
Celecoxib v paracetamol Any Paracetamol 4,000 mg 2 1,056 0.7 1.0 0.73 (0.19–2.7)
Celecoxib v rofecoxib Any Rofecoxib 25 mg 3 769 1.0 0.8 1.3 (0.29-5.7)
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 16 21,825 0.8 1.4 0.64 (0.49–0.83) 173 (115–350)b
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 17 30,921 1.2 1.9 0.75 (0.62–0.90) 144 (100–250)b
Celecoxib (any dose) v any active Any Any active comparator 21 32,610 1.2 1.9 0.76 (0.64–0.91) 156 (110–280)b
aRelative risk: bold indicates statistically significant difference. bNNTp (number-needed-to-treat to prevent one event) is indicated by bold. cNNH (number-needed-to-treat to harm one patient). CI, confidence interval; NSAID, nonsteroidal anti-inflammatory drug.
Table 5b Gastrointestinal adverse events reported in studies of arthritis patients (part 2)
Number of Incidence of events (%)
Outcome/ comparisons Celecoxib daily dose Comparator and daily dose Trials Patients Celecoxib Comparator Relative riska (95% CI) NNTpb or NNHc (95% CI)
Abdominal pain
Celecoxib v placebo Any Placebo 19 9,919 3.6 2.9 1.2 (0.92–1.5)
Celecoxib v paracetamol Any Paracetamol 4,000 mg 2 1,056 0.9 2.0 0.45 (0.15–1.3)
Celecoxib v rofecoxib Any Rofecoxib 25 mg 5 2,671 2.7 4.2 0.64 (0.42–0.97)a 67 (35–920)b
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 19 22,615 5.3 7.8 0.75 (0.68–0.83)a 41 (32–57)b
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 20 31,711 6.6 10.0 0.76 (0.70–0.82)a 29 (25–36)b
Celecoxib (any dose) v any active Any Any active comparator 26 35,302 6.2 9.2 0.75 (0.70–0.81)a 33 (28–41)b
Dyspepsia
Celecoxib v placebo Any Placebo 19 9,919 6.9 4.8 1.30 (1.08–1.6)a 46 (32–84)c
Celecoxib v paracetamol Any Paracetamol 4,000 mg 2 1,056 2.9 2.2 1.34 (0.63–2.9)
Celecoxib v rofecoxib Any Rofecoxib 25 mg 5 2,671 4.4 4.9 0.89 (0.63–1.3)
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 19 22,615 5.7 7.3 0.79 (0.71–0.88)a 61 (43–100)b
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 20 31,711 8.1 10.7 0.84 (0.78–0.90)a 39 (31–52)b
Celecoxib (any dose) v any active Any Any active comparator 26 35,302 7.8 9.9 0.85 (0.79–0.91)a 48 (37–68)b
Diarrhoea
Celecoxib v placebo Any Placebo 17 9,510 5.1 3.5 1.45 (1.16–1.82)a 53 (37–97)c
Celecoxib v paracetamol Any Paracetamol 4,000 mg 2 1,056 2.2 4.6 0.48 (0.24–0.95)a 41 (22–450)b
Celecoxib v rofecoxib Any Rofecoxib 25 mg 5 2,671 4.1 4.4 0.93 (0.65–1.3)
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 17 22,071 4.3 4.9 0.96 (0.85–1.1)
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 18 31,167 5.8 6.9 0.96 (0.88–1.1)
Celecoxib (any dose) v any active Any Any active comparator 24 34,758 5.6 6.6 0.95 (0.87–1.03)
Clinical ulcers and bleeds
Celecoxib v placebo Any Placebo 16 9,321 0.03 0.05 3 events
Celecoxib v paracetamol Any Paracetamol 4,000 mg 2 1,056 0.0 0.0 0 events
Celecoxib v rofecoxib Any Rofecoxib 25 mg 3 897 0.3 0.0 1 event
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 17 22,075 0.2 0.6 0.35 (0.22–0.56)a 250 (170–450)b
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 18 31,171 0.4 0.9 0.61 (0.46–0.81)a 200 (140–320)b
Celecoxib (any dose) v any active Any Any active comparator 22 32,508 0.4 0.8 0.61 (0.46–0.81)a 210 (150–350)b
aRelative risk: bold indicates statistically significant difference. bNNTp (number-needed-to-treat to prevent one event) is indicated by bold. cNNH (number-needed-to-treat to harm one patient). CI, confidence interval; NSAID, nonsteroidal anti-inflammatory drug.
Table 6 Cardio-renal adverse events reported in studies of patients treated for arthritis
Number of Incidence of events (%)
Outcome/ comparisons Celecoxib daily dose Comparator and daily dose Trials Patients Celecoxib Comparator Relative riska (95% CI) NNTpb or NNHc (95% CI)
Myocardial infarction
Celecoxib v placebo Any Placebo 16 9,315 0.12 0.07 10 events
Celecoxib v paracetamol Any Paracetamol 4,000 mg 2 1,056 0.00 0.00 0 events
Celecoxib v rofecoxib Any Rofecoxib 25 mg 5 2,667 0.00 0.08 1 event
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 16 21,818 0.15 0.04 1.9 (0.87–4.1) 23 events
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 17 30,220 0.22 0.14 1.6 (0.93–2.6) 56 events
Celecoxib (any dose) v any active Any Any active comparator 23 34,174 0.19 0.13 1.4 (0.87–2.3) 57 events
Celecoxib (any dose) v any comparator Any Any comparator 30 38,499 0.18 0.12 1.4 (0.88–2.2) 59 events
Celecoxib (any dose) v noncoxib comparator Any Any noncoxib comparator 28 36,316 0.19 0.12 1.4 (0.88–2.2) 57 events
Cardiac failure
Celecoxib v placebo Any Placebo 16 9,834 0.06 0.03 5 events
Celecoxib v paracetamol Any Paracetamol 4,000 mg 2 1,056 0.00 0.00 0 events
Celecoxib v rofecoxib Any Rofecoxib 25 mg 5 2,671 0.15 0.60 10 events
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 15 21,859 0.06 0.15 0.54 (0.29–1.02) 21 events
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 17 30,917 0.11 0.20 0.70 (0.43–1.1) 45 events
Celecoxib (any dose) v any active Any Any active comparator 23 34,512 0.11 0.23 0.64 (0.41–1.0) 55 events
Raised creatinine (above 1.3 × upper limit of normal)
Celecoxib v placebo Any Placebo 5 2,776 1.3 0.7 1.65 (0.69–4.0)
Celecoxib v paracetamol Any Paracetamol 4,000 mg
Celecoxib v rofecoxib Any Rofecoxib 25 mg
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 9 15,319 0.3 0.5 0.78 (0.46–1.3)
Celecoxib (any dose) v NSAID Any NSAID to maximum daily
Celecoxib (any dose) v any active Any Any active comparator 10 15,657 0.3 0.5 0.79 (0.47–1.3)
Hypertension and aggravated hypertension
Celecoxib v placebo Any Placebo 16 9,321 1.0 0.6 1.4 (0.85–2.4)
Celecoxib v paracetamol Any Paracetamol 4,000 mg 2 1,056 0.2 0.6 4 events
Celecoxib v rofecoxib Any Rofecoxib 25 mg 5 2,671 3.5 4.6 0.75 (0.52–1.1)
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 16 22,518 1.3 1.4 0.92 (0.73–1.2)
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 17 30,921 1.6 1.6 1.1 (0.90–1.3)
Celecoxib (any dose) v any active Any Any active comparator 23 34,512 1.7 1.8 1.0 (0.86–1.2)
Oedema at any site
Celecoxib v placebo Any Placebo 16 9,321 2.6 1.4 1.9 (1.4–2.7) 79 (54–145)c
Celecoxib v paracetamol Any Paracetamol 4,000 mg 2 1,056 2.3 1.8 1.3 (0.56–3.0)
Celecoxib v rofecoxib Any Rofecoxib 25 mg 5 2,671 18.0 25.0 0.72 (0.62–0.83) 14 (10–25)b
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 16 21,825 2.4 2.5 0.98 (0.82–1.2)
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 17 30,921 2.9 3.5 0.92 (0.81–1.05)
Celecoxib (any dose) v any active Any Any active comparator 23 34,512 3.8 5.4 0.84 (0.76–0.92) 62 (48–87)b
aRelative risk: bold indicates statistically significant difference. bNNTp (number-needed-to-treat to prevent one event) is indicated by bold. cNNH (number-needed-to-treat to harm one patient; CI, confidence interval; NSAID, nonsteroidal anti-inflammatory drug.
Table 7 Analysis of changes to haematological parameters in patients treated for arthritis
Number of Incidence of events (%)
Outcome/ comparisons Celecoxib daily dose Comparator and daily dose Trials Patients Celecoxib Comparator Relative riska (95% CI) NNTpb or NNHc (95% CI)
Haemoglobin fall of 20 g/L or more
Celecoxib v placebo Any Placebo 5 3,577 0.8 0.5 1.5 (0.56–4.0)
Celecoxib v paracetamol Any Paracetamol 4,000 mg
Celecoxib v rofecoxib Any Rofecoxib 25 mg
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 10 15,746 1.1 2.2 0.71 (0.55–0.91)a 92 (66–150)b
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 10 16,180 1.1 2.2 0.72 (0.56–0.92)a 93 (67–150)b
Celecoxib (any dose) v any active Any Any active comparator 11 16,990 1.1 2.1 0.72 (0.56–0.92)a 100 (71–170)b
Haematocrit fall of 5% or more
Celecoxib v placebo Any Placebo 9 6,442 8.1 6.5 1.20 (0.98–1.5)
Celecoxib v paracetamol Any Paracetamol 4,000 mg
Celecoxib v rofecoxib Any Rofecoxib 25 mg 2 962 12.6 17.1 0.74 (0.54–1.01)
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 12 6,910 9.9 15.4 0.77 (0.68–0.88)a 18 (14–25)b
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 12 8,038 9.9 15.4 0.78 (0.69–0.89)a 18 (14–25)b
Celecoxib (any dose) v any active Any Any active comparator 14 8,970 10.1 15.6 0.78 (0.69–0.88)a 18 (14–25)b
aRelative risk: bold indicates statistically significant difference. bNNTp (number-needed-to-treat to prevent one event) is indicated by bold. cNNH (number-needed-to-treat to harm one patient). CI, confidence interval; NSAID, nonsteroidal anti-inflammatory drug.
Table 8 Endoscopically detected ulcers in patients treated for arthritis, with and without aspirin
Number of Incidence of events (%)
Outcome/ comparisons Celecoxib daily dose Comparator and daily dose Trials Patients Celecoxib Comparator Relative riska (95% CI) NNTpb or NNHc (95% CI)
Analysis irrespective of aspirin use
Celecoxib v placebo Any Placebo 2 1,737 3.9 2.2 1.8 (0.89–3.6)
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 6 4,135 4.6 16.3 0.30 (0.24–0.37)a 8.6 (7.4–10)b
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 6 4,565 4.5 16.3 0.29 (0.24–0.36)a 8.4 (7.3–10)b
Analysis without aspirin use
Celecoxib v placebo Any Placebo 2 1,537 3.3 1.9 1.8 (0.79–3.9)
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 5 3,053 4.5 17.6 0.28 (0.22–0.36)a 7.6 (6.5–9.1)b
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 5 3,440 4.2 17.6 0.28 (0.22–0.36)a 7.5 (6.4–8.9)b
Analysis with aspirin use
Celecoxib v placebo Any Placebo 2 200 7.9 4.1 1.7 (0.45–6.3)
Celecoxib (200/400) v NSAID 200–400 mg NSAID to maximum daily 5 344 10.0 23.8 0.47 (0.27–0.83)a 7.3 (4.6–17)b
Celecoxib (any dose) v NSAID Any NSAID to maximum daily 5 387 9.9 23.8 0.48 (0.28–0.83)a 7.2 (4.7–16)b
aRelative risk: bold indicates statistically significant difference. bNNTp (number-needed-to-treat to prevent one event) is indicated by bold. cNNH (number-needed-to-treat to harm one patient). CI, confidence interval; NSAID, nonsteroidal anti-inflammatory drug.
==== Refs
Fear J Hillman M Chamberlain MA Tennant A Prevalence of hip problems in the population aged 55 years and over: access to specialist care and future demand for hip arthroplasty Br J Rheumatol 1997 36 74 76 9117180 10.1093/rheumatology/36.1.74
Elliott AM Smith BH Penny KI Smith WC Chambers WA The epidemiology of chronic pain in the community Lancet 1999 354 1248 1252 10520633 10.1016/S0140-6736(99)03057-3
Crichton B Green M GP and patient perspectives on treatment with non-steroidal anti-inflammatory drugs for the pain in osteoarthritis Curr Med Res Opin 2002 18 92 96 12017216 10.1185/030079902125000345
Fraenkel L Bogardus ST Concato J Wittink DR Treatment options in knee osteoarthritis. The patient's perspective Arch Intrn Med 2004 164 1299 1304 10.1001/archinte.164.12.1299
Scholes D Stergachis A Penna PM Normand EH Hansten PD Nonsteroidal anti-inflammatory drug discontinuation in patients with osteoarthritis J Rheumatol 1995 22 708 712 7791168
Moore RA Phillips CJ Cost of NSAID adverse effects to the NHS J Med Econ 1999 2 45 55
Tramer MR Moore RA Reynolds DJ McQuay HJ Quantitative estimation of rare adverse effects which follow a biological progression – a new model applied to chronic NSAID use Pain 2000 85 169 182 10692616 10.1016/S0304-3959(99)00267-5
Henry D Page J Whyte I Nanra R Hall C Consumption of non-steroidal anti-inflammatory drugs and the development of functional renal impairment in elderly subjects. Results of a case-control study Br J Clin Pharmacol 1997 44 85 90 9241101 10.1046/j.1365-2125.1997.00631.x
Griffin MR Yared A Ray WA Nonsteroidal antiinflamatory drugs and acute renal failure in elderly persons Am J Epidemiol 2000 151 488 496 10707917
Page J Henry D Consumption of NSAIDs and the development of congestive heart failure in elderly patients: an underrecognized public health problem Arch Intern Med 2000 160 777 784 10737277 10.1001/archinte.160.6.777
Garcia-Rodriguez A Hernandez-Diaz S Nonsteroidal antiinflammatory drugs as a trigger of clinical heart failure Epidemiology 2003 14 240 246 12606892 10.1097/00001648-200303000-00020
Jadad AR Moore RA Carroll D Jenkinson C Reynolds DJ Gavaghan DJ McQuay HJ Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials 1996 17 1 12 8721797 10.1016/0197-2456(95)00134-4
Smith LA Oldman AD McQuay HJ Moore RA Teasing apart quality and validity in systematic reviews: an example from acupuncture trials in chronic neck and back pain Pain 2000 86 119 132 10779669 10.1016/S0304-3959(00)00234-7
Moher D Cook DJ Eastwood S Olkin I Rennie D Stroup DF Improving the quality of reports of meta-analyses of randomised controlled: the QUOROM statement Lancet 1999 354 1896 1900 10584742 10.1016/S0140-6736(99)04149-5
Moore RA Gavaghan D Tramer MR Collins SL McQuay HJ Size is everything – large amounts of information are needed to overcome random effects in estimating direction and magnitude of treatment effects Pain 1998 78 209 216 9870574 10.1016/S0304-3959(98)00140-7
White WB Faich G Borer JS Makuch RW Cardiovascular thrombotic events in arthritis trials of the cyclooxygenase-2 inhibitor celecoxib Am J Cardiol 2003 92 411 418 12914871 10.1016/S0002-9149(03)00659-3
Gavaghan DJ Moore RA McQuay HJ An evaluation of homogeneity tests in meta-analyses in pain using simulations of individual patient data Pain 2000 85 415 424 10781914 10.1016/S0304-3959(99)00302-4
Sterne JA Gavaghan D Egger M Publication and related bias in meta-analysis: power of statistical tests and prevalence in the literature J Clin Epidemiol 2000 53 1119 1129 11106885 10.1016/S0895-4356(00)00242-0
Tang J-L Liu JLY Misleading funnel plot for detection of bias in meta-analysis J Clin Epidemiol 2000 53 477 484 10812319 10.1016/S0895-4356(99)00204-8
L'Abbe KA Detsky AS O'Rourke K Meta-analysis in clinical research Ann Intern Med 1987 107 224 233 3300460
Morris JA Gardner MJ Gardner MJ, Altman DG Calculating confidence intervals for relative risk, odds ratios and standardised ratios and rates Statistics with Confidence – Confidence Intervals and Statistical Guidelines 1995 London: British Medical Journal 50 63
Cook RJ Sackett DL The number needed to treat: a clinically useful measure of treatment effect BMJ 1995 310 452 454 7873954
Langman MJ Jensen DM Watson DJ Harper SE Zhao PL Quan H Bolognese JA Simon TJ Adverse upper gastrointestinal effects of rofecoxib compared with NSAIDs JAMA 1999 282 1929 1933 10580458 10.1001/jama.282.20.1929
Goldstein JL Silverstein FE Agrawal NM Hubbard RC Kaiser J Maurath CJ Verburg KM Geis GS Reduced risk of upper gastrointestinal ulcer complications with celecoxib, a novel COX-2 inhibitor Am J Gastroenterol 2000 95 1681 1690 10925968
Whelton A Maurath CJ Verburg KM Geis GS Renal safety and tolerability of celecoxib, a novel cyclooxygenase-2 inhibitor Am J Ther 2000 7 159 175 11317165
Reicin AS Shapiro D Sperling RS Barr E Yu Q Comparison of cardiovascular thrombotic events in patients with osteoarthritis treated with rofecoxib versus nonselective nonsteroidal anti-inflammatory drugs (ibuprofen, diclofenac, and nabumetone) Am J Cardiol 2002 89 204 209 11792343 10.1016/S0002-9149(01)02201-9
Garner S Fidan D Frankish R Judd M Towheed T Wells G Tugwell P Rofecoxib for rheumatoid arthritis (Cochrane Review) The Cochrane Library 2004 3 Chichester, UK: John Wiley & Sons
Garner S Fidan D Frankish R Judd M Shea B Towheed T Wells G Tugwell P Celecoxib for rheumatoid arthritis (Cochrane Review) The Cochrane Library 2004 3 Chichester, UK: John Wiley & Sons
Deeks JJ Smith LA Bradley MD Efficacy, tolerability, and upper gastrointestinal safety of celecoxib for treatment of osteoarthritis and rheumatoid arthritis: systematic review of randomised controlled trials BMJ 2002 325 1 8 12098707 10.1136/bmj.325.7365.619
Edwards JE McQuay HJ Moore RA Efficacy and safety of valdecoxib for treatment of osteoarthritis and rheumatoid arthritis: systematic review of randomised controlled trials Pain 2004 111 286 296 15363872 10.1016/j.pain.2004.07.004
Clarke M Halsey J DICE2: a further investigation of the effects of chance in life, death and subgroup analyses Int J Clin Pract 2001 55 240 242 11406908
Edwards IR Aronson JK Adverse drug reactions: definitions, diagnosis, and management Lancet 2000 356 1255 1259 11072960 10.1016/S0140-6736(00)02799-9
Edwards JE McQuay HJ Moore RA Collins SL Reporting of adverse effects in clinical trials should be improved: lessons from acute postoperative pain J Pain Symptom Manage 1999 18 427 437 10641469 10.1016/S0885-3924(99)00093-7
Miceli-Richard C Le Bars M Schmidely N Dougados M Paracetamol in osteoarthritis of the knee Ann Rheum Dis 2004 63 923 930 15249319 10.1136/ard.2003.017236
Goldstein J Bjarnason I Moore A Spalding W Cawkwell G Comparison of hemoglobin and hematocrit changes associated with valdecoxib and nonspecific NSAIDs: pooled analysis of randomized trials in patients with OA or RA Ann Rheum Dis 2004 372
Goldstein J Bjarnason I Spalding W Cawkwell G Moore A Long-term NSAIDs but not COX-2 specific inhibitors are associated with anemia [abstract] Gastroenterology 2004 A1 15125448
Laine L Maller ES Yu C Quan H Simon T Ulcer formation with low-dose enteric-coated aspirin and the effect of COX-2 selective inhibition: a double-blind trial Gastroenterology 2004 127 395 402 15300570 10.1053/j.gastro.2004.05.001
Silverstein FE Faich G Goldstein JL Simon LS Pincus T Whelton A Makuch R Eisen G Agrawal NM Stenson WF Gastrointestinal toxicity with celecoxib vs nonsteroidal anti-inflammatory drugs for osteoarthritis and rheumatoid arthritis: the CLASS study: a randomized controlled trial. Celecoxib Long-term Arthritis Safety Study JAMA 2000 284 1247 1255 10979111 10.1001/jama.284.10.1247
Bombardier C Laine L Reicin A Shapiro D Burgos-Vargas R Davis B Day R Ferraz MB Hawkey CJ Hochberg MC Comparison of upper gastrointestinal toxicity of rofecoxib and naproxen in patients with rheumatoid arthritis N Engl J Med 2000 343 1520 1528 11087881 10.1056/NEJM200011233432103
Mamdani M Rochon PA Juurlink DN Kopp A Anderson GM Naglie G Austin PC Laupacis A Observational study of upper gastrointestinal haemorrhage in elderly patients given selective cyclo-oxygenase-2 inhibitors or conventional non-steroidal anti-inflammatory drugs BMJ 2002 325 624 629 12242172 10.1136/bmj.325.7365.624
MacDonald TM Morant SV Goldstein JL Burke TA Pettitt D Channelling bias and the incidence of gastrointestinal haemorrhage in users of meloxicam, coxibs, and older, non-specific non-steroidal anti-inflammatory drugs Gut 2003 52 1265 1270 12912856 10.1136/gut.52.9.1265
Nørgård B Pedersen L Johnsen SP Tarone RE McLaughlin JK Friis S Sorensen HT COX-2-selective inhibitors and the risk of upper gastrointestinal bleeding in high-risk patients with previous gastrointestinal diseases: a population-based case-control study Aliment Pharmacol Ther 2004 19 817 825 15043523 10.1111/j.1365-2036.2004.01913.x
Arboleya LR de la Figuera E Soledad Garcia M Aragon B VICOXX Study Group Experience of rofecoxib in patients with osteoarthritis previously treated with traditional non-steroidal anti-inflammatory drugs in Spain: results of phase 2 of the VICOXX study Curr Med Res Opin 2003 19 288 297 12841921 10.1185/030079903125001910
Hawkey C Laine L Simon T Beaulieu A Maldonado-Cocco J Acevedo E Shahane A Quan H Bolognese J Mortensen E Comparison of the effect of rofecoxib (a cyclooxygenase 2 inhibitor), ibuprofen, and placebo on the gastroduodenal mucosa of patients with osteoarthritis Arthritis Rheum 2000 43 370 377 10693877 10.1002/1529-0131(200002)43:2<370::AID-ANR17>3.0.CO;2-D
Hawkey CJ Laine L Simon T Quan H Shingo S Evans J Rofecoxib Rheumatoid Arthritis Endoscopy Study Group Incidence of gastroduodenal ulcers in patients with rheumatoid arthritis after 12 weeks of rofecoxib, naproxen, or placebo: a multicentre, randomised, double blind study Gut 2003 52 820 826 12740337 10.1136/gut.52.6.820
Harley C Wagner S Persistence with COX-2 inhibitors in managed care: an analysis of claims data Manag Care Interface 2003 16 38 45 14606259
Laine L Wogen J Yu H Gastrointestinal health care resource utilization with chronic use of COX-2-specific inhibitors versus traditional NSAIDs Gastroenterology 2003 125 389 395 12891540 10.1016/S0016-5085(03)00900-4
Gøtzsche PC Reporting of outcomes in arthritis trials measured on ordinal and interval scales is inadequate in relation to meta-analysis Ann Rheum Dis 2001 60 349 352 11247864 10.1136/ard.59.6.407
Barden J Meta-analysis to investigate methodological issues in clinical trials using pain as an exemplar DPhil Thesis 2004 Oxford University, Nuffield Department of Anesthetics
Loke YK Derry S Reporting of adverse drug reactions in randomised controlled trials – a systematic survey BMC Clin Pharmacol 2001 1 3 11591227 10.1186/1472-6904-1-3
| 15899051 | PMC1174947 | CC BY | 2021-01-04 16:02:36 | no | Arthritis Res Ther. 2005 Mar 24; 7(3):R644-R665 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1704 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17051589903910.1186/ar1705Research ArticleThe mechanism of low-concentration sodium nitroprusside-mediated protection of chondrocyte death Kim Hyun A [email protected] Ki Byoung [email protected] Sang-cheol [email protected] Department of Internal Medicine, Hallym University Sacred Heart Hospital, Kyunggi-do, Korea2 Department of Orthopedic Surgery, Hallym University Sacred Heart Hospital, Kyunggi-do, Korea3 Department of Internal Medicine, Hanyang University College of Medicine, Seoul, Korea2005 1 3 2005 7 3 R526 R535 26 11 2004 22 12 2004 22 1 2005 1 2 2005 Copyright © 2005 Kim et al.; licensee BioMed Central LtdThis is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Sodium nitroprusside (SNP), a widely used nitric oxide donor, has recently been shown to mediate chondrocyte apoptosis by generating reactive oxygen species, whereas more potent nitric oxide donors do not induce chondrocyte apoptosis. The present study was performed to investigate the protective effect of a low concentration of SNP upon the cytotoxicity of chondrocytes to higher concentrations of SNP, and to elucidate the underlying mechanism. Human osteoarthritis chondrocytes were cultured as monolayers, and first-passage cells were used for the experiments. Chondrocyte death induced by 1 mM SNP was completely inhibited by pretreating with 0.1 mM SNP. This protective effect of SNP was replicated by the guanosine-3',5'κ-cyclic monophosphate analog, DBcGMP. Protection from chondrocyte death conferred by 0.1 mM SNP was mediated by heme oxygenase 1 (HO-1), as was revealed by the increased expression of HO-1 in 0.1 mM SNP pretreated chondrocytes and by the reversal of this protective effect by the HO-1 inhibitor, zinc protoporphyrin. SNP-mediated chondrocyte protection correlated with the downregulation of both extracellular signal-regulated protein kinase 1/2 and p38 kinase activation. SNP at 0.1 mM induced significant NF-κB activation as revealed by electrophoretic mobility shift assays, and the inhibition of NF-κB by MG132 or Bay 11-7082 nullified 0.1 mM SNP-mediated chondrocyte protection. The upregulation of p53 and the downregulation of Bcl-XL and Mcl-1 by 1 mM SNP were reversed by 0.1 mM SNP pretreatment at the protein level by western blotting. Our study shows that priming with 0.1 mM SNP confers complete protection against cell death induced by 1 mM SNP in human articular chondrocytes. This protective effect was found to be correlated with the upregulation of both HO-1 and NF-κB and with the concomitant downregulation of both extracellular signal-regulated protein kinase 1/2 and p38 activation.
==== Body
Introduction
Articular cartilage consists of chondrocytes, the only cell type present, which are responsible for repairing tissue damage. Chondrocyte death and the pertinent signaling pathway involved have therefore been the focus of interest recently as pathogenetic factors leading to joint cartilage degradation in various forms of arthritides [1,2]. Several stimuli involved in the pathophysiology of arthritis, including nitric oxide (NO), Fas receptor ligation, and ceramide, have been reported to induce chondrocyte death in vitro [3-5].
The pathogenetic involvement of NO in arthritis was first demonstrated when levels of nitrite, a stable end product of NO metabolism, were shown to be elevated in serum and synovial fluid samples of rheumatoid arthritis patients and osteoarthritis patients [6]. Moreover, because osteoarthritic cartilage produces large amounts of NO, it could serve as a powerful initiator of chondrocyte death. In addition to the negative effects of NO on cartilage matrix synthesis (i.e. the inhibition of cartilage matrix macromolecule neosynthesis), the enhancement of matrix metalloproteinase activity, and the reduction of IL-1 receptor antagonist synthesis, NO may be an important mediator of cartilage degradation. However, the precise role of NO in the induction of chondrocyte death is debatable. For example, treatment with NO donors consistently induces cell death in cultured chondrocytes [3,7], whereas the production of high levels of endogenous NO by the overexpression of inducible NO synthase in transfected chondrocytes was not found to cause cell death [8]. This discrepancy may be attributed to the use of chemical NO donors, which not only generate reactive nitrogen species but also produce various secondary reactions depending on the cellular milieu in vitro. A recent study that employed diazeniumdiolates, which have been shown to be reliable sources of NO, demonstrated that exogenous NO is not cytotoxic to cultured chondrocytes per se, and that NO can even be protective under certain conditions of oxidative stress [9]. In addition, nitrite was found to exert a protective effect upon hypochlorous acid-induced chondrocyte toxicity, thus suggesting that NO has a novel cytoprotective role in inflamed joints [10].
This paradoxical effect of NO on cytotoxicity indicates that previous results using sodium nitroprusside (SNP) or S-nitroso-N-acetyl-L-penicillamine (SNAP) as NO donors should be cautiously interpreted. It has recently been reported that a low concentration of SNP exerts a protective effect against the cytotoxicity induced by higher concentrations of SNP, or against glucose deprivation in hepatocytes [11,12]. The objective of this study was to investigate the influence of low SNP concentrations upon the cytotoxicity induced by higher concentrations of SNP in chondrocytes. We also explored the mechanism of this low-concentration SNP-mediated cytoprotection.
Materials and methods
Reagents
Nitrate/nitrite colorimetric assay kits were purchased from Cayman Chemical (Ann Arbor, MI, USA). Ly83583 was purchased from Calbiochem (San Diego, CA, USA), 1-hydroxy-2-oxo-3-(3-aminopropyl)-3-isopropyl-1-triazene (NOC-5), SNAP, SB202190, PD98059, MG132 and Bay 11-7082 were from Alexis (Carlsbad, CA, USA), and zinc protoporphyrin (ZnPP) and cobalt protoporphyrin (CoPP) were from Frontier Scientific (Logan, UT, USA). Anti-phospho extracellular signal-regulated protein kinase (ERK) 1/2, anti-phospho-p38, anti-ERK 1/2, and anti-p38 were purchased from New England Biolabs (Beverly, MA, USA), anti-Bcl-2 from Transduction Laboratories (Lexington KY, USA), anti-Bax from Pharmingen (San Diego, CA, USA), and anti-heme oxygenase 1 (anti-HO-1), anti-Bcl-XL, anti-Mcl-1, anti-p53, anti-CIAP1 and anti-CIAP2 from Santa Cruz (Santa Cruz, CA, USA). All other reagents were obtained from Sigma (St Louis, MO, USA) unless specified otherwise.
Chondrocyte culture
Cartilage samples were obtained from the femoral condyle and from the tibial plateau of the knees of osteoarthritis patients at the time of joint replacement surgery. All cartilage samples were procured after obtaining oral informed consent from the patients and institutional approval. Pieces of articular cartilage were cut, minced, and incubated sequentially with pronase and collagenase in DMEM until they had been digested. Released cells were seeded at 4 × 106/plate in 10 cm culture plates in DMEM supplemented with 10% FCS, 1% L-glutamine, and 1% Fungizone (Gibco, Grand Island, NY, USA) and in DMEM supplemented with penicillin/streptomycin (150 units/ml and 50 mg/ml, respectively). Confluent chondrocytes were split once after about 7 days and were seeded at high density, and these first-passage adherent chondrocytes were then used in subsequent experiments.
Nitrate/nitrite quantification
NOC-5 was dissolved in 10 mM NaOH to produce a 200 mM stock solution and was stored at -20°C. SNP, 3-morpholinosydnonimine (SIN-1), and SNAP were freshly dissolved in water before each experiment. All NO donor compounds were diluted with DMEM and added directly to cultured chondrocytes. The final products of NO in vivo are nitrite and nitrate, the sum of which can be used as an index of total NO production. Chondrocyte culture media were harvested after being incubated for 24 hours with the respective NO donors, and were then analyzed using a nitrate/nitrite colorimetric assay kit as recommended by the manufacturer. Briefly, nitrate was converted to nitrite using nitrate reductase, and then Griess reagents were added to form a deep-purple azo compound. Absorbance was measured at 540 nm using a plate reader to determine the nitrite concentrations. The detection limit of the assay was 1 μM.
Quantification and verification of cell death
Cell death was quantitated using the 3-(4,5-dimethylthiazol-2-yl)-2,5 diphenyltetrazdium bromide (MTT) assay, as previously described [13]. Briefly, chondrocytes were seeded at 4 × 104/100 μl/well in 96-well microtiter plates. Cell death was induced by treating with 1 mM SNP. To protect cells, chondrocytes were treated with 0.1 mM SNP, 50 μM CoPP, or 1 mM dibutylyl guanosine-3',5'-cyclic monophosphate (DBcGMP) 14 hours prior to being treated with 1 mM SNP. MTT was then added to each well to a final concentration of 0.125 mg/ml after they had been incubated with 1 mM SNP for 24 hours, and plates were incubated at 37°C for a further 3 hours. The formazan product obtained was solubilized with 100 μl dimethylsulfoxide and optical densities were read at 595 nm. Percentage cell survival was calculated by taking the optical density of cells post-treatment, dividing this by the optical density of the untreated control cells, and multiplying by 100. Cell death was also verified by flow cytometry. Chondrocytes were trypsinized after treatment and were sedimented, and the cell pellets obtained were washed and stained with 100 μg/ml propidium iodide solution for 15 min. For each sample, 104 cells were analyzed by FACS II flow cytometry (Becton Dickinson, Mountain View, CA, USA).
Western blot
Cellular proteins were extracted in lysis buffer containing 50 mM sodium acetate, pH 5.8, 10% v/v SDS, 1 mM ethylene diaminetetraacetic acid, 1 mM phenylmethylsulfonyl fluoride, and 1 μg/ml aprotinin at 4°C. Samples were electrophoresed on 12% SDS-polyacrylamide gel, and transferred to polyvinylidene difluoride membranes. Blots were blocked with Tris-buffered saline containing 5% non-fat milk at room temperature for 1 hour, and then incubated with the respective antibodies overnight at 4°C. Finally, blots were incubated with 1:5000 peroxidase-conjugated goat anti-mouse or anti-rabbit IgG (Biorad, Hercules, CA, USA) for 1 hour. Bound immunoglobulin was detected by enhanced chemiluminescence (Amersham, Bucks, UK).
Electrophoretic mobility shift assay
Nuclear extracts from chondrocytes were prepared from 2 × 106 cells, as described previously with minor modification [14]. Briefly, cells were incubated on ice for 15 min with homogenization buffer containing 10 mM HEPES-KOH, 4 mM MgCl2, 10 mM KCl, 1 mM NaF, 0.5 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, and 20 μg/ml leupeptin. After adding detergent, the lysates were centrifuged at 3000g for 5 min. Pellets were resuspended in extraction buffer containing 20 mM HEPES-KOH, 1.5 mM MgCl2, 420 mM NaCl, 25% glycerol, 1 mM NaF, 0.5 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 20 μg/ml leupeptin, and 0.2 mM ethylene diaminetetraacetic acid. After incubation on ice and centrifugation, supernatants were collected, the protein content was measured, and 5 μg portions of extracts were used for the binding reaction. A consensus double-stranded NF-κB probe was obtained from Promega (Madison, WI, USA), and was end-labeled using γ-32P-adenosine-5-triphosphate. After incubating nuclear extracts in 2 μl gel binding buffer (Promega), end-labeled probe was added (100,000 cpm/sample). Samples were then incubated for 20 min and were loaded onto 4% nondenaturing polyacrylamide gels. Electrophoresis was run for 3 hours at 4°C. Protein complexes were identified by autoradiography.
Data analysis
Data are expressed as means ± standard deviations. The paired t test was used to compare controls and treatment conditions, and significance was accepted at a confidence level of 95% (P < 0.05).
Results
Chondrocyte death does not correlate with the amount of NO released by NO donors
A nitrate/nitrite assay kit was used to determine the amount of NO generated by the various NO donor compounds, SNP, NOC-5, SIN-1, and SNAP [15]. As was reported previously [9], the different NO donors released variable degrees of NO in the culture medium; SNP was the least efficient NO producer (Fig. 1a). A one millimolar concentration of SNP yielded about 12% of the NO produced by 1 mM diazeniumdiolate, NOC-5. However, 1 mM SNP led to almost complete chondrocyte death, whereas the same concentration of NOC-5 caused no appreciable cell death 24 hours after treatment (Fig. 1b). Other NO donors, SIN-1 and SNAP, also induced significant chondrocyte death at 2 mM concentrations. The amounts of NO produced by 2 mM SIN-1 or 2 mM SNAP were 10-fold and 8.9-fold higher than that produced by 2 mM SNP, respectively, but the levels of cell death induced were not as profound as that produced by 2 mM SNP. These results demonstrate that the amount of NO produced by a NO donor is not correlated with chondrocyte death.
0.1 mM SNP protects chondrocytes from death induced by 1 mM SNP
It was previously reported that pretreatment of hepatocytes with a low dose of SNP significantly inhibited high-dose SNP-induced hepatocyte death [11,12]. In order to determine whether this phenomenon also occurs in chondrocytes, we treated chondrocytes with a low, noncytotoxic concentration of SNP (i.e. 0.1 mM). As shown in Fig. 2a, priming the chondrocytes with 0.1 mM SNP for 14 hours completely inhibited the cell death induced by 1 mM SNP. However, pretreatment with concentrations higher than 0.2 mM SNP did not confer protection (data not shown). Pretreatment with 0.1 mM SNP for 1–6 hours was also protective (data not shown), but because the degree of protection was greatest for the 14-hour pretreatment, chondrocytes were pretreated with 0.1 mM SNP for 14 hours in all subsequent experiments.
Inhibition of cell death was also verified by fluorescence-activated cell sorting analysis of treated chondrocytes stained with propidium iodide (Fig. 2b). Because low concentrations of SNP are known to protect a murine macrophage cell line via the cGMP signaling pathway [16], we investigated whether cGMP is also protective in chondrocytes. Chondrocytes were thus pretreated with 1 mM DBcGMP, a cell-permeable cGMP analog, for 14 hours before administering 1 mM SNP. As is shown in Fig. 2a, pretreatment with DBcGMP led to the complete inhibition of 1 mM SNP-mediated chondrocyte death. In addition, pretreatment with LY83583, a soluble guanylate cyclase inhibitor, negated the protective effect of 0.1 mM SNP pretreatment, thus implicating the cGMP pathway in 0.1 mM SNP-mediated chondrocyte cytoprotection.
NOC-5 protects chondrocytes from 1 mM SNP-induced death
We also investigated whether 0.1 mM SNP-mediated cytoprotection is replicated by other NO donors. Low concentrations (0.1–0.5 mM) of SIN-1 or SNAP did not protect from the cell death induced by 1 mM SNP, but rather acted synergistically with SNP to enhance the cytotoxicity of subsequent 1 mM SNP treatment (data not shown). On the other hand, NOC-5 slightly inhibited 1 mM SNP-induced chondrocyte death, with maximal effect at 0.3 mM (Fig. 3).
The protection conferred by low concentration SNP is mediated by HO-1 upregulation
Because 0.1 mM SNP inhibited chondrocyte cytotoxicity induced by 1 mM SNP more so than the other NO donors examined, we investigated the mechanism of 0.1 mM SNP-mediated protection. Heme oxygenase is a rate-limiting enzyme in heme catabolism, and leads to the generation of bilirubin, free iron, and carbon monoxide (CO) [17-19]. HO-1 is inhibited by various metalloprotoporphyrins (e.g. ZnPP and tin protoporphyrin). Previous reports reveal that HO-1 overexpression is cytoprotective in multiple models including endotoxemia, shock, and ischemia/reperfusion [20-22]. Pretreatment of chondrocytes with the HO-1 inducer CoPP at 50 μM reproduced the cytoprotective effect of 0.1 mM SNP upon 1 mM SNP-induced cell death (Fig. 4a). Conversely, the co-treatment of chondrocytes with the HO-1 inhibitor ZnPP and 0.1 mM SNP attenuated the cytoprotective effect of 0.1 mM SNP. HO-1 was time-dependently induced by 0.1 mM SNP in chondrocytes, and western blot revealed that HO-1 is significantly upregulated after pretreatment with 0.1 mM SNP for 14 hours (Fig. 4b,c). CoPP at 50 μM upregulated HO-1 as was expected, and DBcGMP was also found to upregulate HO-1 in chondrocytes (Fig. 4c).
The protection conferred by 0.1 mM SNP is correlated with the downregulation of ERK 1/2 and p38 kinase activation
SNP at 1 mM caused the upregulation of both ERK 1/2 and p38 phosphorylation followed by chondrocyte death (Fig. 5a), and priming with 0.1 mM SNP reversed this pattern of mitogen-activated protein (MAP) kinase activation, by downregulating both ERK 1/2 and p38 phosphorylation. Pretreatment with the ERK 1/2 inhibitor PD98059 partially protected chondrocytes from death mediated by 1 mM SNP. The P38 kinase inhibitor SB202190 protected 1 mM SNP-mediated chondrocyte death only at 10 μM, which may inhibit pathways other than p38 (Fig. 5b). This result shows that the protection conferred by 0.1 mM SNP correlates with the downregulation of both ERK 1/2 and p38 kinase activation, but only the activation of ERK 1/2 was found to be directly responsible for chondrocyte death induced by 1 mM SNP.
The protection conferred by 0.1 mM SNP is negated by NF-κB suppression
Because NF-κB activation plays a pivotal role in protecting chondrocytes from apoptosis induced by death signals [23,24], the role of NF-κB activation in the protective effect of 0.1 mM SNP was examined. Activation of NF-κB by 0.1 mM SNP pretreatment was verified by electrophoretic mobility shift assay (Fig. 6a). Co-treatment with the NF-κB inhibitor Bay 11-7082 and with 0.1 mM SNP completely negated the protective effect of 0.1 mM SNP (Fig. 6a). Because Bay 11-7082 was found to be cytotoxic to chondrocytes (data not shown), another NF-κB inhibitor MG132, which is not cytotoxic to chondrocytes, was also tested. It was found that MG132 co-treatment also negated the protection conferred by 0.1 mM SNP. This result implies that NF-κB activation participates in the chondrocyte protection mediated by 0.1 mM SNP.
The protection conferred by 0.1 mM SNP correlates with the upregulation of Bcl-2 family proteins and the downregulation of p53
The Bcl-2 family proteins MCl-1 and Bcl-XL were both downregulated during the cell death induced by 1 mM SNP (Fig. 7). This downregulation was reversed by priming chondrocytes with 0.1 mM SNP. On the contrary, p53 was upregulated during 1 mM SNP-mediated chondrocyte death, but was downregulated by 0.1 mM SNP pretreatment (Fig. 7). The expressions of other Bcl-2 family members, such as Bcl-2 and Bax, or of the IAP family, c-IAP1, c-IAP2, or XIAP, were unaffected (data not shown).
Discussion
The mechanism of SNP-mediated chondrocyte death has been extensively investigated, and has usually been viewed as a NO-mediated form of chondrocyte apoptosis. In line with a previous result [9], our result shows that SNP is the least potent in terms of producing exogenous NO in chondrocyte culture, yet it is the most potent inducer of chondrocyte death. We cannot rule out the role played by NO in SNP-mediated chondrocyte death, because it is not possible to quench the NO produced by SNP treatment selectively. However, it is believed unlikely that NO is the sole mediator of SNP-induced chondrocyte death and peroxynitrite, a reaction product of NO and superoxide anions, or the primary byproducts of the decomposition of SNP, such as the cyanide anion or pentacyanoferrate complex, might contribute to its cytotoxicity [25,26].
Cytotoxic concentrations of SNP are associated with a 20-fold increase in NO production versus noncytotoxic concentrations, which contrasts with the actions of other nontoxic NO donors, which increase NO concentrations several hundred fold. It was of interest to find that pretreatment with 0.1 mM SNP led to complete chondrocyte protection against the toxic effect of 1 mM SNP. NOC-5, a diazeniumdiolate, also inhibited 1 mM SNP-induced chondrocyte death. However, despite the much higher level of NO formed by NOC-5, the degree of protection it conferred was smaller than that conferred by 0.1 mM SNP. It is thus also likely that the protection conferred by low-concentration SNP is not solely explained by NO production. It remains for further research to identify the other cytoprotective component mediated by low-concentration SNP.
In the present study, we used chondrocytes obtained from osteoarthritis patents at the advanced stage, because it was not possible to obtain sufficient chondrocytes from normal cartilage to carry out the in vitro experimentation. Although it is not possible to extrapolate our results to normal chondrocytes, a limited experiment utilizing normal cartilage obtained from a femoral head revealed that 0.1 mM SNP protected chondrocytes from cell death induced by 1 mM SNP to the same degree as was observed in osteoarthritis chondrocytes (data not shown).
To elucidate the signaling mechanism involved in low-concentration SNP-mediated cytoprotection, cGMP dependence was first examined. In the present study, the soluble guanylate cyclase inhibitor LY83583 was found to inhibit the cytoprotective effect of 0.1 mM SNP, whereas DBcGMP, a cell-permeable cGMP analog, attenuated the cell death induced by 1 mM SNP, indicating a cGMP-mediated cytoprotective mechanism. A previous study showed that the protection afforded by DBcGMP against SNP-induced death in RAW264 cells is mediated by protein kinase G activation, which results in the inhibition of cytochrome c release [16]. Because the inhibition of cytoprotection by Ly83583 was incomplete in the present study, and because other NO donors such as SIN-1 and SNAP, which also induce cGMP, failed to protect chondrocytes, other cytoprotective pathways were also examined.
Recent evidence has demonstrated the critical importance of HO-1 expression in the mediation of antioxidant, anti-inflammatory, and anti-apoptotic effects [19,27,28]. HO-1 is distributed ubiquitously and is induced strongly by a variety of physiologic and pathophysiologic stimuli, including heme, heavy metals, inflammatory cytokines, endotoxins, and NO [12]. The pretreatment of chondrocytes with the HO-1 inducer CoPP reproduced the cytoprotective effect of 0.1 mM SNP against 1 mM SNP-induced cell death, whereas the co-treatment of chondrocytes with the HO-1 inhibitor ZnPP and 0.1 mM SNP inhibited this cytoprotective effect. Moreover, HO-1 was found to be induced by 0.1 mM SNP treatment in chondrocytes.
The mechanism by which HO-1 protects from cell death has been postulated to involve several mechanisms, although the role of CO produced by the HO-1 degradation of heme has received most attention. Pharmacologic CO donors have also been demonstrated to protect hepatocytes from the death induced by glucose deprivation or anti-Fas [29]. Zuckerbraun and colleagues [30] recently showed that CO mediates hepatocyte protection by activating NF-κB, which in the presence of an inflammatory stimulus upregulates inducible NO synthase and leads to NO production. This mechanism implies a synergy between CO and NO in the provision of cytoprotection. Increased HO-1 activity also results in the generation of bilirubin, an antioxidant capable of scavenging peroxy radicals and inhibiting lipid peroxidation [29]. Finally, ferritin is another catalytic byproduct of HO-1 induction, and sequesters the free iron produced during heme catalysis, which reduces intracellular free iron and thus has an anti-oxidant effect [31]. The downstream process of cytoprotection conferred by the upregulation of HO-1 in human chondrocytes warrants further study. HO-1 was recently detected in human cartilage and in chondrocytes, and was found to be downregulated by proinflammatory cytokines and to be upregulated by anti-inflammatory cytokine, suggesting that HO-1 is a component of the protective mechanisms in human cartilage [32].
In a previous report, cell death and the dedifferentiation of chondrocytes was demonstrated to be regulated oppositely by two MAP kinase subtypes, ERK 1/2 and p38 kinase [33]. In rabbit chondrocytes, SNP increased both p38 kinase and ERK activation, and SNP-induced p38 kinase functioned as an induction signal for apoptosis and in the maintenance of the chondrocyte phenotype, whereas ERK activity caused dedifferentiation and operated as a prosurvival signal. Although our results show that high-dose SNP induces both p38 and ERK phosphorylation in line with the previous report [33], the downregulation of ERK 1/2 phosphorylation by low-concentration SNP was associated with chondrocyte protection rather than cell death in our human chondrocyte cultures.
The role played by ERK inhibition in chondrocyte death is not without controversy. Whereas one report showed that the blocking of MAP kinase kinase upstream of ERK by U0126 induces chondrocyte death, another report showed that ERK 1/2 or p38 kinase inhibition prevents SNP-induced chondrocyte death [34,35]. We found that the inhibition of ERK 1/2 leads to partial protection against 1 mM SNP-mediated chondrocyte death, but SB202190 at low concentrations, which specifically suppresses p38 activation, did not suppress it. This discrepancy probably stems from the differences in culture conditions, and concentrations of the inhibitors used. According to our result, although the protection conferred by 0.1 mM SNP correlates with the downregulation of both ERK 1/2 and p38 kinase activation, only the activation of ERK 1/2 is directly responsible for chondrocyte death induced by 1 mM SNP.
Finally, the role played by NF-κB activation in 0.1 mM SNP-mediated chondrocyte protection was investigated because NF-κB has been reported to serve as a survival signal in both tumor necrosis factor alpha and anti-Fas-mediated chondrocyte death [23,24]. NF-κB activation was observed after pretreating 0.1 mM SNP in human chondrocytes. Pretreating with the NF-κB inhibitors MG132 or Bay 11-7085 completely abolished the protection conferred by 0.1 mM SNP. Because this inhibition of the protective effect of SNP was greater than that conferred by either HO-1 or cyclic guanylase inhibitor, we believe that NF-κB has a pivotal role in the protective mechanism signaled by low-dose SNP in chondrocytes.
Of the regulators of cell survival, the expressions of p53, Bcl-XL, and Mcl-1 were significantly affected by 0.1 mM SNP pretreatment. The upregulation of p53 induced by 1 mM SNP was downregulated by 0.1 mM SNP pretreatment. Although we did not determine the mechanistic role of p53 phosphorylation, it is generally recognized that the phosphorylation of p53 leads to its accumulation, and that p53 is phosphorylated either indirectly or directly by c-Jun N terminal kinase, by p38 kinase, or by ERK [36-38]. We hypothesize that 1 mM SNP induced p38 kinase and ERK activity in chondrocytes and phosphorylated p53, resulting in p53 accumulation, and that this was negated by 0.1 mM SNP pretreatment via the downmodulation of these MAP kinases. Of the Bcl-2 family members, the downregulations of Bcl-XL and Mcl-1, both anti-apoptotic species, by 1 mM SNP was reversed by 0.1 mM SNP.
Despite the marked improvements made in our understanding of the mechanisms of chondrocyte apoptosis over the past several years, it is unclear whether chondrocyte apoptosis is the major mechanism of cartilage degradation or merely a byproduct of tissue degeneration. Thus, whether the modulation of apoptosis represents a feasible therapeutic target for the treatment of osteoarthritis is not obvious at the moment. A recent report showing that the intra-articular instillation of the pan-caspase inhibitor zVAD-fmk into the knees of rabbits induced to osteochondral injury led to a significant reduction in chondrocyte apoptosis implies that apoptosis inhibitors could be used to inhibit chondrocyte death in traumatic cartilage injury [39].
Conclusion
The present study shows that the widely used NO donor SNP at 1 mM concentration mediates chondrocyte death strongly despite its relatively poor ability to produce NO compared with other NO donors. Pretreating chondrocytes with SNP at 0.1 mM (a noncytotoxic concentration) protects the cells against 1 mM SNP cytotoxicity. This protective pathway was found to be related to four factors: cyclic GMP, HO-1, MAP kinase, and NF-κB. The study elucidates the survival pathway inherent in chondrocytes, and provides strategic information for the development of new therapeutics based on the regulation of chondrocyte death
Abbreviations
CO = carbon monoxide; CoPP = cobalt protoporphyrin; DBcGMP = dibutylyl guanosine-3',5'-cyclic monophosphate; DMEM = Dulbecco's modified Eagle's medium; ERK = extracellular signal-regulated protein kinase; FCS = fetal calf serum; HO-1 = heme oxygenase 1; IL = interleukin; MAP = mitogen-activated protein; MTT = 3-(4,5-dimethylthiazol-2-yl)-2,5 diphenyltetrazdium bromide; NF = nuclear factor; NO = nitric oxide; NOC-5 = 1-hydroxy-2-oxo-3-(3-aminopropyl)-3-isopropyl-1-triazene; SIN-1 = 3-morpholinosydnonimine; SNAP = S-nitroso-N-acetyl-L-penicillamine; SNP = sodium nitroprusside; ZnPP = zinc protoporphyrin.
Competing interests
The author(s) declare that there are no competing interests.
Authors' contributions
HAK conceived of the study, participated in its design, and supervised the experimental procedure. KBL provided samples, participated in the design of the study, and drafted the manuscript. S-cB performed the data analysis and drafted the manuscript.
Acknowledgements
This study was supported by a grant from the Korean Health 21 R & D Project, Korean Ministry of Health and Welfare (grant number 01-PJ3-PG6-01GN11-0002) and by the Korean Science and Engineering Foundation (grant number R04-2003-000-10006-0).
Figures and Tables
Figure 1 Quantification of nitrate/nitrite levels and the cell death induced by different nitric oxide (NO) donor compounds. Chondrocytes were obtained from the femoral condyle and the tibial plateau of knee osteoarthritis patients, cultured in monolayers, and seeded at 4 × 104/100 μl/well in 96-well microtiter plates. First-passage chondrocytes were used in the subsequent experiments. Chondrocytes were treated with the respective NO donor compounds for 24 hours. (a) Chondrocyte culture media were harvested and analyzed with a nitrate/nitrite colorimetric assay kit, as described in Materials and methods. (b) Cell death was quantitated using the 3-(4,5-dimethylthiazol-2-yl)-2,5 diphenyltetrazdium bromide (MTT) assay. MTT was added to each well after the collecting medium. Percentage cell survival was calculated by dividing the optical density of treated cells by the optical density of untreated control cells, and multiplying by 100. Cell survival in control culture was set at 100%. * P < 0.05 versus control. Data shown are the means and standard deviations of duplicate experiments on three different donors. NOC-5, 1-hydroxy-2-oxo-3-(3-aminopropyl)-3-isopropyl-1-triazene; SIN-1, 3-morpholinosydnonimine; SNAP, S-nitroso-N-acetyl-L-penicillamine.
Figure 2 Protective effect of low-concentration sodium nitroprusside (SNP) on human articular chondrocytes. Cell death was induced by treating chondrocytes with 1 mM SNP for 24 hours. To protect them from cell death, chondrocytes were treated with 0.1 mM SNP or 1 mM dibutylyl guanosine-3',5'-cyclic monophosphate (DBcGMP) 14 hours prior to 1 mM SNP treatment. To inhibit cyclic guanylase, 1 μM Ly83583 was added with 0.1 mM of SNP. (a) Cell death was quantitated using the 3-(4,5-dimethylthiazol-2-yl)-2,5 diphenyltetrazdium bromide assay. Cell survival in control cultures was set at 100%. Data shown are the means and standard deviations of triplicate experiments from at least three different donors. * P < 0.05 versus control. (b) Cell death was verified by propidium iodide staining and fluorescence-activated cell sorting analysis. Chondrocytes were trypsinized after treatment and were sedimented. Cell pellets obtained were washed and stained in 100 μg/ml propidium iodide solution for 15 min. For each sample, 104 cells were analyzed. Data are representative of samples from four different donors. Percentage values denote propidium iodide positive (dead) chondrocytes.
Figure 3 Effect of 1-hydroxy-2-oxo-3-(3-aminopropyl)-3-isopropyl-1-triazene (NOC-5) on the cytotoxic effect of 1 mM sodium nitroprusside (SNP) in human articular chondrocytes. Cell death was induced by treating chondrocytes with 1 mM SNP for 24 hours. To protect them from cell death, chondrocytes were treated with various concentrations of NOC-5 for 14 hours prior to 1 mM SNP treatment. Cell death was quantitated by the 3-(4,5-dimethylthiazol-2-yl)-2,5 diphenyltetrazdium bromide assay. Cell survival in control cultures was set at 100%. Data are the means and standard deviations of triplicate experiments on nine different donors. * P < 0.05 versus treatment with 1 mM SNP without NOC-5 pretreatment.
Figure 4 Protective effect of heme oxygenase 1 (HO-1) on human chondrocytes. (a) Cell death was induced by treating chondrocytes with 1 mM sodium nitroprusside (SNP) for 24 hours. For HO-1 induction, chondrocytes were treated with 50 μM cobalt protoporphyrin (CoPP) 14 hours prior to treating them with 1 mM SNP. For HO-1 inhibition, chondrocytes were treated with 1 μM zinc protoporphyrin (ZnPP) along with 0.1 mM SNP. Cell death was quantitated using the 3-(4,5-dimethylthiazol-2-yl)-2,5 diphenyltetrazdium bromide assay. Cell survival in control culture was set at 100%. Data shown are the means and standard deviations of triplicate experiments from at least four different donors. * P < 0.05 versus control. (b) Induction of HO-1 by 0.1 mM SNP treatment in human chondrocytes was analyzed by western blotting. Protein was extracted from chondrocytes after the indicated incubation periods and 20 μg each protein sample was separated by 12% SDS-PAGE and blotted with anti-HO-1 antibody. Data are representative of two samples from different donors. (c) Upregulation of HO-1 by pretreating chondrocytes with 0.1 mM SNP, 50 μM CoPP, or 1 mM dibutylyl guanosine-3',5'-cyclic monophosphate (DBcGMP). Chondrocytes were treated or not treated with the indicated chemicals for 14 hours and were then treated with 1 mM SNP for 2 hours. Protein was extracted from chondrocytes and 20 μg each protein sample was separated by 12% SDS-PAGE and blotted with anti-HO-1 antibody. The data shown are representative of five samples from different donors.
Figure 5 The regulation of extracellular signal-regulated protein kinase (ERK) 1/2 and p38 phosphorylation on 0.1 mM sodium nitroprusside (SNP) pretreated chondrocytes. (a) Chondrocytes were treated with 0.1 mM SNP 14 hours prior to treatment with 1 mM SNP and the phosphorylations of ERK 1/2 and p38 were analyzed after 4 hours by western blotting. Protein was extracted from chondrocytes and 20 μg each protein sample was separated by 12% SDS-PAGE and blotted with anti-phospho-ERK 1/2 or anti-phospho-p38. The expressions of ERK 1/2 and p38 were also determined by western blot analysis. Data are representative of samples from four different donors. (b) Effect of ERK 1/2 and p38 kinase inhibition on the chondrocyte death induced by 1 mM SNP. Cell death was induced by treating chondrocytes with 1 mM SNP for 24 hours. To inhibit ERK 1/2 or p38 kinase, chondrocytes were pretreated with 10, 20, or 50 μM PD98059 or with 1, 5, or 10 μM SB202190, respectively, for 2 hours before treating them with 1 mM SNP. Cell death was quantitated by the 3-(4,5-dimethylthiazol-2-yl)-2,5 diphenyltetrazdium bromide assay. Cell survival in control culture was set at 100%. Data are the means and standard deviations of duplicate experiments from at least three different donors. * P < 0.05 versus treatment with 1 mM SNP without pretreatment.
Figure 6 NF-κB activation was correlated with chondrocyte survival mediated by low-concentration sodium nitroprusside (SNP). (a) Activation of NF-κB in 0.1 mM SNP-treated chondrocytes. Chondrocytes were treated with 0.1 mM SNP for 14 hours with or without NF-κB inhibitors and the activation of NF-κB was analyzed by electrophoretic mobility shift assay. Nuclear extracts were prepared from 2 × 106 cells, and 5 μg portions of extracts were used for the binding reaction. Nuclear extracts were incubated in gel binding buffer with radiolabeled consensus double-stranded NF-κB probe, and samples were loaded onto 4% nondenaturing polyacrylamide gel. Protein complexes were identified by autoradiography. Data are representative of three samples from different donors. (b) Effect of the inhibition of NF-κB activation on the protective effect of 0.1 mM SNP on human chondrocytes. Cell death was induced by treating chondrocytes with 1 mM SNP for 24 hours. To inhibit NF-κB, chondrocytes were co-treated with 20 μM Bay 11-7082 or MG132 and 0.1 mM SNP for 14 hours before treating with 1 mM SNP. Cell death was quantitated by the 3-(4,5-dimethylthiazol-2-yl)-2,5 diphenyltetrazdium bromide assay. Cell survival in the control culture was set at 100%. Data are the means and standard deviations of triplicate experiments from at least three different donors. * P < 0.05 versus control.
Figure 7 Regulation of apoptosis-related proteins in sodium nitroprusside (SNP)-treated chondrocytes. Chondrocytes were treated with 0.1 mM SNP 14 hours prior to being treated with 1 mM SNP, and the expressions of Mcl-1, Bcl-XL, and p53 were analyzed after 4 hours by western blot. Protein was extracted from chondrocytes and 20 μg each protein sample was separated by 12% SDS-PAGE and blotted with respective antibodies. Data are representative of five samples from different donors.
==== Refs
Kim HA Song YW Apoptotic chondrocyte death in rheumatoid arthritis Arthritis Rheum 1999 42 1528 1537 10403282 10.1002/1529-0131(199907)42:7<1528::AID-ANR28>3.0.CO;2-9
Blanco FJ Guitian R Vazquez-Martul E de Toro FJ Galdo F Osteoarthritis chondrocytes die by apoptosis. A possible pathway for osteoarthritis pathology Arthritis Rheum 1998 41 284 289 9485086 10.1002/1529-0131(199802)41:2<284::AID-ART12>3.0.CO;2-T
Blanco FJ Ochs RL Schwarz H Lotz M Chondrocyte apoptosis induced by nitric oxide Am J Pathol 1995 146 75 85 7856740
Hashimoto S Setareh M Ochs RL Lotz M Fas/Fas ligand expression and induction of apoptosis in chondrocytes Arthritis Rheum 1997 40 1749 1755 9336406
Sabatini M Rolland G Leonce S Thomas M Lesur C Perez V de Nanteuil G Bonnet J Effects of ceramide on apoptosis, proteoglycan degradation, and matrix metalloproteinase expression in rabbit articular cartilage Biochem Biophys Res Commun 2000 267 438 444 10623638 10.1006/bbrc.1999.1983
Jang D Murrell GA Nitric oxide in arthritis Free Radic Biol Med 1998 24 1511 1519 9641270 10.1016/S0891-5849(97)00459-0
DelCarlo M Loeser RF Increased oxidative stress with aging reduces chondrocyte survival: correlation with intracellular glutathione levels Arthritis Rheum 2003 48 3419 3430 14673993 10.1002/art.11338
Studer RK Levicoff E Georgescu H Miller L Jaffurs D Evans CH Nitric oxide inhibits chondrocyte response to IGF-I: inhibition of IGF-IRbeta tyrosine phosphorylation Am J Physiol Cell Physiol 2000 279 C961 C969 11003576
Del Carlo M Loeser RF Nitric oxide-mediated chondrocyte cell death requires the generation of additional reactive oxygen species Arthritis Rheum 2002 46 394 403 11840442 10.1002/art.10056
Whiteman M Rose P Siau JL Halliwell B Nitrite-mediated protection against hypochlorous acid-induced chondrocyte toxicity: a novel cytoprotective role of nitric oxide in the inflamed joint? Arthritis Rheum 2003 48 3140 3150 14613276 10.1002/art.11284
Choi BM Pae HO Chung HT Nitric oxide protects nitric oxide-mediated apoptosis via heme oxygenase-1 induction Free Radic Biol Med 2003 34 1136 1145 12706494 10.1016/S0891-5849(03)00064-9
Choi BM Pae HO Kim YM Chung HT Nitric oxide-mediated cytoprotection of hepatocytes from glucose deprivation-induced cytotoxicity: involvement of heme oxygenase-1 Hepatology 2003 37 810 823 12668974 10.1053/jhep.2003.50114
Xu Y Bialik S Jones BE Iimuro Y Kitsis RN Srinivasan A Brenner DA Czaja MJ NF-kappaB inactivation converts a hepatocyte cell line TNF-alpha response from proliferation to apoptosis Am J Physiol 1998 275 C1058 C1066 9755059
Schreiber E Matthias P Muller M Schaffner W Rapid detection of octamer binding proteins with 'mini-extracts' prepared from a small number of cells Nucl Acid Res 1989 17 6419
Huie RE Padmaja S The reaction of NO with superoxide Free Radic Res Commun 1993 18 195 199 8396550
Yoshioka Y Yamamuro A Maeda S Nitric oxide at a low concentration protects murine macrophage RAW264 cells against nitric oxide-induced death via cGMP signaling pathway Br J Pharmacol 2003 139 28 34 12746220 10.1038/sj.bjp.0705206
Maines MD The heme oxygenase system: a regulator of second messenger gases Annu Rev Pharmacol Toxicol 1997 37 517 554 9131263 10.1146/annurev.pharmtox.37.1.517
Ponka P Cell biology of heme Am J Med Sci 1999 318 241 256 10522552 10.1097/00000441-199910000-00004
Otterbein LE Choi AM Heme oxygenase: colors of defense against cellular stress Am J Physiol Lung Cell Mol Physiol 2000 279 L1029 L1037 11076792
Otterbein L Sylvester SL Choi AM Hemoglobin provides protection against lethal endotoxemia in rats: the role of heme oxygenase-1 Am J Respir Cell Mol Biol 1995 13 595 601 7576696
Tamion F Richard V Lacoume Y Thuillez C Intestinal preconditioning prevents systemic inflammatory response in hemorrhagic shock. Role of HO-1 Am J Physiol Gastrointest Liver Physiol 2002 283 G408 G414 12121889
Amersi F Buelow R Kato H Ke B Coito AJ Shen XD Zhao D Zaky J Melinek J Lassman CR Upregulation of heme oxygenase-1 protects genetically fat Zucker rat livers from ischemia/reperfusion injury J Clin Invest 1999 104 1631 1639 10587527
Kuhn K Lotz M Regulation of CD95 (Fas/APO-1)-induced apoptosis in human chondrocytes Arthritis Rheum 2001 44 1644 1653 11465715 10.1002/1529-0131(200107)44:7<1644::AID-ART287>3.0.CO;2-S
Kim HA Song WY TNF-alpha-mediated apoptosis in chondrocytes sensitized by MG132 or actinomycin D Biochem Biophys Res Commun 2002 295 937 944 12127985 10.1016/S0006-291X(02)00789-1
Fitzhugh AL Keefer LK Diazeniumdiolates: pro- and antioxidant applications of the 'NONOates' Free Radic Biol Med 2000 28 1463 1469 10927170 10.1016/S0891-5849(00)00251-3
Bates JN Baker MT Guerra R JrHarrison DG Nitric oxide generation from nitroprusside by vascular tissue. Evidence that reduction of the nitroprusside anion and cyanide loss are required Biochem Pharmacol 1991 42 S157 S165 1768273 10.1016/0006-2952(91)90406-U
Brouard S Otterbein LE Anrather J Tobiasch E Bach FH Choi AM Soares MP Carbon monoxide generated by heme oxygenase 1 suppresses endothelial cell apoptosis J Exp Med 2000 192 1015 1026 11015442 10.1084/jem.192.7.1015
Poss KD Tonegawa S Reduced stress defense in heme oxygenase 1-deficient cells Proc Natl Acad Sci USA 1997 94 10925 10930 9380736 10.1073/pnas.94.20.10925
Llesuy SF Tomaro ML Heme oxygenase and oxidative stress. Evidence of involvement of bilirubin as physiological protector against oxidative damage Biochim Biophys Acta 1994 1223 9 14 8061058 10.1016/0167-4889(94)90067-1
Zuckerbraun BS Billiar TR Otterbein SL Kim PK Liu F Choi AM Bach FH Otterbein LE Carbon monoxide protects against liver failure through nitric oxide-induced heme oxygenase 1 J Exp Med 2003 198 1707 1716 14657222 10.1084/jem.20031003
Balla G Jacob HS Balla J Rosenberg M Nath K Apple F Vercellotti GM Ferritin: a cytoprotective antioxidant stratagem of endothelium J Biol Chem 1992 267 18148 18153 1517245
Fernandez P Guillen MI Gomar F Alcaraz MJ Expression of heme oxygenase-1 and regulation by cytokines in human osteoarthritic chondrocytes Biochem Pharmacol 2003 66 2049 2052 14599563 10.1016/S0006-2952(03)00543-4
Kim SJ Ju JW Oh CD Yoon YM Song WK Kim JH Yoo YJ Bang OS Kang SS Chun JS ERK-1/2 and p38 kinase oppositely regulate nitric oxide-induced apoptosis of chondrocytes in association with p53, caspase-3, and differentiation status J Biol Chem 2002 277 1332 1339 11689560 10.1074/jbc.M107231200
Shakibaei M Schulze-Tanzil G de Souza P John T Rahmanzadeh M Rahmanzadeh R Merker HJ Inhibition of mitogen-activated protein kinase kinase induces apoptosis of human chondrocytes J Biol Chem 2001 276 13289 13294 11278768 10.1074/jbc.M010859200
Notoya K Jovanovic DV Reboul P Martel-Pelletier J Mineau F Pelletier JP The induction of cell death in human osteoarthritis chondrocytes by nitric oxide is related to the production of prostaglandin E2 via the induction of cyclooxygenase-2 J Immunol 2000 165 3402 3410 10975859
Fuchs SY Adler V Pincus MR Ronai Z MEKK1/JNK signaling stabilizes and activates p53 Proc Natl Acad Sci USA 1998 95 10541 10546 9724739 10.1073/pnas.95.18.10541
She QB Chen N Dong Z ERKs and p38 kinase phosphorylate p53 protein at serine 15 in response to UV radiation J Biol Chem 2000 275 20444 20449 10781582 10.1074/jbc.M001020200
Sayed M Kim SO Salh BS Issinger OG Pelech SL Stress-induced activation of protein kinase CK2 by direct interaction with p38 mitogen-activated protein kinase J Biol Chem 2000 275 16569 16573 10747897 10.1074/jbc.M000312200
Costouros JG Dang AC Kim HT Inhibition of chondrocyte apoptosis in vivo following acute osteochondral injury Osteoarthritis Cartilage 2003 11 756 759 13129695 10.1016/S1063-4584(03)00157-2
| 15899039 | PMC1174948 | CC BY | 2021-01-04 16:02:37 | no | Arthritis Res Ther. 2005 Mar 1; 7(3):R526-R535 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1705 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17061589904010.1186/ar1706Research ArticleSynovial microparticles from arthritic patients modulate chemokine and cytokine release by synoviocytes Berckmans René J [email protected] Rienk [email protected] Maarten C [email protected] Marianne CL [email protected] Desirée [email protected] Tom JM [email protected] Augueste [email protected] Paul P [email protected] Department of Clinical Chemistry, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands2 Department of Clinical Immunology and Rheumatology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands2005 1 3 2005 7 3 R536 R544 13 10 2004 1 11 2004 26 1 2005 2 2 2005 Copyright © 2005 Berckmans et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Synovial fluid from patients with various arthritides contains procoagulant, cell-derived microparticles. Here we studied whether synovial microparticles modulate the release of chemokines and cytokines by fibroblast-like synoviocytes (FLS). Microparticles, isolated from the synovial fluid of rheumatoid arthritis (RA) and arthritis control (AC) patients (n = 8 and n = 3, respectively), were identified and quantified by flow cytometry. Simultaneously, arthroscopically guided synovial biopsies were taken from the same knee joint as the synovial fluid. FLS were isolated, cultured, and incubated for 24 hours in the absence or presence of autologous microparticles. Subsequently, cell-free culture supernatants were collected and concentrations of monocyte chemoattractant protein-1 (MCP-1), IL-6, IL-8, granulocyte/macrophage colony-stimulating factor (GM-CSF), vascular endothelial growth factor (VEGF) and intracellular adhesion molecule-1 (ICAM-1) were determined. Results were consistent with previous observations: synovial fluid from all RA as well as AC patients contained microparticles of monocytic and granulocytic origin. Incubation with autologous microparticles increased the levels of MCP-1, IL-8 and RANTES in 6 of 11 cultures of FLS, and IL-6, ICAM-1 and VEGF in 10 cultures. Total numbers of microparticles were correlated with the IL-8 (r = 0.91, P < 0.0001) and MCP-1 concentrations (r = 0.81, P < 0.0001), as did the numbers of granulocyte-derived microparticles (r = 0.89, P < 0.0001 and r = 0.93, P < 0.0001, respectively). In contrast, GM-CSF levels were decreased. These results demonstrate that microparticles might modulate the release of chemokines and cytokines by FLS and might therefore have a function in synovial inflammation and angiogenesis.
==== Body
Introduction
Cell-derived microparticles, predominantly from platelets and erythrocytes, are present in human blood. The presence of such microparticles has been associated with the activation of coagulation [1-3]. We demonstrated recently that synovial fluid from the inflamed joints of rheumatoid arthritis (RA) and arthritis control (AC) patients also contains cell-derived microparticles. These microparticles originate from monocytes and granulocytes, and to a smaller extent from lymphocytes [4]. Synovial microparticles are strongly procoagulant via an initiation mechanism dependent on tissue factor and factor VII(a). We therefore proposed that such microparticles might contribute to the local formation of fibrin clots, the so-called rice bodies.
Fibroblast-like synoviocytes (FLS) have a key function in the development of sustained inflammation and angiogenesis in arthritic joints [5-8]. On activation in vitro by cytokines or bacterial lipopolysaccharides, FLS produce chemokines including monocyte chemoattractant protein-1 (MCP-1) [9,10], IL-8 [11-13] and RANTES [11,14], cytokines such as IL-6 [12,13] and granulocyte/macrophage colony-stimulating factor (GM-CSF) [13,15,16], and angiogenic factors such as vascular endothelial growth factor (VEGF) [17,18].
The presence of leukocyte-derived microparticles in blood has been associated with systemic inflammatory disorders, such as pre-eclampsia [19], sepsis with multiple organ failure [20], and meningococcal septic shock [21], and leukocyte-derived microparticles – but not platelet-derived microparticles – trigger the expression of IL-6 and MCP-1 by endothelial cells [22,23]. However, it is unknown whether leukocytic microparticles contribute to local inflammation. We therefore determined whether isolated synovial microparticles of arthritis patients trigger the release of (pro-) inflammatory and angiogenic mediators by cultured autologous FLS from inflamed joints of RA and AC patients.
Materials and methods
Patients
Paired synovial fluid, plasma and synovial tissue specimens were collected from eight RA and three undifferentiated AC patients. The diagnosis of AC patients stayed unchanged during 1 year of follow-up. The RA patients fulfilled the criteria of the 1987 Criteria of the American College of Rheumatology. The study was approved by the Medical Ethical Committee of the Academical Medical Center of the University of Amsterdam, and informed consent was obtained to participate in the present study. The demographic and clinical data are summarized in Table 1.
Reagents and assays
Anti-CD4 labeled with phycoerythrin (PE; CLB-T4/2 6D10, IgG1) and anti-CD66e-PE (CLB-gran/10 IH4Fc, IgG1) were obtained from the Central Laboratory of the Netherlands Red Cross Blood Transfusion Service (CLB; Amsterdam, The Netherlands), anti-glycophorin A-PE (JC159, IgG1) was from DakoCytomation (Glostrup, Denmark). Anti-CD8-PE (Leu™-2a, IgG1), anti-CD14-PE (MφP9, IgG2b), anti-CD20-PE (L27, IgG1), anti-CD61-PE (VI-PL2, IgG1) and IgG1-PE (X40) were from Becton Dickinson (BD, San Jose, CA, USA), and anti-IgG2b-PE (MCG2b) was from Immuno Quality Products (Groningen, The Netherlands). IL-6, IL-8 and intracellular adhesion molecule-1 (ICAM-1; Diaclone Research, Besançon, France) and MCP-1, RANTES, VEGF and GM-CSF (BioSource International, Camarillo, CA, USA) were determined by ELISA. IL-1β was obtained from Roche Diagnostics (Mannheim, Germany)
Collection of the synovial biopsy and culture of FLS
Synovial tissue was collected from an actively inflamed joint by small-needle arthroscopy under local anesthesia with a 2.5 mm biopsy forceps to sample from different areas throughout the knee joint [24]. Synovial tissue was placed in Dulbecco's modified Eagle's medium (Life Technologies, Paisley, Renfrewshire, UK) supplemented with 10% FCS, 50 μg/ml streptomycin, 50 IU/ml penicillin and 2 mM L-glutamine and subjected to tissue digestion within 2 hours, as described previously [25]. The cells were cultured at 37°C and 5% CO2. After the second passage, FLS were seeded into 24-well flat-bottomed plates (Costar, Acton, MA) and maintained for 24 hours in culture medium containing 1% FCS.
Collection of synovial fluid and blood samples
Immediately before the arthroscopy, we collected synovial fluid (4.5 ml) from the same joint and also venous blood (4.5 ml) into tubes containing 0.5 ml of 3.2% sodium citrate (BD). Immediately after collection, a further 0.5 ml of 3.2% sodium citrate was added to the synovial fluid to prevent clotting. Cells were removed from both blood and synovial fluid by centrifugation for 20 min at 1,550 g and 20°C. For all determinations, aliquots of cell-free plasma and synovial fluid were snap-frozen in liquid nitrogen for at least 15 min and stored at -80°C until use.
Microparticle isolation
For flow-cytometric analysis, cell-free synovial fluid aliquots (250 μl) were thawed on melting ice and centrifuged for 30 min at 17,570 g and 20°C to pellet the microparticles. Supernatant (225 μl) was removed and microparticles were resuspended in 225 μl PBS (154 mM NaCl, 1.4 mM phosphate, pH 7.4), containing 10.9 mM trisodium citrate. After centrifugation for 30 min, supernatant (225 μl) was again removed and microparticles were resuspended in 150 μl of PBS/citrate buffer. For the FLS experiments, microparticles were isolated from 1 ml of synovial fluid by centrifugation for 1 hour at 17,570 g and 20°C. Supernatant (975 μl) was removed and replaced by 975 μl of PBS containing trisodium citrate. Microparticles were resuspended and again pelleted by centrifugation for 1 hour at 17,570 g and 20°C. Again, 975 μl of supernatant was removed and microparticles were resuspended in the remaining 25 μl. This microparticle suspension was added to a final volume of 1 ml of culture medium in which FLS had been maintained for 24 hours. Where indicated, a higher concentration of microparticles was also tested for its ability to activate FLS when sufficient synovial fluid was available. These microparticles, isolated from 3 ml of synovial fluid, were also concentrated into 25 μl of PBS containing trisodium citrate. Microparticle suspensions were each added to FLS cultures from the same donor to mimic the situation in vivo as much as possible.
Incubation of FLS with microparticles
FLS were quiescent after incubation for 24 hours in medium containing 1% FCS. After 24 hours, this medium (1 ml) was replaced by culture medium containing 1% FCS without any other addition (1 ml; control), or by 975 μl of culture medium plus (1) 25 μl of IL-1β (125 pg/ml final concentration), (2) 25 μl of microparticle suspension or (3) 25 μl of microparticle-free synovial fluid that had been diluted 1:9 in PBS (that is, containing 2.5 μl of the original synovial fluid; this quantity was chosen arbitrarily to correct for both the onefold (unconcentrated) and threefold concentrated microparticle suspensions that, after washing of the microparticles, still contained about 0.7 and 2.1 μl of synovial fluid, respectively). Because individual FLS cultures showed a considerable variation in (mediator) response to the positive control, namely IL-1β, we expressed the response of each FLS culture to microparticles as a percentage of the IL-1β-induced response.
Flow-cytometric analysis
Microparticles were measured by flow cytometry with a method that differed slightly from that used previously [4]. In the present study, the microparticles were not washed by centrifugation after being labeled with antibodies because this resulted in a selective loss of microparticle populations. In brief, 5 μl of the microparticle suspension was added to a mixture of PBS (45 μl) containing 2.5 mM CaCl2 and 5 μl of PE-labeled mAb, and incubated for 15 min in the dark at ambient temperature (20 to 22°C). The following (final concentrations) of mAbs were used: anti-CD4-PE (0.5 μg/ml), anti-CD8-PE (0.25 μg/ml), anti-CD14-PE (0.25 μg/ml), anti-CD20-PE (0.5 μg/ml), anti-CD61-PE (0.5 μg/ml), anti-CD66e-PE (0.25 μg/ml) and anti-glycophorin A-PE (0.25 μg/ml). PE-labeled IgG1 and IgG2b (both at 0.5 μg/ml) were used as isotype-specific control antibodies. After incubation, 900 μl of PBS/CaCl2 was added. Samples were analyzed on a FACSCalibur (BD) and data were analyzed with CellQuest™ Pro software (version 4.02; BD). Both forward scatter and side scatter were set at logarithmic gain. Microparticles were identified by forward scatter, side scatter and binding of cell-specific mAb. The number of microparticles per liter of plasma or synovial fluid was estimated by using the number of events (N) of cell-specific mAb-binding microparticles after correction for control antibody binding: number/liter = N × (150/5) × (955/67) × (106/250). The lower detection limit of the particle count was previously established as 107 microparticles per liter. In this formula, 150 (μl) is the final volume of the washed microparticle suspension, 5 (μl) is the volume of this suspension that is used for each labeling, 955 (μl) is the total volume of the microparticle suspension after labeling before fluorescence-activated cell sorting analysis, 67 (μl) is the average volume of the labeled microparticle suspension that is analyzed by the flow cytometer in 1 min, 106 is the conversion from μl to liter, and 250 (μl) is the original volume of the plasma or synovial fluid sample used for microparticle isolation.
Statistical analysis
Data were analyzed with GraphPad Prism for Windows, release 3.02 (San Diego, CA, USA). Differences in the concentrations of chemokines, cytokines and VEGF between synovial fluid and plasma as well as in culture supernatants were analyzed with the Wilcoxon signed-rank test. Two-tailed significance levels were considered significant at P < 0.05. All data are presented as medians (range).
Results
Cellular origin of synovial microparticles
Previously, we found no differences between the numbers and cellular origin of microparticles in synovial fluid from RA and AC patients [4]. For all cell-specific antigens tested, the microparticle numbers of the three AC patients fell within the range of the RA patients, which is consistent with these earlier observations. The data in Table 2 therefore summarize the microparticle numbers for RA and AC patients together. Most microparticles originated from monocytes (CD14) and granulocytes (CD66e). Microparticles derived from platelets (CD61) and erythrocytes (glycophorin A) were below detection level (less than 107/l) in synovial fluid from all patients, except in one RA patient who had a low but detectable number (1.7 × 107/l) of platelet-derived microparticles. One other RA patient had a relatively high number of erythrocyte-derived microparticles (3.1 × 109/l). Microparticles from CD4+ cells were found in six RA patients and all AC patients. Microparticles from CD8+ T cells were present in the synovial fluid of five RA patients and one AC patient. Microparticles from B cells were found in two RA patients only.
Synovial microparticles stimulate FLS
FLS were quiescent after incubation for 24 hours in medium containing 1% FCS. The concentrations of all markers studied in the FLS culture supernatants are summarized in Table 3. In comparison with the control (unstimulated), IL-1β significantly increased the levels of all mediators tested, whereas the addition of microparticle-free synovial fluid affected especially the soluble ICAM-1 (sICAM-1) levels. This increase was due to its presence in the synovial fluid itself. Addition of microparticles to FLS significantly increased the levels of MCP-1 (P = 0.010), sICAM-1 (P = 0.010), IL-8 (P = 0.008), IL-6 (P = 0.042), VEGF (P = 0.001) and RANTES (P = 0.031). In contrast, the concentrations of GM-CSF decreased (P = 0.002).
In six patients (three RA and three AC patients), we also tested a threefold higher (final) concentration of synovial microparticles. In comparison with the 'onefold' concentration, levels of sICAM-1 (P = 0.031), IL-8 (P = 0.031) and IL-6 (P = 0.031) increased further and GM-CSF (P = 0.016) decreased further (Table 3). Levels of MCP-1 (P = 0.156), VEGF (P = 0.078) and RANTES (P = 0.062) also tended to increase further, but these differences did not reach statistical significance.
Because individual microparticle suspensions were tested in (autologous) FLS cultures and considerable differences were observed in the responsiveness of these individual cell cultures, the individual responses of FLS cultures are also shown (Fig. 1). The response is expressed as either an increase or a decrease relative to the control, namely the 24-hour incubation of FLS with the microparticle-free synovial fluid. Although variation between FLS cultures is apparent, the individual data substantiate the conclusions above as based on group analysis.
Concentrations of MCP-1, IL-6, IL-8, RANTES, sICAM-1, VEGF and GM-CSF in vivo
For comparison, the concentrations of the various mediators were also determined in both synovial fluid and plasma from RA and AC patients. Because only 2 values (of 36) of the AC patients fell outside the RA range, namely MCP-1 in synovial fluid and sICAM-1 in plasma from the same AC patient, all data are summarized in Table 4. In comparison with plasma, levels of MCP-1 (P = 0.008), IL-6 (P = 0.002), IL-8 (P = 0.002) and VEGF (P = 0.002) were elevated in synovial fluid, those of RANTES and ICAM-1 were decreased (P = 0.001 and P = 0.006, respectively), and GM-CSF concentrations were similar (P = 0.125). Figure 2 shows that both the total number of microparticles (Fig. 2a; r = 0.91; P < 0.0001) and the numbers of granulocyte-derived microparticles (Fig. 2b; r = 0.89, P < 0.0001) were correlated with the IL-8 concentrations, whereas the numbers of monocyte-derived microparticles were not (Fig. 2c; r = 0.04; P = 0.89). In addition, concentrations of MCP-1 were correlated with total numbers of microparticles (r = 0.81, P < 0.0001) and numbers of granulocyte-derived microparticles (r = 0.93, P < 0.0001), but again not with the numbers of monocyte-derived microparticles (r = 0.06; P = 0.82; data not shown). No other correlations were found between microparticle numbers and concentrations of mediators.
Discussion
The present study shows that synovial fluid microparticles trigger FLS to release chemokines, cytokines and other mediators of inflammation. The extent to which these changes are solely induced by microparticles remains to be shown. We cannot exclude from our present data the possibility that the activation of FLS is due in part to synergistic actions of the microparticles with one or more mediators released by FLS themselves under these conditions. Neither can we exclude the possibility that microparticles activate FLS in synergy with one or more mediators already present in the synovial fluid. Nevertheless, the release of IL-8 and MCP-1 was correlated directly to both the total number of microparticles and the number of granulocyte-derived microparticles. This suggests that microparticles might trigger FLS to release these mediators. Although no correlations were found between microparticle numbers and sICAM-1, IL-6, VEGF and RANTES, a threefold increased concentration of microparticles tended to induce a higher response.
On the basis of these data it is tempting to speculate that synovial fluid microparticles promote synovial inflammation and neoangiogenesis in arthritic joints. The FLS are localized in the intimal lining layer, which directly contacts the synovial fluid compartment. Thus, synovial fluid microparticles may interact directly with the FLS, thereby modulating the release of an array of proinflammatory cytokines and chemokines. This may lead to further cell activation, neoangiogenesis and cell recruitment, constituting a proinflammatory amplification loop. Consistent with this notion is the observation that the removal of synovial fluid by arthroscopic lavage has a positive therapeutic effect in RA [26]. In addition, it has previously been shown that intra-articular injection of corticosteroids is more effective after arthrocentesis [27]. This has been explained by the effects of removal of fluid containing various proinflammatory cytokines.
At present, we can only speculate how synovial microparticles trigger FLS to produce and/or release proinflammatory mediators. Synovial microparticles originate mainly from leukocytes [4]. In vitro, leukocytic microparticles trigger the release of IL-6 and MCP-1 from endothelial cells [22,23]. Microparticles can contain bioactive lipids such as oxidized phospholipids, arachidonic acid and lysophosphatidic acid [28,29]. In particular, both arachidonic acid and lysophosphatidic acid are present in microparticles previously exposed to secretory phospholipase A2 (sPLA2) [30]. Arachidonic acid is transferred directly from microparticles to endothelial cells, resulting in the production of IL-6 [29]. It is unknown whether lysophosphatidic acid, a multifunctional lipid mediator that induces cell proliferation, migration and survival, is also directly transferred [31]. Synovial microparticles have been exposed to high levels of sPLA2 in vivo and are therefore likely to contain elevated levels of bioactive lipids. Thus, we propose that synovial microparticles might directly transfer bioactive lipids to FLS, thereby modulating the production and/or release of proinflammatory mediators. For this transfer, a direct interaction between microparticles and the FLS is essential. Because microparticles expose an array of cell-type-specific adhesion receptors, a direct interaction is likely. Alternatively, we cannot exclude the possibility that synovial microparticles might also contain inflammatory cytokines, because monocyte-derived microparticles generated in vitro were recently demonstrated to contain IL-1β [32].
Finally, the present study again showed that elevated levels of microparticles from granulocytes, monocytes and lymphocytes are present in the synovial fluid of arthritic patients. At present it is unknown why such elevated numbers of microparticles occur under these conditions. Apoptotic cells expose phosphatidylserine. Macrophages expose phosphatidylserine receptors, which efficiently initiate the recognition and subsequent removal of apoptotic cells [33,34]. It is also likely that microparticles are removed from the circulation by means of such receptors. However, synovial microparticles bind less annexin V than microparticles from plasma [4]. This decreased binding is due either to a decreased exposure of phosphatidylserine or to the presence of high levels of sPLA2, which competes with annexin V for binding to phosphatidylserine [35,36]. The removal of microparticles by phagocytic cells might thus be impaired in inflamed joints, resulting in the prolonged presence of microparticles and therefore in the continued stimulation of the FLS.
Conclusion
The results of the present study suggest that microparticles modulate the release of chemokines and cytokines by FLS. However, their biological relevance, compared with or in synergy with other biological mediators in synovial fluid, remains to be determined. The beneficial effect of arthrocentesis and arthroscopic lavage in RA might be explained, at least in part, by the removal of synovial fluid microparticles.
Abbreviations
AC = arthritis control; ELISA = enzyme-linked immunosorbent assay; FCS = fetal calf serum; FLS = fibroblast-like synoviocytes; GM-CSF = granulocyte/macrophage colony-stimulating factor; IL = interleukin; mAb = monoclonal antibody; MCP = monocyte chemoattractant protein; PBS = phosphate-buffered saline; PE = phycoerythrin; RA = rheumatoid arthritis; sICAM-1 = soluble intracellular adhesion molecule 1; sPLA2 = secretory phospholipase A2; VEGF = vascular endothelial growth factor.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
RB wrote the manuscript, guided by RN and AS, with clinical input and final correction by PT. RB, RN and AS devised the experimental design. The selection of patients and collection of synovial biopsy and blood materials were performed by MK. All experiments were performed by RB and MS except the culture of synoviocytes, which was performed by DP and TS. Supervision was fulfilled by AS and PT, with daily supervision by RN. The manuscript was read and approved by all authors.
Figures and Tables
Figure 1 Responses of individual cultures of fibroblast-like synoviocytes from rheumatoid arthritis (RA; n = 8) and arthritis control (AC; n = 3) patients to their autologous synovial microparticles. All individual patient data for the markers studied are expressed as the concentration of the mediator in the presence of microparticles concentrated either onefold (black bars) or threefold (open bars) divided by the concentration of mediator in the presence of microparticle-free synovial fluid. ICAM-1, intracellular adhesion molecule-1; MCP-1, monocyte chemoattractant protein-1.
Figure 2 Correlation between microparticle numbers and IL-8 concentrations. Correlations are shown between IL-8 produced by FLS in response to total microparticles (a), granulocyte-derived microparticles (b) and monocyte-derived microparticles (c). Note that data obtained with FLS in response to onefold and threefold concentrated microparticle suspensions are included.
Table 1 Demographic and clinical data of the rheumatoid arthritis patients and arthritis controls
Parameter RA patients (n = 8) AC patients (n = 3)
Age (years) 58 (34–69) 56 (49–68)
Sex (no. of males/females) 4/4 3/0
Disease duration (months) 60 (4–360) 2 (1–12)
Rheumatoid factor 7 positive; 1 negative 1 positive; 2 negative
Tender joint count 9 (5–15) 1 (1–2)
Swollen joint count 11 (5–19) 2 (1–23)
ESR (mm/h) 46 (25–69) 38 (28–43)
Erosive disease 6 positive; 2 negative None
No. of DMARDs 4.5 (1–5) 0
Leukocytes in SF (109/l) 6.3 (4.5–7.0) 4.3 (4.2–4.5)
CRP (mg/l) 34 (8–97) 4 (<3–26)
Results are medians, with ranges in parentheses. AC, arthritis control; CRP, C-reactive protein in plasma; DMARDs, disease-modifying antirheumatic drugs; ESR, erythrocyte sedimentation rate; RA, rheumatoid arthritis; SF, synovial fluid.
Table 2 Microparticle numbers in synovial fluid from patients with arthritic joints
Origin mAb Synovial fluid
CD4+ cells CD4 191 (<10–711)
CD8+ cells CD8 <10 (<10–331)
Monocytic cells CD14 1,315 (57–13,326)
B-cells CD20 <10 (<10–104)
Platelets CD61 <10 (<10–17)
Erythrocytes Glycophorin A <10 (<10–3,104)
Granulocytes CD66e 2,380 (<10–20,864)
Results are medians, with ranges in parentheses. Data are the numbers (× 106/l) of marker-positive microparticles from all arthritic patients (n = 11).
Table 3 Effect of synovial microparticles on the release of inflammatory mediators by fibroblast-like synoviocytes from arthritic patients (n = 11)
Mediator Control P* MP-free synovial fluid MP (1×) Nx/Nt P† MP (3×) Nx/Nt P‡
Unstimulated IL-1β
MCP-1 (pg/ml) 456 4,754 0.001 469 488 6/11 0.010 900 4/6 0.156
(355–1,292) (2,492–6,081) (293–1,241) (338–1,481) higher (346–2,326) higher
sICAM-1 (ng/ml) 0.09 0.40 0.007 1.04 2.00 8/11 0.010 6.07 6/6 0.031
(0–0.3) (0–0.76) (0.34–1.84) (0.35–4.07) higher (0.91–11.75) higher
IL-8 (pg/ml) 0 8,642 0.001 26 301 5/11 0.008 790 6/6 0.031
(0–564) (2,954–18,330) (0–528) (0–707) higher (0–2,100) higher
IL-6 (pg/ml) 74 4949 0.001 110 136 7/11 0.042 436 6/6 0.031
(24–1,710) (1,870–22,797) (30–1,176) (34–1,937) higher (44–3,766) higher
VEGF (pg/ml) 48 79 0.014 34 74 10/11 0.001 111 6/6 0.078
(11–102) (7–141) (1–97) (28–138) higher (27–161) higher
GM-CSF (pg/ml) 32 53 0.004 31 22 10/11 0.002 18 6/6 0.016
(28–40) (40–72) (26–70) (14–43) lower (14–25) lower
RANTES (pg/ml) 0 138 0.001 0 0.2 4/11 0.031 4.2 5/6 0.062
(0–74) (46–277) (0–58) (0–86) higher (0–32) higher
Results are medians, with ranges in parentheses. Concentrations of mediators were determined in the culture supernatant of the fibroblast-like synoviocytes (FLS) by ELISA as described in the Materials and methods section. FLS were incubated for 24 hours with 1 ml of culture medium containing 1% FCS (negative control), 975 μl of culture medium supplemented with either (1) 25 μl of interleukin (IL)-1β (final concentration 125 pg/ml; positive control), (2) 25 μl (onefold (1×) or threefold (3×) concentrated) microparticles (MP), or (3) MP-free synovial fluid. P*, positive versus negative control; P†, MP (1×) versus MP-free synovial fluid; P‡, MP (3×) versus MP (1×). Nx/Nt, number of individual culture supernatants that contained elevated or decreased concentrations of mediators after incubation for 24 hours with isolated MP compared with MP-free synovial fluid, divided by the number of patients studied. GM-CSF, granulocyte/macrophage colony-stimulating factor; sICAM-1, soluble intracellular adhesion molecule-1; MCP-1, monocyte chemoattractant protein-1; VEGF, vascular endothelial growth factor.
Table 4 Concentrations of inflammatory mediators in synovial fluid and plasma from arthritic patients (n = 11)
Mediator Concentration P
Synovial fluid Plasma
MCP-1 (pg/ml) 134 (36–522) 34 (15–62) 0.008
sICAM-1 (ng/ml) 706 (226–1,085) 871 (657–1,691) 0.006
IL-8 (pg/ml) 614 (<50–24,630) <50 0.002
IL-6 (pg/ml) 13,897 (35–43,131) 11 (0–57) 0.002
VEGF (pg/ml) 1,604 (528–2,506) 23 (<5–69) 0.002
RANTES (pg/ml) 7 (<5–35) 3,986 (2,920–10,037) 0.001
GM-CSF (pg/ml) <2 (<2–39) <2 (<2–28) 0.125
Results are medians, with ranges in parentheses. Concentrations of all mediators were determined by ELISA as described in the Materials and methods section. GM-CSF, granulocyte/macrophage colony-stimulating factor; MCP-1, monocyte chemoattractant protein-1; sICAM-1, soluble intracellular adhesion molecule-1; VEGF, vascular endothelial growth factor.
==== Refs
Holme PA Solum NO Brosstad F Roger M Abdelnoor M Demonstration of platelet-derived microvesicles in blood from patients with activated coagulation and fibrinolysis using a filtration technique and western blotting Thromb Haemost 1994 72 666 671 7900071
Berckmans RJ Nieuwland R Boing AN Romijn FP Hack CE Sturk A Cell-derived microparticles circulate in healthy humans and support low grade thrombin generation Thromb Haemost 2001 85 639 646 11341498
Katopodis JN Kolodny L Jy W Horstman LL De Marchena EJ Tao JG Haynes DH Ahn YS Platelet microparticles and calcium homeostasis in acute coronary ischemias Am J Hematol 1997 54 95 101 9034282 10.1002/(SICI)1096-8652(199702)54:2<95::AID-AJH1>3.0.CO;2-Z
Berckmans RJ Nieuwland R Tak PP Boing AN Romijn FP Kraan MC Breedveld FC Hack CE Sturk A Cell-derived microparticles in synovial fluid from inflamed arthritic joints support coagulation exclusively via a factor VII-dependent mechanism Arthritis Rheum 2002 46 2857 2866 12428225 10.1002/art.10587
Smith RS Smith TJ Blieden TM Phipps RP Fibroblasts as sentinel cells. Synthesis of chemokines and regulation of inflammation Am J Pathol 1997 151 317 322 9250144
Paleolog EM Miotla JM Angiogenesis in arthritis: role in disease pathogenesis and as a potential therapeutic target Angiogenesis 1998 2 295 307 14517450 10.1023/A:1009229508096
Firestein GS Invasive fibroblast-like synoviocytes in rheumatoid arthritis. Passive responders or transformed aggressors? Arthritis Rheum 1996 39 1781 1790 8912499
Pap T Muller-Ladner U Gay RE Gay S Fibroblast biology. Role of synovial fibroblasts in the pathogenesis of rheumatoid arthritis Arthritis Res 2000 2 361 367 11094449 10.1186/ar113
Villiger PM Terkeltaub R Lotz M Production of monocyte chemoattractant protein-1 by inflamed synovial tissue and cultured synoviocytes J Immunol 1992 149 722 727 1624809
Akahoshi T Wada C Endo H Hirota K Hosaka S Takagishi K Kondo H Kashiwazaki S Matsushima K Expression of monocyte chemotactic and activating factor in rheumatoid arthritis. Regulation of its production in synovial cells by interleukin-1 and tumor necrosis factor Arthritis Rheum 1993 36 762 771 8507217
Rathanaswami P Hachicha M Sadick M Schall TJ McColl SR Expression of the cytokine RANTES in human rheumatoid synovial fibroblasts. Differential regulation of RANTES and interleukin-8 genes by inflammatory cytokines J Biol Chem 1993 268 5834 5839 7680648
Nanki T Nagasaka K Hayashida K Saita Y Miyasaka N Chemokines regulate IL-6 and IL-8 production by fibroblast-like synoviocytes from patients with rheumatoid arthritis J Immunol 2001 167 5381 5385 11673556
Georganas C Liu H Perlman H Hoffmann A Thimmapaya B Pope RM Regulation of IL-6 and IL-8 expression in rheumatoid arthritis synovial fibroblasts: the dominant role for NF-κB but not C/EBP β or c-Jun J Immunol 2000 165 7199 7206 11120852
Volin MV Shah MR Tokuhira M Haines GK Woods JM Koch AE RANTES expression and contribution to monocyte chemotaxis in arthritis Clin Immunol Immunopathol 1998 89 44 53 9756723 10.1006/clin.1998.4590
Hamilton JA Piccoli DS Cebon J Layton JE Rathanaswani P McColl SR Leizer T Cytokine regulation of colony-stimulating factor (CSF) production in cultured human synovial fibroblasts. II. Similarities and differences in the control of interleukin-1 induction of granulocyte-macrophage CSF and granulocyte-CSF production Blood 1992 79 1413 1419 1372187
Nawata Y Eugui EM Lee SW Allison AC IL-6 is the principal factor produced by synovia of patients with rheumatoid arthritis that induces B-lymphocytes to secrete immunoglobulins Ann NY Acad Sci 1989 557 230 238 2786697
Kasama T Shiozawa F Kobayashi K Yajima N Hanyuda M Takeuchi HT Mori Y Negishi M Ide H Adachi M Vascular endothelial growth factor expression by activated synovial leukocytes in rheumatoid arthritis: critical involvement of the interaction with synovial fibroblasts Arthritis Rheum 2001 44 2512 2524 11710707 10.1002/1529-0131(200111)44:11<2512::AID-ART431>3.0.CO;2-O
Jackson JR Minton JA Ho ML Wei N Winkler JD Expression of vascular endothelial growth factor in synovial fibroblasts is induced by hypoxia and interleukin 1beta J Rheumatol 1997 24 1253 1259 9228120
VanWijk MJ Nieuwland R Boer K van der Post JA VanBavel E Sturk A Microparticle subpopulations are increased in preeclampsia: possible involvement in vascular dysfunction? Am J Obstet Gynecol 2002 187 450 456 12193942 10.1067/mob.2002.124279
Joop K Berckmans RJ Nieuwland R Berkhout J Romijn FP Hack CE Sturk A Microparticles from patients with multiple organ dysfunction syndrome and sepsis support coagulation through multiple mechanisms Thromb Haemost 2001 85 810 820 11372673
Nieuwland R Berckmans RJ McGregor S Boing AN Romijn FP Westendorp RG Hack CE Sturk A Cellular origin and procoagulant properties of microparticles in meningococcal sepsis Blood 2000 95 930 935 10648405
Mesri M Altieri DC Endothelial cell activation by leukocyte microparticles J Immunol 1998 161 4382 4387 9780216
Mesri M Altieri DC Leukocyte microparticles stimulate endothelial cell cytokine release and tissue factor induction in aJNK1 signaling pathway J Biol Chem 1999 274 23111 23118 10438480 10.1074/jbc.274.33.23111
Youssef PP Kraan M Breedveld F Bresnihan B Cassidy N Cunnane G Emery P Fitzgerald O Kane D Lindblad S Quantitative microscopic analysis of inflammation in rheumatoid arthritis synovial membrane samples selected at arthroscopy compared with samples obtained blindly by needle biopsy Arthritis Rheum 1998 41 663 669 9550475 10.1002/1529-0131(199804)41:4<663::AID-ART13>3.0.CO;2-L
Aupperle KR Bennett BL Boyle DL Tak PP Manning AM Firestein GS NF-κB regulation by IκB kinase in primary fibroblast-like synoviocytes J Immunol 1999 163 427 433 10384145
van Oosterhout M Sont JK van Laar JM Superior effect of arthroscopic lavage compared with needle aspiration in the treatment of inflammatory arthritis of the knee Rheumatology (Oxford) 2003 42 102 107 12509621
Weitoft T Uddenfeldt P Importance of synovial fluid aspiration when injecting intra-articular corticosteroids Ann Rheum Dis 2000 59 233 235 10700435 10.1136/ard.59.3.233
Barry OP Pratico D Lawson JA FitzGerald GA Transcellular activation of platelets and endothelial cells by bioactive lipids in platelet microparticles J Clin Invest 1997 99 2118 2127 9151784
Barry OP Kazanietz MG Pratico D FitzGerald GA Arachidonic acid in platelet microparticles up-regulates cyclooxygenase-2-dependent prostaglandin formation via a protein kinase C/mitogen-activated protein kinase-dependent pathway J Biol Chem 1999 274 7545 7556 10066822 10.1074/jbc.274.11.7545
Fourcade O Simon MF Viode C Rugani N Leballe F Ragab A Fournie B Sarda L Chap H Secretory phospholipase A2 generates the novel lipid mediator lysophosphatidic acid in membrane microvesicles shed from activated cells Cell 1995 80 919 927 7697722 10.1016/0092-8674(95)90295-3
Moolenaar WH van Meeteren LA Giepmans BN The ins and outs of lysophosphatidic acid signaling Bioessays 2004 26 870 881 15273989 10.1002/bies.20081
MacKenzie A Wilson HL Kiss-Toth E Dower SK North RA Surprenant A Rapid secretion of interleukin-1β by microvesicle shedding Immunity 2001 15 825 835 11728343 10.1016/S1074-7613(01)00229-1
Pradhan D Krahling S Williamson P Schlegel RA Multiple systems for recognition of apoptotic lymphocytes by macrophages Mol Biol Cell 1997 8 767 778 9168465
Fadok VA Bratton DL Rose DM Pearson A Ezekewitz RA Henson PM A receptor for phosphatidylserine-specific clearance of apoptotic cells Nature 2000 405 85 90 10811223 10.1038/35011084
Hara S Kudo I Chang HW Matsuta K Miyamoto T Inoue K Purification and characterization of extracellular phospholipase A2 from human synovial fluid in rheumatoid arthritis J Biochem (Tokyo) 1989 105 395 399 2732214
Buckland AG Wilton DC Inhibition of secreted phospholipases A2 by annexin V. Competition for anionic phospholipid interfaces allows an assessment of the relative interfacial affinities of secreted phospholipases A2 Biochim Biophys Acta 1998 1391 367 376 9555096
| 15899040 | PMC1174949 | CC BY | 2021-01-04 16:02:37 | no | Arthritis Res Ther. 2005 Mar 1; 7(3):R536-R544 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1706 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17081589904510.1186/ar1708Research ArticleSignalling pathway involved in nitric oxide synthase type II activation in chondrocytes: synergistic effect of leptin with interleukin-1 Otero Miguel 1Lago Rocío 1Lago Francisca 2Reino Juan Jesús Gomez 34Gualillo Oreste [email protected] NEIRID (NeuroEndocrine Interactions in Rheumatology and Inflammatory Diseases) Laboratory, Santiago University Clinical Hospital, Research Laboratory 4, Santiago de Compostela, Spain2 Laboratory of Molecular and Cellular Cardiology, Santiago University Clinical Hospital, Research Laboratory 1, Santiago de Compostela, Spain3 Rheumatology Division, Santiago University Clinical Hospital, Santiago de Compostela, Spain4 Department of Medicine, School of Medicine, University of Santiago de Compostela, Santiago de Compostela, Spain2005 4 3 2005 7 3 R581 R591 11 8 2004 16 9 2004 14 1 2005 3 2 2005 Copyright © 2005 Otero et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The objective of the present study was to investigate the effect of leptin, alone or in combination with IL-1, on nitric oxide synthase (NOS) type II activity in vitro in human primary chondrocytes, in the mouse chondrogenic ATDC5 cell line, and in mature and hypertrophic ATDC5 differentiated chondrocytes. For completeness, we also investigated the signalling pathway of the putative synergism between leptin and IL-1. For this purpose, nitric oxide production was evaluated using the Griess colorimetric reaction in culture medium of cells stimulated over 48 hours with leptin (800 nmol/l) and IL-1 (0.025 ng/ml), alone or combined. Specific pharmacological inhibitors of NOS type II (aminoguanidine [1 mmol/l]), janus kinase (JAK)2 (tyrphostin AG490 and Tkip), phosphatidylinositol 3-kinase (PI3K; wortmannin [1, 2.5, 5 and 10 μmol/l] and LY294002 [1, 2.5, 5 and 10 μmol/l]), mitogen-activated protein kinase kinase (MEK)1 (PD098059 [1, 5, 10, 20 and 30 μmol/l]) and p38 kinase (SB203580 [1, 5, 10, 20 and 30 μmol/l]) were added 1 hour before stimulation. Nitric oxide synthase type II mRNA expression in ATDC5 chondrocytes was investigated by real-time PCR and NOS II protein expression was analyzed by western blot. Our results indicate that stimulation of chondrocytes with IL-1 results in dose-dependent nitric oxide production. In contrast, leptin alone was unable to induce nitric oxide production or expression of NOS type II mRNA or its protein. However, co-stimulation with leptin and IL-1 resulted in a net increase in nitric oxide concentration over IL-1 challenge that was eliminated by pretreatment with the NOS II specific inhibitor aminoguanidine. Pretreatment with tyrphostin AG490 and Tkip (a SOCS-1 mimetic peptide that inhibits JAK2) blocked nitric oxide production induced by leptin/IL-1. Finally, wortmannin, LY294002, PD098059 and SB203580 significantly decreased nitric oxide production. These findings were confirmed in mature and hypertrophic ATDC5 chondrocytes, and in human primary chondrocytes. This study indicates that leptin plays a proinflammatory role, in synergy with IL-1, by inducing NOS type II through a signalling pathway that involves JAK2, PI3K, MEK-1 and p38 kinase.
==== Body
Introduction
Chondrocytes are the predominant cells in mature cartilage that synthesize and maintain the integrity of cartilage-specific extracellular matrix. In rheumatoid arthritis and osteoarthritis the phenotype of chondrocytes changes, and apoptosis and extracellular matrix degradation occur [1-3]. These severe perturbations in cartilage homeostasis may be mediated in part by nitric oxide (NO). This gaseous mediator is induced by several proinflammatory cytokines, including IL-1.
Leptin, the OB gene product, is a 16 kDa hormone that is synthesized by adipocytes. Leptin regulates food intake and energy expenditure, but it also modulates neuroendrocrine function [4]. It is involved in immune modulation in that it influences the innate immune response by promoting activation of monocyte/macrophages, chemotaxis and activation of neutrophils, and activation of natural killer cells [5]. Furthermore, leptin influences adaptive immunity by increasing the expression of adhesion molecules by CD4+ T cells, and promoting proliferation and secretion of IL-2 by naïve CD4+ T cells [5-7]. Leptin has also been found to influence bone growth [8] and inflammation [9].
High leptin levels are associated with obesity, which is a risk factor for osteoarthritis [10-12]. Interestingly, in patients with osteoarthritis leptin is present in synovial fluid and is expressed by articular chondrocytes [13], and normal human chondrocytes express the functional Ob-Rb leptin receptor isoform [14]. It is unlikely that leptin alone acts on cartilage to trigger an inflammatory response; rather, it may associate with other proinflammatory cytokines to amplify inflammation and enhance damage to cartilage. We recently demonstrated a synergistic effect of leptin with IFN-γ on nitric oxide synthase (NOS) type II activity in cultured chondrocytes that was mediated by the janus kinase (JAK)2 [15]. In the present study we investigated whether leptin synergizes with IL-1, an abundant mediator of inflammation and cartilage destruction [16,17], to activate NOS type II in chondrocytes. To gain further insights into the mechanism of action of this putative synergism, we also analyzed the role played by several intracellular kinases by using specific pharmacological inhibitors.
Materials and methods
Reagents
Foetal bovine serum, tissue culture media, media supplements, mouse and human recombinant leptin, mouse recombinant IL-1, tyrphostin AG490, wortmannin, LY294002, PD098059 and SB203580 were purchased from Sigma (St Louis, MO, USA) unless otherwise specified. RT-PCR reagents were purchased from Invitrogen (Carlsbad, CA, USA) and Stratagene (La Jolla, CA, USA). Tkip (WLVFFVIFYFFR), a suppressor of cytokine signalling (SOCS)-1 mimetic peptide that inhibits JAK2 autophosphorylation, was generously provided by Dr Howard M Johnson (Institute of Food and Agricultural Science, Department of Microbiology and Cell Science, University of Florida, Gainesville, FL, USA).
Cell culture
The clonal chondrogenic cell line ATDC5 was chosen for these studies because it has been shown to be a useful in vitro model for examining the multistep differentiation of chondrocytes. Undifferentiated ATDC5 cells proliferate rapidly until they reach confluence, at which point they undergo growth arrest. When treated with insulin, transferrin and sodium selenite, confluent ATDC5 cells re-enter a proliferative phase and form cartilaginous matrix nodules (mature chondrocytes). As differentiation progresses, these cells undergo a late differentiation phase, becoming hypertrophic, calcifying chondrocytes that synthesize type X collagen and osteopontin – a marker of terminal chondrocyte differentiation [18]. ATDC5 cells were a kind gift from Dr Agamemnon E Grigoriadis (Department of Craniofacial Development, King's College, London Guy's Hospital, London, UK). Unless otherwise specified, cells were cultured in Dulbecco's modified Eagle's medium/Hams' F12 medium supplemented with 5% foetal bovine serum, 10 μg/ml human transferrin, 3 × 10-8 mol/l sodium selenite and antibiotics (50 U/ml penicillin and 50 μg/ml streptomycin).
In some experiments, conducted to demonstrate that leptin/IL-1 synergism does not appear to depend on the differentiation state of the chondrocytes, chondrogenic ATDC5 cells were differentiated into mature and hypertrophic chondrocytes, as described by Thomas and coworkers [19]. Briefly, cells were plated at an initial density of 2 × 104 cells/well in 24-well plates. Cells were cultured in the above-mentioned medium supplemented with 10 μg/ml of human recombinant insulin (Novo Nordisk A/S, Bagsvaerd, Denmark). Culture was continued for a further 15 or 21 days, with replacement of medium every other day. As expected, ATDC5 cultures treated with insulin underwent progressive differentiation from 0 to 21 days as compared with untreated cultures. This differentiation was qualitatively characterized by increased formation of cartilage nodules and enhanced staining with alcian blue dye, which is indicative of cartilage proteoglycan accumulation.
In other experiments (data not shown), the differentiation from days 0 to 21 was further evidenced by sequential increases in type II collagen, aggrecan and type X collagen mRNAs. The early and mature chondrocyte marker type II collagen was expressed in undifferentiated ATDC5 cells; the level began to increase at day 3, peaked at days 7–10 and gradually declined after day 15. The expression profile of aggrecan mimicked that of type II collagen but with a slight delay of a couple of days. The decline in expression of both chondrocyte markers coincided with the onset of late-stage chondrocyte differentiation. The expression of the hypertrophic chondrocyte marker type X collagen began at days 12 and 13. The expression patterns of these early and late chondrocyte markers were consistent with previous findings in ATDC5 cells regarding in vivo chondrocyte differentiation. We do not illustrate findings regarding the differentiation of ATDC5 cells because they are extensively reported in literature [19].
Cartilage harvest and human chondrocyte isolation
Human normal articular cartilage samples were obtained from knee joints of patients undergoing leg amputations from above the knee because of peripheral vascular disease. (Permission from the local ethical committee was granted.) None of the patients had a clinical history of arthritis or any other pathology affecting the cartilage, and the specimens appeared normal on morphological examination (no change in colour and no fibrillation). For chondrocyte isolation, aseptically dissected cartilage was subjected to sequential digestion with pronase (catalogue number 165921; Roche Molecular Biochemicals, Indianapolis, IN, USA) and collagenase P (catalogue number 1213873; Roche Molecular Biochemicals) at a final concentration of 1 mg/ml in Dulbecco's modified Eagle's medium/F12 plus 10% foetal calf serum and sterilized by filtration, in accordance with the manufacturer's instructions. In our hands, this procedure was superior to enzymatic isolation with collagenase alone in terms of chondrocyte yields and capacity for attachment. Cartilage specimens were finely diced in phosphate-buffered saline (PBS), and after removing PBS diced tissue was incubated for 30 min with pronase in a shaking water bath at 37°C. Pronase was subsequently removed from the digestion flask and the cartilage pieces were washed with PBS. After removal of PBS, digestion was continued with addition of collagenase P; this was done over 6–8 hours in a shaking water bath at 37°C. The resulting cell suspension was filtered through a 40 μm nylon cell strainer (BD Biosciences Europe, Erembodegem, Belgium) in order to remove debris. Cells were centrifuged and washed twice with PBS, counted and plated in 24-well tissue culture plates for chondrocyte culture. Cells were serially passaged to obtain a sufficient number of cells and used between the first and second passages.
Cell treatments and nitrite assay
ATDC5 cells and human primary chondrocytes, with a viability greater than 95% as evaluated using the trypan blue exclusion method, were cultured (as described above) in 24-well plates. After 12 hours of starvation in serum-free medium, cells were stimulated for 48 hours with leptin (800 nmol/l), alone or in combination with IL-1 (0.025 ng/ml). We wished to determine whether increased NO production was due to NOS type II activation and to the involvement of JAK2, phosphatidylinositol 3-kinase (PI3K), mitogen-activated protein kinase kinase (MEK)1 and p38 kinase. For this purpose, the following specific pharmacological inhibitors were added 1 hour before cytokine stimulation: aminoguanidine (1 mmol/l) for NOS type II; tyrphostin AG490 (5 and 10 μmol/l) and Tkip (20 and 50 μmol/l) for JAK2; wortmannin (1, 2.5, 5 and 10 μmol/l) and LY294002 (1, 2.5, 5 and 10 μmol/l) for PI3K; PD098059 (1, 5, 10, 20 and 30 μmol/l) for MEK-1; and SB203580 (1, 5, 10, 20 and 30 μmol/l) for p38 kinase. Cytokines and pharmacological inhibitor doses were selected on the basis of prior dose–response experiments (data not shown) or previously published literature [15].
Nitrite accumulation was measured in culture medium using the Griess reaction. Briefly, 100 μl cell culture medium was mixed with 100 μl Griess reagent (equal volumes of 1% [weight/vol] sulfanilamide in 5% [vol/vol] phosphoric acid and 0.1% [weight/vol] naphtylethylenediamine-HCl), incubated at room temperature for 10 min, and then the absorbance at 550 nm was measured using a microplate reader (Titertek-Multiscan, Labsystem, Helsinki, Finland). Fresh culture medium was used as blank in all of the experiments. The amount of nitrite in the samples (in micromolar units) was calculated from a sodium nitrite standard curve freshly prepared in culture medium.
RNA isolation and real-time RT-PCR
ATDC5 chondrogenic cells were seeded in P6 well plates to reach 85–90% confluence. After 8 hours of starvation in serum-free medium, cells were treated with leptin alone or in combination with IL-1. In order to test the involvement of JAK2, PI3K, MEK-1 and p38 kinase on NOS type II mRNA expression, specific inhibitors (tyrphostin AG490 10 μmol/l, wortmannin and LY294002 10 μmol/l, PD098059 30 μmol/l and SB203580 30 μmol/l) were added 1 hour before cytokine stimulation. After 48 hours of treatment, RNA was isolated from cell culture using the Trizol-LS®TM method (Gibco-BRL, Life Technologies, Grand Island, NY USA), in accordance with the manufacturer's instructions. Briefly, 5 × 105 cells were lysed in 1000 μl Trizol-LS® reagent, and recovery of total RNA after isopropanol precipitation was measured using a spectrophotometer (Beckman DU62, Amersham Biosciences, Chalfont St. Giles, UK) at 260 nm.
Analysis of nitric oxide synthase type II gene expression using real-time RT-PCR
Real-time RT-PCR analyses were performed in a fluorescent temperature cycler (MX3000P Real Time PCR System; Stratagene), in accordance with the manufacturer's instructions. Total RNA 1 μg was used for each RT reaction. cDNAs were synthesized using 200 units of Moloney murine leukaemia reverse transcriptase (Gibco-BRL) and 6 μl dNTPs mix (10 mmol/l of each dNTP), 6 μl of first strand buffer (250 mmol/l Tris-HCl [pH 8.3], 375 mmol/l KCl, 15 mmol/l MgCl2; Gibco-BRL), 1.5 μl of 50 mmol/l MgCl2, 0.17 μl random hexamer solution (3 μg/μl; Gibco-BRL) and 0.25 μl of RNAse OutTM (recombinant ribonuclease inhibitor 40 μg/μl; Gibco-BRL), in a total volume of 30 μl. Reaction mixtures were incubated at 37°C for 50 min and at 42°C for 15 min. The RT reaction was terminated by heating at 95°C for 5 min and subsequently quick chilled on ice. The 50 μl amplification mixture (Brilliant SYBR Green QPC Master Mix; Stratagene) contained 2 μl of RT reaction products plus 0.75 μl (30 nmol/l) diluted reference dye, 150 nmol/l of each primer and nuclease-free, PCR grade water to adjust the final volume to 50 μl.
After a first enzyme activation step (95°C for 10 min), reactions were cycled 33 times using the following parameters for NOS type II detection: denaturation at 95°C for 40 s, annealing at 60°C for 1 min and extension at 72°C for 1 min. Mouse glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA (5'-TCCATGACAACTTTGGCATCGTGG-3' for upstream primer and 5'-GTTGCTGTTGAAGTCACAGGAGAC-3' for downstream primer; Genebank M32599) was amplified under the same conditions and was used as a normalizer gene. The amount of PCR products formed in each cycle was evaluated on the basis of SYBR Green I fluorescence. A final extension at 72°C over 10 min was followed by melting curve profiles as follows: 95°C for 1 min, ramping down to 45°C at a rate of 0.2°C/s, and heating slowly (0.5°C/cycle) to 95°C for a total of 81 cycles (30 s/cycle). Fluorescence was measured continuously to confirm amplification of specific transcripts (data not shown).
The oligonucleotide primers specific for mouse NOS type II were as follows: upstream primer 5'-CTCACTGGGACAGCACAGAA-3' and downstream primer 5'-TGGTCAAACTCTTGGGGTTC-3' (from Genbank U43428).
Cycle-to-cycle fluorescence emission readings were monitored and quantified using the second derivative maximum method from the MX3000P Real Time software package (Stratagene). This method determines the crossing points of individual samples using an algorithm that identifies the first turning point of the fluorescence curve. This turning point corresponds to the first maximum of the second derivative curve and correlates inversely with the log of the initial template concentration. NOS type II mRNA levels were normalized with respect to mouse GAPDH level in each sample.
Nitric oxide synthase type II western blot analysis
ATDC-5 chondrogenic cells were seeded in P100 plates until they reached 85–90% confluence. After overnight starvation in serum-free medium, cells were stimulated for 24 hours with leptin (800 nmol/l), alone or in combination with IL-1 (0.025 ng/ml). In order to demonstrate the involvement of JAK2, PI3K, MEK-1 and p38 kinase, the following specific pharmacological inhibitors were added 1 hour before cytokine stimulation: tyrphostin AG490 (5 and 10 μmol/l) and Tkip (20 and 50 μmol/l) for JAK2; LY294002 (1, 5 and 10 μmol/l) for PI3K; PD098059 (1, 10 and 30 μmol/l) for MEK-1; and SB203580 (1, 10 and 30 μmol/l) for p38 kinase. After stimulation, cells were rapidly washed with ice cold PBS and scraped in lysis buffer: 10 mmol/l Tris-HCl (pH 7.5), 5 mmol/l EDTA, 150 mmol/l NaCl, 30 mmol/l sodium pyrophosphate, 50 mmol/l sodium fluoride, 1 mmol/l sodium orthovanadate (Na3VO4), 10% glycerol, 0.5% Triton X-100, 1 mmol/l phenylmethylsulfonilfluoride, aprotinin, leupeptin and pepstatin A (10 mg/ml). Lysed cells were centrifuged at 13000 g for 15 min. Lysates from control or stimulated cells were collected and separated by SDS-PAGE on a 10% polyacrylamide gel. Proteins were subsequently transferred to a polyvinylidene difluoride transfer membrane (Hybond TM-P; Amersham International, Little Chalfont, UK) using a transfer semidry blot cell (BioRad Laboratories, Hercules, CA, USA). Blots were incubated with the appropriate antibody (mouse anti-NOS II antibody; purchased from Upstate Biotech, Lake Placid, NY, USA). Immunoblots were visualized using ECLPlus detection Kit (Amersham-Pharmacia Biotech, Barcelona, Spain) using horseradish peroxidase labelled secondary antibody. To confirm equal load in each sample, after stripping in glycine buffer at pH 3, membranes were reblotted with anti-actin antibody (Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA). The images of autoradiograms were captured and analyzed using a Typhoon 9410 digital variable mode imager (Amersham Biotech, Little Chalfont, UK).
Data analysis
Data are expressed as mean ± standard error of the mean of at least three independent experiments, each with at least three or more independent observations. Statistical analysis was performed using analysis of variance followed by the Student–Newman–Keuls or Bonferroni multiple comparison test with the Instat computerized package (GraphPad Software Inc., San Diego, CA, USA). i < 0.05 was considered statistically significant.
Results
Leptin synergistic effect over IL-1 induced nitrite production in chondrocytes
A leptin concentration of 800 nmol/l was found to be optimal for co-stimulatory experiments. This concentration was selected based on a braod set of previous dose–response experiments (data not shown). Because NOS type II stimulation with IL-1 at 0.05 ng/ml was maximal, a dose of 0.025 ng/ml was selected in order to avoid masking leptin synergism. As shown in Fig. 1, ATDC5 cells and human primary chondrocytes did not accumulate nitrites when stimulated with leptin alone; however, leptin was able to increase significantly nitrite accumulation induced by IL-1 when cells were co-stimulated with both cytokines (Fig 1a,c). This result was confirmed in terms of protein expression. Indeed, a clear-cut increase in levels of NOS type II protein was observed when cells were co-stimulated with leptin and IL-1 (Fig. 1b).
To confirm whether NO formation was produced via NOS type II, ATDC5 cells and human chondrocytes were incubated for 48 hours with both cytokines in the presence of the NOS type II inhibitor aminoguanidine (1 mmol/l), added 1 hour before cytokine administration. Aminoguanidine completely inhibited nitrite accumulation in the culture supernatant of human primary chondrocytes (Fig. 1c) and ATDC5 cells (Fig. 1d).
Janus kinase-2 inhibition blocks leptin/IL-1 induced nitric oxide production and nitric oxide synthase type II protein expression
We also investigated the role played by JAK2 in nitrite production evoked by co-stimulation with leptin and IL-1 by using tyrphostin AG490. This JAK2 inhibitor, added 1 hour before cytokine co-stimulation, completely blocked nitrite production (Fig. 2a). This result was confirmed in terms of protein expression, because cell pretreatment with tyrphostin AG490 significantly decreased NOS II protein expression in leptin/IL-1 co-stimulated cells (Fig. 2d). Intriguingly, tyrphostin AG490 was also able to inhibit nitrite accumulation induced by IL-1 alone, suggesting that leptin synergizes with fundamental pathways in IL-1 responses. To gain further insights into the involvement of JAK2, Tkip (a 12-mer SOCS-1 mimetic peptide that binds to the autophosphorylation site of JAK2) was added to ATDC5 cells 1 hour before they were stimulated with leptin or IL-1, or both cytokines. Tkip at 50 μmol/l was able to blunt completely leptin/IL-1 induced nitrite accumulation and NOS II protein expression (Fig. 2b,e). A lipophilic irrelevant peptide, MuIFN-γ95–125 (AKFEVNNPQVQRQAFNELIRVVHQLLPESSL), was used as control. Intriguingly, Tkip was also able to inhibit, in a dose–response manner, nitrite accumulation and NOS II protein expression in ATDC5 cells stimulated with IL-1 alone (Fig. 2c,e).
Effect of the specific signalling pathways inhibitors LY294002, PD098059 and SB203580 on leptin/IL-1 co-stimulation
In order to define the signalling pathway involved in the synergistic induction of NOS type II mediated by co-stimulation with leptin and IL-1 in cultured ATDC5 cells, we evaluated the effects of specific pharmacological inhibitors on other kinases, specifically PI3K, MEK-1 and p38 kinase.
We first investigated the effect of a specific inhibitor of PI3K, namely LY294002 (1, 2.5, 5 and 10 μmol/l) on leptin/IL-1 induced NO production. The addition of LY294002 1 hour before cytokine co-stimulation resulted in significant and dose-dependent decreases in NO production and NOS type II protein expression (Fig. 3a,a1).
In order to test whether MEK-1 (the mitogen-activated protein kinase [MAPK] kinase involved in extracellular signal-regulated kinase [ERK]-1 and ERK-2 phosphorylation/activation) participates in NOS type II induction via leptin/IL-1 co-stimulation, we used the specific MEK-1 inhibitor PD98059. When this inhibitor was added 1 hour before cytokine co-stimulation, significant dose-dependent decreases in NO production and NOS II protein expression were observed (Fig. 3b,b1).
Finally, because it has been shown that p38 kinase is involved in apoptotic processes induced by NO in chondrocytes, we tested whether this MAPK is also involved in NOS type II synergistic activation stimulated by leptin/IL-1. For this purpose, we used the specific p38 kinase inhibitor SB203580. Addition of this inhibitor 1 hour before leptin/IL-1 co-stimulation caused significant and dose-dependent decreases in NO production and NOS II protein expression (Fig. 3c,c1 [lower panel]).
Leptin synergism does not depend on chondrocyte differentiation state
In order to determine whether leptin/IL-1 synergism and its signalling pathway depend on the differentiation state of chondrocytes, we conducted similar experiments in mature and hypertrophic chondrocytes. We differentiated ATDC5 cells (see Materials and methods, above) into mature and hypertrophic chondrocytes, and tested co-stimulation and treatments with all specific inhibitors. Nitrite accumulation, evaluated in 15-day (mature) and in 21-day (hypertrophic) differentiated ATDC5 cells at 24 and 48 hours after treatment, was similar to that observed in the ATDC5 chondrogenic undifferentiated cell line (Fig. 4a–d). Note that in order to evaluate the involvement of PI3K, in some experiments we also used wortmannin at 10 μmol/l (a classical but not very specific PI3K inhibitor), yielding results similar to those obtained with LY294002.
Finally, a similar pattern was observed in human cultured primary chondrocytes. In these cells, leptin induced a strong increase in nitrite accumulation over that induced by IL-1, and the synergistic response was significantly inhibited by tyrphostin AG490, wortmannin, LY294002, PD98059 and SB203580 (Fig. 5).
Effect of leptin/IL-1 co-stimulation on nitric oxide synthase type II RNA expression
We finally studied NOS II mRNA expression in order to determine whether NO increase/inhibition was due to modulation of NOS type II mRNA expression. As shown in Fig. 6, NOS type II mRNA, evaluated using real-time PCR, was strongly expressed when cells were co-stimulated with leptin plus IL-1, and this expression was significantly reduced by tyrphostin AG490, wortmannin, LY294002, PD098059 and SB203580.
Discussion
In the present study we investigated the effect of leptin on NO production stimulated by IL-1. We found that leptin had a synergistic effect in the ATDC5 murine chondrogenic cell line, in differentiated mature and hypertrophic ATDC5 chondrocytes, and in human primary chondrocytes.
Leptin has been classified as a cytokine-like hormone, because of its structure and the homology of its receptors with members of the class I cytokine receptor superfamily. A proinflammatory role for leptin has previously been proposed. Several data show that leptin levels are increased by proinflammatory cytokine administration and in animal models of acute inflammation [9]. In addition, leptin regulates not only humoral but also cellular immune responses in antigen-induced arthritis models [20]. Nevertheless, there are only few reports of a direct action of leptin at the cellular level in cartilage [14,15].
NO controls a variety of cartilage functions, including loss of chondrocyte phenotype, chondrocyte apoptosis, and extracellular matrix degradation [2,3]. NOS type II is mainly expressed by immune cells in response to a wide range of proinflammatory cytokines [21,22]. In vitro, human articular cartilage is able to produce large amounts of NO [23], which can be enhanced by proinflammatory cytokines. In addition, NO production can be significantly increased by the presence of leptin, as shown in our previous work [15] and in the present study.
Here, we show that the IL-1 induced production of NO by ATDC5 murine chondrocytes and by human chondrocytes is significantly enhanced by leptin. It is noteworthy that, apart from blood, several sources of leptin and IL-1 have been identified in or around the joints in pathological conditions. IL-1 is produced by inflamed synovium and periarticular fat pad [24]. Interestingly, multipotent stromal cells from the infrapatellar fat produce leptin [25]. In addition, osteoarthritic human chondrocytes produce leptin, and leptin administration in rats induces over-expression of this hormone by articular chondrocytes [13]. Thus, in patients with inflammatory synovitis or osteoarthritis, there is a unique microenvironment in the cartilage characterized by elevated levels of both leptin and IL-1, due not only to local production but also to systemic increase [10,13,26]. It is conceivable that in this scenario leptin plays a significant proinflammatory role, as suggested by the findings presented here. Of further interest is our previous report [15] of the co-stimulatory effect of leptin and IFN-γ at the chondrocyte level.
We previously established that the early event in leptin/IFN-γ synergistic NOS type II activation was the involvement of JAK2 [15]; the present results confirm that JAK2 activation is also an early step in leptin/IL-1 induced NOS type II co-stimulation. The fact that tyrphostin AG490 blocks the leptin/IL-1 response implies that leptin synergizes with critical pathways in IL-1 response. It was surprising that tyrphostin AG490 also blocked the response to IL-1 alone, because JAK2 is not known to be required for IL-1 receptor transduction, and so one would expect the effect of tyrphostin AG490 to be partial. However, our results are in agreement with those reported by other investigators [27,28].
We also used Tkip in our experiments; Tkip is a 12-mer SOCS-1 mimetic lipophilic peptide (WLVFFVIFYFFR) that inhibits JAK2 autophosphorylation [29]. Interestingly, the behaviour of this peptide was similar to that of tyrphostin AG490 in terms of NOS II inhibition. It is conceivable that this peptide, because of its SOCS-1 mimetic properties, could inhibit IL-1/Toll-like receptor function in chondrocytes. SOCS-1 is a negative regulator of lipopolysaccharide-induced macrophage activation [30,31] and has been shown to bind to IL-1 receptor associated kinase [32]. This disrupts the cascade that leads to nuclear factor-κB (NF-κB) signalling and causes NOS inhibition. Of note, it has been demonstrated that tyrphostin AG490 inhibits IL-1 induced NF-κB activation in concentrations that also inhibit NOS II mRNA and protein synthesis. These findings suggest that JAK2 is required for NF-κB activation, which in turn mediates IL-1 induced NOS II expression in chondrocytes [28].
To gain further insights into the mechanism by which leptin, together with IL-1, promotes NO production, we evaluated the roles played by downstream signalling cascades using specific pharmacological inhibitors. First, we analyzed the involvement of PI3K. The role played by this kinase in the activation of NOS type II is quite controversial and remains the subject of debate. A number of studies support the view that PI3K activity down-regulates NOS type II, but there are several caveats to this view. For instance, insulin-like growth factor-II stimulates NOS type II expression and activity in myoblasts via a PI3K-dependent mechanism involving IκBα degradation and increased p65 NF-κB DNA binding activity [33], which is in agreement with recent evidence indicating that PI3K/protein kinase B is involved in NF-κB activation [34,35]. In addition, PI3K homologues (mammalian target of rapamycin/FKBP12–rapamycin associated protein) have been implicated in the phosphorylation and activation of NOS type II [36]. It should therefore be stressed that the interaction between NOS type II and PI3K may vary depending on the cell model, and so this interaction may be subject to tissue-specific regulation.
Our results also suggest that ERK-1/2 and p38 kinase play pivotal roles in the activation of NOS type II mediated by leptin/IL-1 co-stimulation. We found that ERK-1/2-specific pharmacological inhibition significantly decreased NO production induced by leptin/IL-1 co-stimulation in cultured chondrocytes. This result is in agreement with previous studies that associated ERK-1/2 activation with NOS type II induction by a combination of proinflammatory stimuli [37-40].
Finally, we found that the blockade of p38 kinase caused a significant decrease in NO production, NOS II mRNA expression and NOS II protein level. These data are concordant with previous reports that implicate p38 kinase in NOS type II upregulation in macrophages [41], neural cells [42,43] and chondrocytes [44].
Synergistic interactions of IL-1 with other soluble factors are not novel and have been reported in chondrocytes and other cell types. For instance, IL-1 synergizes with oncostatin M to induce markedly the expression of matrix metalloproteinase (MMP)-1, MMP-3, MMP-8 and MMP-13, as well as aggrecanase ADAM-TS4 [45]. In addition, a synergistic induction of MMP-1 by IL-1 and oncostatin M has been observed in human chondrocytes via a novel mechanism that involves STAT (signal transducer and activator of transcription) and activator protein-1 [46].
As far as we are aware, this is the first report that demonstrates the cooperative interaction between leptin and IL-1 in the induction of NO production in activated chondrocytes.
Conclusion
The present study shows that in human and ATDC5 murine cultured chondrocytes, leptin, together with IL-1, significantly increases the production of NO by a mechanism that implies upregulation of NOS type II mRNA and protein. In this modulation, an intracellular signalling pathway that involves JAK2 tyrosine kinase, PI3K and two members or the MAPK pathway (ERK and p38) is at play. The functional interplay of these pathways may be important for the onset as well as the maintenance of inflammatory responses in cartilage.
Abbreviations
ERK = extracellular signal-regulated kinase; GAPDH = glyceraldehyde-3-phosphate dehydrogenase; IFN = interferon; IL = interleukin; JAK = janus kinase; MAPK = mitogen-activated protein kinase; MEK = mitogen-activated protein kinase kinase; MMP = matrix metalloproteinase; NF-κB = nuclear factor-κB; NO = nitric oxide; NOS = nitric oxide synthase; PBS = phosphate-buffered saline; PI3K = phosphatidylinositol 3-kinase; RT-PCR = reverse transcription polymerase chain reaction; SOCS = suppressor of cytokine signalling.
Competing interests
The author(s) declare that they have no competing interests.
Acknowledgements
This work was supported by grants from Spanish Ministry of Health (FIS 01/3137 and PI-020431). Oreste Gualillo and Francisca Lago are recipients of a research contract from Spanish Ministry of Health, Instituto de Salud Carlos III (EXP 00/3051 and 99/3040). Miguel Otero is a recipient of a predoctoral fellowship funded by Xunta de Galicia. Rocío Lago is a recipient of a fellowship funded by Instituto de Salud Carlos III (Red Temática G03/152). We would like to thank Prof. Carlos Dieguez for his helpful advice and for his continued support during the realization of this work. The authors are very grateful to Dr Antonio Mera from Rheumatology Division and to Dr Jorge Fernadez Noya from Vascular Surgery Division of Santiago Univeristy Clinical Hospital for helping us in harvesting human tissues.
Figures and Tables
Figure 1 Leptin synergizes with IL-1 in inducing nitric oxide synthase (NOS) type II. Synergistic effect of leptin (OB) on nitrite (NO2-) accumulation and NOS type II protein expression induced by IL-1. Stimulations were conducted in serum-free conditions (a,b) in ATDC5 chondrogenic cells and (c) in human primary chondrocytes. NO2- accumulation is selectively inhibited by aminoguanidine (AG) both in (d) ATDC5 cells and in (panel c) human primary chondrocytes. Values are expressed as mean ± standard error of the mean. WB, western blot.
Figure 2 Janus kinase (JAK)2 inhibition blocks leptin/IL-1-induced nitric oxide (NO) production and nitric oxide synthase (NOS) type II protein expression. Effect of tyrphostin AG490 and Tkip on NO production and NOS II protein expression. The effect of tyrphostin AG490 was evaluated in terms of (a) nitrite accumulation in ATDC5 cells stimulated with leptin and IL-1, and in terms of (d) NOS II protein expression. The effect of Tkip was evaluated by nitrite accumulation in (b) leptin/IL-1 ATDC5 co-stimulated cells and in (c) IL-1 stimulated cells (panel c). (e) Effect of Tkip on NOS type II protein expression in leptin/IL-1 co-stimulated cells.
Figure 3 Involvement of phosphatidylinositol 3-kinase (PI3K), mitogen-activated protein kinase kinase (MEK)-1 and p38-kinase in leptin/IL-1-induced nitric oxide synthase (NOS). Dose-dependent effect of (a,a1) LY294002, (b,b1) PD098059 and (c,c1) SB203580 on nitrite (NO2-) production and NOS type II protein expression in stimulated and unstimulated ATDC5 cells. Stimulations were conducted in serum-free conditions. Each inhibitor was added 1 hour before cytokine co-stimulation. Values are expressed as mean ± standard error of the mean. OB, leptin; WB, western blot.
Figure 4 Leptin synergism does not depend upon chondrocyte differation state. Effect of different inhibitors on nitrite (NO2-) accumulation in 15-day differentiated ATDC5 cells stimulated or not with leptin, alone or in combination with IL-1, during (a) 24 and (b) 48 hours. The effect of inhibitors was also evaluated in 21-day differentiated ATDC5 cells, after (c) 24 or (d) 48 hours of stimulation with leptin and IL-1 (alone or in combination). Values are expressed as mean ± standard error of the mean. OB, leptin.
Figure 5 Leptin acts synergistically with IL-1 in human primary chondrocytes. Nitrite (NO2-) accumulation in leptin (OB)/IL-1 co-stimulated human primary chondrocytes. Stimulations were conducted in serum-free conditions in the presence or absence of tyrphostin AG490, wortmannin, LY294002, PD98059 and SB203580 inhibitors. Values are expressed as mean ± standard error of the mean.
Figure 6 Effect of leptin/IL-1 co-stimulation on nitric oxide synthase (NOS) type II mRNA expression. Real-time RT-PCR analysis of the expression of the inducible NOS type II mRNA in leptin (OB)/IL-1 co-stimulated ATDC5 cells. Stimulations (24 hours) were conducted in serum-free conditions. Specific inhibitors were added 1 hour before cytokine co-stimulation. Values are expressed as mean ± standard error of the mean.
==== Refs
Goldring MB The role of the chondrocyte in osteoathritis Arthritis Rheum 2000 43 1916 1926 11014341 10.1002/1529-0131(200009)43:9<1916::AID-ANR2>3.0.CO;2-I
Kim SJ Ju JW Oh CD Yoon YM Song WK Kim JH Yoo YJ Bang OS Kang SS Chun JS ERK-1/2 and p38 kinase oppositely regulate nitric oxide-induced apoptosis of chondrocytes in association with p53, caspase-3 and differentiation status J Biol Chem 2002 277 1332 1339 11689560 10.1074/jbc.M107231200
Sasaki K Hattori T Fujisawa T Takahashi K Inoue H Takigawa M Nitric oxide mediates interleukin-1-induced gene expression of matrix metalloproteinases and basic fibroblast growth factor in cultured rabbit articular chondrocytes J Biochem (Tokyo) 1998 123 431 439 9538225
Sandoval DA Davis SN Leptin: metabolic control and regulation J Diabetes Complications 2003 17 108 113 12614978 10.1016/S1056-8727(02)00167-8
La Cava A Matarese G The weight of leptin in immunity Nature Rev Immunol 2004 4 371 379 15122202 10.1038/nri1350
Fantuzzi G Faggioni R Leptin in the regulation of immunity, inflammation and hematopoiesis J Leukoc Biol 2000 68 437 446 11037963
Matarese G La Cava A The intricate interface between immune system and metabolism Trends Immunol 2004 25 193 200 15039046 10.1016/j.it.2004.02.009
Steppan CM Crawford DT Chidsey-Frink KL Ke H Swick AG Leptin is a potent stimulator of bone growth in ob/ob mice Regul Pept 2000 92 73 78 11024568 10.1016/S0167-0115(00)00152-X
Gualillo O Eiras S Lago F Dieguez C Casanueva FF Elevated serum leptin concentrations induced by experimental acute inflammation Life Sci 2000 67 2433 2441 11065166 10.1016/S0024-3205(00)00827-4
Felson DT Anderson JJ Naimark A Walker AM Meenan RF Obesity and knee osteoarthritis: the Framingham Study Ann Intern Med 1988 109 18 24 3377350
Karlson EW Mandl LA Aweh GN Sangha O Liang MH Grodstein F Total hip replacement due to osteoarthritis: the importance of age, obesity an other modifiable risk factors Am J Med 2003 114 93 98 12586227 10.1016/S0002-9343(02)01447-X
Oliveria SA Felson DT Cirillo PA Reed JI Walker AM Body weight, body mass index and incident symptomatic osteoarthritis of the hand, hip and knee Epidemiology 1999 10 161 166 10069252 10.1097/00001648-199903000-00010
Dumond H Presle N Terlain B Mainard D Loeuille D Netter P Pottie P Evidence for a key role of leptin in osteoarthritis Arthritis Rheum 2003 48 3118 3129 14613274 10.1002/art.11303
Figenschau Y Knutsen G Shahazeydi S Johansen O Sveinbjornsson B Human articular chondrocytes express functional leptin receptors Biochem Biophys Res Commun 2001 287 190 197 11549273 10.1006/bbrc.2001.5543
Otero M Gómez-Reino JJ Gualillo O Synergistic induction of nitric synthase type II: in vitro effect of leptin and interferon-γ in human chondrocytes and ATDC5 chondrogenic cells Arthritis Rheum 2003 48 404 409 12571850 10.1002/art.10811
Dayer JM The pivotal role of interleukin-1 in the clinical manifestations of rheumatoid arthritis Rheumatology 2003 42 Suppl ii3 ii10 12817089
Goldring SR Pathogenesis of bone and cartilage destruction in rheumatoid arthritis Rheumatology 2003 42 Suppl ii11 ii16 12817090
Akiyama H Shukunami C Nakamura T Hiraki Y Differential expressions of BMP family genes during chondrogenic differentiation of mouse ATDC5 cells Cell Struct Funct 2000 25 195 204 10984103 10.1247/csf.25.195
Thomas DP Sunters A Gentry A Grigoriadis AE Inhibition of chondrocyte differentiation in vitro by constitutive and inducible overexpression of the c-fos proto-oncogene J Cell Sci 2000 113 439 450 10639331
Busso N So A Chobaz-Peclat V Morard C Martinez-Soria E Talabot-Ayer D Gabay C Leptin signaling deficiency impairs humoral and cellular immune responses and attenuates experimental arthritis J Immunol 2002 168 875 882 11777985
Brune B von Knethen A Sandau KB Nitric oxide and its role in apoptosis Eur J Pharmacol 1998 351 261 272 9721017 10.1016/S0014-2999(98)00274-X
Del Carlo M Loeser RF Nitric oxide-mediated chondrocyte cell death requires the generation of additional reactive oxygen species Arthritis Rheum 2002 46 394 403 11840442 10.1002/art.10056
Vuolteenaho K Moilanen T Al-Saffar N Knowles RG Moilanen E Regulation of the nitric oxide production resulting from the glucocorticoid-insensitive expression of NOS type II in human osteoarthritic cartilage Osteoarthritis Cartilage 2001 9 597 605 11597171 10.1053/joca.2001.0431
Ushiyama T Chano T Inoue K Matsusue Y Cytokine production in the infrapatellar fat pad: another source of cytokines in knee synovial fluids Ann Rheum Dis 2003 62 108 112 12525378 10.1136/ard.62.2.108
Wickham MQ Erickson GR Gimble JM Vail TP Guilak F Multipotent stromal cells derived from the infrapatellar fat pad of the knee Clin Orthop 2003 412 196 212 12838072
Bokarewa M Bokarewa D Hultgren O Tarkowski A Leptin consumption in the inflamed joints of patients with rheumatoid arthritis Ann Rheum Dis 2003 62 952 956 12972473 10.1136/ard.62.10.952
Doi M Shichiri M Katsuyama K Ishimaru S Hirata Y Cytokine-activated Jak-2 is involved in inducible nitric oxide synthase expression independent from NF-kappaB activation in vascular smooth muscle cells Atherosclerosis 2002 160 123 132 11755929 10.1016/S0021-9150(01)00578-0
Mendes AF Caramona MM Carvalho AP Lopes MC Role of mitogen-activated protein kinases and tyrosine kinases on IL-1-Induced NF-kappaB activation and iNOS expression in bovine articular chondrocytes Nitric Oxide 2002 6 35 44 11829533 10.1006/niox.2001.0378
Flowers LO Johnson HM Mujtaba MG Ellis MR Haider SM Subramaniam PS Characterization of a peptide inhibitor of Janus kinase 2 that mimics suppressor of cytokine signaling 1 function J Immunol 2004 172 7510 7518 15187130
Kinjyo I Hanada T Inagaki-Ohara K Mori H Aki D Ohishi M Yoshida H Kubo M Yoshimura A SOCS1/JAB is a negative regulator of LPS-induced macrophage activation Immunity 2002 17 583 5891 12433365 10.1016/S1074-7613(02)00446-6
Nakagawa R Naka T Tsutsui H Fujimoto M Kimura A Abe T Seki E Sato S Takeuchi O Takeda K SOCS-1 participates in negative regulation of LPS responses Immunity 2002 17 677 687 12433373 10.1016/S1074-7613(02)00449-1
Akira S Takeda K Toll-like receptor signalling Nat Rev Immunol 2004 4 499 511 15229469 10.1038/nri1391
Kaliman P Canicio J Testar X Palacin M Zorzano A Insulin-like growth factor-II, phosphatidylinositol 3-kinase, nuclear factor-kappaB and inducible nitric-oxide synthase define a common myogenic signaling pathway J Biol Chem 1999 274 17437 17444 10364173 10.1074/jbc.274.25.17437
Sizemore N Leung S Stark GR Activation of phosphatidylinositol 3-kinase in response to interleukin-1 leads to phosphorylation and activation of the NF-kappaB p65/RelA subunit Mol Cell Biol 1999 19 4798 4805 10373529
Kane LP Shapiro VS Stokoe D Weiss A Induction of NF-kappaB by the Akt/PKB kinase Curr Biol 1999 9 601 604 10359702 10.1016/S0960-9822(99)80265-6
Salh B Wagey R Marotta A Tao JS Pelech S Activation of phosphatidylinositol 3-kinase, protein kinase B, and p70 S6 kinases in lipopolysaccharide-stimulated Raw 264.7 cells: differential effects of rapamycin, Ly294002, and wortmannin on nitric oxide production J Immunol 1998 161 6947 6954 9862729
Blanchette J Jaramillo M Olivier M Signalling events involved in interferon-gamma-inducible macrophage nitric oxide generation Immunology 2003 108 513 522 12667213 10.1046/j.1365-2567.2003.01620.x
Kan H Xie Z Finkel MS TNF-alpha enhances cardiac myocyte NO production through MAP kinase-mediated NF-kappaB activation Am J Physiol 1999 277 H1641 H1646 10516205
Kristof AS Marks-Konczalik J Moss J Mitogen-activated protein kinases mediate activator protein-1-dependent human inducible nitric-oxide synthase promoter activation J Biol Chem 2001 276 8445 8452 11112784 10.1074/jbc.M009563200
Han IO Kim HS Kim HC Joe EH Kim WK Synergistic expression of inducible nitric oxide synthase by phorbol ester and interferon-gamma is mediated through NF-kappaB and ERK in microglial cells J Neurosci Res 2003 73 659 669 12929133 10.1002/jnr.10706
Ajizian SJ English BK Meals EA Specific inhibitors of p38 and extracellular signal-regulated kinase mitogen-activated protein kinase pathways block inducible nitric oxide synthase and tumor necrosis factor accumulation in murine macrophages stimulated with lipopolysaccharide and interferon-gamma J Infect Dis 1999 179 939 944 10068590 10.1086/314659
Bhat NR Zhang P Lee JC Hogan EL Extracellular signal-regulated kinase and p38 subgroups of mitogen-activated protein kinases regulate inducible nitric oxide synthase and tumor necrosis factor-alpha gene expression in endotoxin-stimulated primary glial cultures J Neurosci 1998 18 1633 1641 9464988
Da Silva J Pierrat B Mary JL Lesslauer W Blockade of p38 mitogen-activated protein kinase pathway inhibits inducible nitric-oxide synthase expression in mouse astrocytes J Biol Chem 1997 272 28373 28380 9353295 10.1074/jbc.272.45.28373
Badger AM Cook MN Lark MW Newman-Tarr TM Swift BA Nelson AH Barone FC Kumar S SB 203580 inhibits p38 mitogen-activated protein kinase, nitric oxide production, and inducible nitric oxide synthase in bovine cartilage-derived chondrocytes J Immunol 1998 161 467 473 9647257
Koshy PJ Lundy CJ Rowan AD Porter S Edwards DR Hogan A Clark IM Cawston TE The modulation of matrix metalloproteinase and ADAM gene expression in human chondrocytes by interleukin-1 and oncostatin M: a time-course study using real-time quantitative reverse transcription-polymerase chain reaction Arthritis Rheum 2002 46 961 967 11953973 10.1002/art.10212
Catterall JB Carrere S Koshy PJ Degnan BA Shingleton WD Brinckerhoff CE Rutter J Cawston TE Rowan AD Synergistic induction of matrix metalloproteinase 1 by interleukin-1alpha and oncostatin M in human chondrocytes involves signal transducer and activator of transcription and activator protein 1 transcription factors via a novel mechanism Arthritis Rheum 2001 44 2296 2310 11665970 10.1002/1529-0131(200110)44:10<2296::AID-ART392>3.0.CO;2-9
| 15899045 | PMC1174950 | CC BY | 2021-01-04 16:02:36 | no | Arthritis Res Ther. 2005 Mar 4; 7(3):R581-R591 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1708 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17091589904310.1186/ar1709Research ArticleProliferation and differentiation potential of chondrocytes from osteoarthritic patients Tallheden Tommi [email protected] Catherine [email protected] Camilla [email protected]ögren-Jansson Eva [email protected] Lars [email protected] Lars 2Brittberg Mats [email protected] Anders [email protected] Department of Clinical Chemistry and Transfusion Medicine, Sahlgrenska University Hospital, Gothenburg, Sweden2 Department Orthopaedics, Sahlgrenska University Hospital, Gothenburg, Sweden2005 3 3 2005 7 3 R560 R568 6 9 2004 18 10 2004 30 12 2004 3 1 2005 Copyright © 2005 Tallheden et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Autologous chondrocyte transplantation (ACT) has been shown, in long-term follow-up studies, to be a promising treatment for the repair of isolated cartilage lesions. The method is based on an implantation of in vitro expanded chondrocytes originating from a small cartilage biopsy harvested from a non-weight-bearing area within the joint. In patients with osteoarthritis (OA), there is a need for the resurfacing of large areas, which could potentially be made by using a scaffold in combination with culture-expanded cells. As a first step towards a cell-based therapy for OA, we therefore investigated the expansion and redifferentiation potential in vitro of chondrocytes isolated from patients undergoing total knee replacement. The results demonstrate that OA chondrocytes have a good proliferation potential and are able to redifferentiate in a three-dimensional pellet model. During the redifferentiation, the OA cells expressed increasing amounts of DNA and proteoglycans, and at day 14 the cells from all donors contained type II collagen-rich matrix. The accumulation of proteoglycans was in comparable amounts to those from ACT donors, whereas total collagen was significantly lower in all of the redifferentiated OA chondrocytes. When the OA chondrocytes were loaded into a scaffold based on hyaluronic acid, they bound to the scaffold and produced cartilage-specific matrix proteins. Thus, autologous chondrocytes are a potential source for the biological treatment of OA patients but the limited collagen synthesis of the OA chondrocytes needs to be further explained.
==== Body
Introduction
Adult articular cartilage consists of a delicate system of cells and matrix proteins, which have the function of creating a viscoelastic tissue with high biomechanical stability and low friction. Even though the cartilage is exposed to continuous mechanical wear, there is surprisingly low turnover in cells and extracellular matrix [1], which could be a reason for the inability of adult articular cartilage to respond to injuries and subsequently repair lesions. This low potential of self-repair has led to the development of several techniques such as mosaic plastic, microfracture, periosteal transplantation and autologous chondrocyte transplantation (ACT), all seeking to create a functional and painless repair of articular cartilage defects.
In ACT, culture-expanded chondrocytes are transplanted under a cover of periosteum [2]; the method was initially aimed at the treatment of small isolated lesions. However, 10 years later, the indication has been expanded to include lesions up to 20 cm2 in size. This first generation of cell-based treatment has been followed by a second or third generation, consisting of culture-expanded cells loaded on a membrane or into a biodegradable scaffold before implantation [3,4]. One major advantage in using scaffolds as cell carriers is that the cells can be positioned in the lesion, thereby ensuring that the cells become evenly distributed in the defect. Subsequently, the degradation time of the scaffold needs to be controlled. This can be made by different combinations of poly-L-lactic acid and poly-(lactic-co-glycollic acid) [5] or by the esterification of hyaluronic acid [6,7]. The scaffold made of hyaluronic acid has additionally been shown to degrade into chondrogenically active components [8].
Another major advantage of using a scaffold for delivery of the cells is the potential for treating larger defects. This is especially interesting for young (under 60 years old) and active patients with developed osteoarthritis (OA), who at present lack an appropriate treatment alternative. The aetiology of OA has been suggested to contain a phenotypic alteration of the chondrocytes [9] and disturbance in the proteoglycan metabolism due to systematic, mechanical or unknown reasons. Chondrocytes isolated from OA cartilage have been shown to be more metabolically active than cells isolated from non-OA regions in the same joint [10], whereas chondrocytes isolated from less severe grades of OA cartilage synthesize normal matrix components [11].
When chondrocytes are isolated from their three-dimensional (3D) environment in the articular cartilage and expanded in monolayer cultures, the cells dedifferentiate and gradually lose their specific phenotype [12,13]. We have shown previously that dedifferentiated cells from ACT patients have the ability to differentiate into several mesenchymal phenotypes [14] and that during redifferentiation towards the chondrogenic phenotype the cells express genes known to be involved in the embryonic formation of cartilage [15].
We therefore proposed, as a first step towards cell-based treatments for OA, that culture-expanded cells from patients diagnosed for OA have the capacity to proliferate and produce matrix proteins in the same quantity as ACT chondrocytes when placed in a differentiation model.
Materials and methods
Cartilage harvest
Cartilage biopsies were harvested with a curved chisel from macroscopically affected and unaffected surplus cartilage from seven patients with OA (age 64 to 83 years), with OA grades 3 to 5 on the Ahlbäck scale [16], undergoing total knee replacement. The affected side was considered to be the femoral condyle on the concave side of the knee deformity; that is, the medial condyle in varus deformity and the lateral in valgus knees. In all patients the hip–knee–ankle angle was determined from standing whole-leg radiographs (an angle of more than 180° indicates a valgus knee deformity). The harvested biopsies were transported to the cell culture laboratory in sterile saline solution (0.9% NaCl; Fresenius Kabi, Uppsala, Sweden) supplemented with gentamicin sulphate (50 mg/l; Gibco, Paisley, Renfrewshire, UK) and amphotericin B (250 μg/ml; Gibco). Part of the cartilage biopsy was processed for histology, blinded and scored by two independent experienced researchers in accordance with a modified (biopsies without subchondral bone) Mankin scale [17], with a maximum score of 13. The rest of the biopsy was used for cell culture as described below. The donation of surplus cartilage was approved by the ethical committee at the Medical Faculty at Gothenburg University.
Cell culture
The chondrocytes were isolated from the surrounding matrix by mechanical mincing of the tissue with scalpel followed by enzymatic treatment overnight with collagenase (0.8 mg/ml; Worthington Biochemical Corp, Lakewood, NJ, USA) in Ham's F-12 medium (Invitrogen, Lidingö, Sweden), at 37°C in 7% CO2/93% air. The isolated cells were seeded at 104 cells/cm2 in culture flasks (Costar; Corning Incorporated, Corning, NY, USA) in DMEM/F12 medium (Invitrogen) supplemented with L-ascorbic acid (0.025 mg/ml; Apotekets produktionsenhet, Umeå, Sweden), gentamicin sulphate (50 mg/l; Gibco), amphotericin B (250 μg/ml) and L-glutamine (2 mM; Gibco) with the addition of 10% human serum [18]. In brief, the human serum was collected in transfusion bags (dry pack; JMS, Singapore) from healthy blood donors. The serum was left to coagulate overnight at 4 to 8°C, centrifuged, sterile filtered, divided into aliquots and frozen until use. The first medium change was made on day 6 and thereafter twice a week. When the cells reached 80% confluence, they were subcultured and frozen. Thawed cells were subcultured into new flasks (Costar) at a density of 4 × 103 cells/cm2.
Three-dimensional pellet culture
After passage 1, the cells were cultured in a 3D pellet culture system as described previously [15,19]. On days 7 and 14, the pellets were fixed in Histofix™ (Histolab Products AB, Göteborg, Sweden), dehydrated and embedded in paraffin. Sections 5 μm thick were cut and placed on microscope slides (Superfrost Plus; Menzel-Gläser, Braunschweig, Germany), deparaffinized and stained with Alcian blue/van Gieson or immunohistochemically with anti-type I collagen and anti-type II collagen antibodies.
Immunohistochemistry of pellets
Deparaffinized sections were digested with hyaluronidase, 8,000 units/ml (Sigma, St Louis, MO) in 0.1 M PBS for 60 min at 37°C and blocked with 3% BSA (Sigma) in PBS for 5 min. The primary antibodies (anti-type I and II collagen; ICN Biomedicals, Aurora, OH, USA), diluted 1:150 in PBS containing 3% BSA, were incubated with the sections for 1 hour at room temperature (20–22°C). The secondary antibody, peroxidase-conjugated goat anti-mouse (1:150; Jackson Immunoresearch Laboratories, West Grove, PA, USA) were applied to the sections for 1 hour at room temperature. A substrate kit (Vector VIP; Vector Laboratories, Burlingame, CA, USA) was used for visualization and the results were analysed with a Nikon Optiphot2-pol microscope (Nikon Instruments Inc, Melville, NY, USA). Goat cartilage and bone explants were used as a positive control; for a negative control the primary antibodies were omitted.
Biochemical analysis of pellets
On days 7 and 14, pellets were digested in papain (Sigma) solution (0.3 mg/ml in 20 mM sodium phosphate buffer, pH 7.4, containing 1 mM EDTA and 2 mM dithiothreitol) for 60 min at 60°C. The digested pellets were then mechanically dissolved by vortex-mixing and further analysed for DNA, glycosaminoglycan and hydroxyproline content as described previously [15]. All biochemical analyses was performed on triplicate pellets.
Cells in scaffold
Culture-expanded cells (passage 2), 106/cm2 or 5.0 × 106/cm2, were seeded on human serum precoated Hyaff-11 scaffolds (thickness 2 mm; Fidia Advanced Biopolymers, Abano Terme, Italy) in 100 μl in Ham's F12 medium (Invitrogen) supplemented with 20% human serum. After incubation overnight at 37°C in 7% CO2/93% air, the scaffolds were cultured in serum-free medium [15] in non-adherent dishes (Falcon four-well IVF; Becton Dickinson, Le Pont De Claix, France) for 14 days. After fixation, the scaffolds were embedded in paraffin, sectioned (10 μm thickness), stained with Alcian blue/van Gieson and analysed immunohistochemically for type II collagen as described above.
Isolation of total RNA
Total RNA was isolated from cells cultured in a monolayer (passage 1) and from day 7 pellets with the use of an RNeasy mini kit (Qiagen, Hilden, Germany) in accordance with the manufacturer's description. Before RNA isolation, the pellets were collected in a 1.5 ml micro-tube (Sahrstedt, Nümbrecht, Germany) containing RLT buffer (Qiagen) and disrupted by sonication. To remove cell debris and cartilage matrix proteins a QIAshredder column (Qiagen) was used. Contaminating genomic DNA was removed from the isolated RNA by using a DNA-free kit (Ambion, Huntingdon, UK) and total RNA content and purity were determined spectrophotometrically at 260 and 280 nm. In general, A260/A280 ratios of about 2 were considered to indicate acceptable purity of the samples [20].
Real-time PCR
Expression patterns of four cartilage genes were analysed by real-time PCR with an ABI PRISM 7000 (Applied Biosystems, Foster City, CA, USA) sequence detector and software system. TaqMan MGB probes (FAM dye-labelled) and primers for type I collagen (Hs00164004_m1) and type X collagen (Hs00166657_m1) were ordered from Applied Biosystems assays-on-demand (20× assay mixes). The gene-specific primers and probes for type II collagen 5'-TGG TGT CAA AGG TCA CAG AGG TTAT-3', antisense 5'-GGA ACC ACT CTC ACC CTT CACA-3', probe 5'-TCC CTT AGC ACC GTC CAG GCC TG-3', were designed by using Primer Express Software version 2.0 (Applied Biosystems). All genes were designed to amplify fragments of 70 to 150 base pairs; as endogenous control, 18S rRNA labelled with VIC/TAMRA was used (Applied Biosystems).
Reverse transcription in vitro was performed with 500 ng of total RNA with the use of random hexamer primers and TaqMan Reverse Transcription reagents (Applied Biosystems). Real-time PCR was performed with 5 μl of diluted (1:10) cDNA corresponding to 10 ng of RNA, 15 μl of TaqMan Universal PCR master mixture (Applied Biosystems), 1× assay-on-demand mixes of primers and TaqMan MGB probes. All samples were analysed in triplicate and PCR was performed in optical 96-well microtitre plates (Applied Biosystems). After an initial denaturation step at 95°C for 10 min, the cDNA products were amplified with 40 PCR cycles consisting of a denaturation step at 95°C for 15 s and an extension step at 60°C for 1 min.
To analyse the real-time PCR data, a standard curve method was used. The data were analysed with ABI Prism 7000 SDS software (Applied Biosystems). For each sample, the Ctsample values were determined as the cycle number at which all samples were in the exponential phase of amplification. By using the formulas below, a value (Y) was obtained as a measure of the gene expression correlated to the standard curve for that particular gene: X = (Ctsample - Intercept value)/Slope value; X10 = Y. The Y value for each cDNA sample and target sequence was divided by the Y value from the housekeeping gene (18S) for that particular sample to derive a ΔCt value (PE-ABI; Sequence Detector User Bulletin 2).
Statistical analysis
Biochemical differences between donors and chondrocytes isolated from affected and unaffected were analysed with a two-sided Student's t-test (two-sample equal variance). P < 0.05 was considered significant. All analyses were performed with cell samples from at least three separate donors unless otherwise indicated; as a comparison, surplus cells from three or four donors undergoing ACT were used [15].
Results
After histological preparation, four of the seven isolated biopsies were evaluated on the Mankin scale for severity of OA [17]. The score in these samples varied from 1.5 to 11, and in two of the patients a significant difference was found between the affected pathological and unaffected non-pathological side of the joint (Fig. 1). After mechanical and enzymatic isolation of the chondrocytes from the biopsies, no difference could be observed in the average number of cells per milligram of cartilage between the affected and unaffected sides (Table 1). These numbers did not differ from the average number of cells isolated from ACT patients [21].
In the primary cultures of the isolated chondrocytes from the unaffected and affected sides, floating matrix fragments were initially found in the affected cultures. These fragments did not seem to affect the proliferation ability and disappeared after the first change of medium. The cells from the unaffected and affected sides expanded with, on average, 0.21 and 0.22 cell doublings per day, respectively. After one passage (4.3 cell doublings) and 3 weeks of culture, 106 primary cells isolated from a 400 mg biopsy were expanded into 20 million cells.
When the expanded cells were cultured in serum-free medium in a redifferentiation model they formed spherical pellets overnight. During this shift from two-dimensional culture to 3D culture, the cells expressed increasing amounts of type II and type X collagen, whereas the expression of type I collagen was unchanged or slightly decreased (Fig. 2). No difference in expression of these typical cartilage genes was observed between affected, unaffected and ACT donors.
The shift from a proliferative to a matrix-synthesizing state was also demonstrated by an increase in the size of pellets from day 1 to day 14. The histological sections of these pellets showed flattened cells on the surface and round cells in the centre (Fig. 3). Spindle-shaped cells were found in the central part of the pellet in some donors, and the frequency of spindle-shaped cells was greater in samples isolated from biopsies with high Mankin scores. In the pellets, sulphated proteoglycans were detected by Alcian blue/van Gieson staining at both days 7 and 14 in all donors (Fig. 3). Metachromatic staining was normally found throughout the whole pellets, but slightly weaker staining was found in the day 7 pellet from the sample with the highest Mankin score (data not shown).
The increase in pellet size during the culture period was accompanied by a significant increase in DNA amounts in all samples, except from one donor with Mankin score 11 on the affected side, between days 7 and 14 (data not shown). During the same period there was an increase in proteoglycan accumulation in each cell in 63% of all samples. At day 14 no difference could be observed in the amount of proteoglycans per cell in pellets from the unaffected and affected sides (P > 0.05). In a comparison between OA chondrocytes and those from patients undergoing ACT only one sample, from the affected side of a female aged 81 years, had a significantly lower content of proteoglycans (P < 0.05; Fig. 4a).
The ability of the culture-expanded cells to form collagens in the pellet model was analysed biochemically and by immunoreactivity to type I and type II collagen (Fig. 3). The total amounts of collagen per cell were significantly lower in all OA samples than in those from ACT patients (Fig. 4b). By immunohistochemistry, type II collagen was detected in all donors at day 14, on both affected and unaffected sides, without any correlation with Mankin score. In the immunohistochemical analysis for type I collagen, both samples from one donor (a male aged 74 years) with Mankin scores of 6 (unaffected) and 11 (affected), stained positive at days 7 and 14. The other samples were only weakly positive at day 14.
To test the potential of using Hyaff-11 as a scaffold for the delivery of chondrocytes, the scaffold was seeded from the top with two different concentrations of cell suspensions of OA samples and samples from ACT patients. After the use of this technique, the chondrocytes could be detected throughout the whole thickness of the scaffold, but higher concentrations of cells were observed on the side of the scaffold from which the cells had been seeded (Fig. 5). This cell distribution was more obvious in the scaffolds seeded with OA chondrocytes than in those seeded with ACT chondrocytes.
Attached to the hyaluronic acid, the chondrocytes redifferentiated within the scaffold, as seen by the secretion of proteoglycans and the synthesis of type II collagen (Fig. 5). The expression of cartilage proteins was more obvious on the surface of the scaffolds seeded with the high cell density than those seeded with the low cell density, as shown by the increased intensity in staining with Alcian blue/van Gieson and in staining for type II collagen (Fig. 5).
Discussion
Chondrocytes isolated from OA cartilage are able to proliferate in a monolayer and redifferentiate in 3D models, demonstrating properties similar to those of non-OA chondrocytes used for ACT. This indicates that culture-expanded autologous chondrocytes from OA patients could potentially be used for resurfacing articular cartilage.
In this paper we studied the potential of chondrocytes isolated from patients with developed OA. During the initial monolayer culture, chondrocytes are extracted from their normal 3D environment and exposed to an artificial environment consisting of a plastic surface, culture medium and serum. The plastic provides a substrate for the growth of the anchorage-dependent cells and the culture medium stabilizes pH and osmolarity and supplements the cells with trace compounds and energy sources (pyruvate and glucose). The added serum contains high levels of growth factors released by cells and platelets during the coagulation process of whole blood and has the ability to stimulate cell proliferation [18]. In this artificial environment enriched in growth factors the chondrocytes proliferate, dedifferentiate and lose their phenotype. The ability of these dedifferentiated cells to redifferentiate into the chondrogenic phenotype has been proven to be affected by the growth factors used during expansion [22] and the number of cell divisions [23].
In ACT treatments, 106 cells/cm2 are implanted into the defects under a covering of periosteum or type I/III collagen membrane [24]. If this treatment were to be used for OA patients, most probably both the femoral condyle and the tibial plateau would need restoration. This would mean that surfaces about at least 25 cm2 in size should be treated. In the present study we were able to obtain, from a 400 mg cartilage biopsy taken from OA patients, 20 million cells within 3 weeks of culture. This means that without exceeding the number of cell divisions, which could possibly hamper the redifferentiation potential [25], it would be necessary to harvest about 500 mg of cartilage, which correlates to a circular biopsy 7.2 mm in diameter on the basis of calculations of normal hyaline cartilage [26]. The data in this study indicate that the biopsy could be harvested either from a non-weight-bearing area or from the actual affected area during a cleanout pre-arthroscopic procedure. However, 106 cells/cm2 greatly exceeds the cell density in adult cartilage, and the number of cells actually needed for a successful scaffold-assisted cartilage repair has not been defined.
It has previously been demonstrated in several studies that cells isolated from OA cartilage have limited proliferation capacity [27] and malfunctioning proteoglycan synthesis [10,11]. It was therefore a great surprise to us that we observed a proliferation rate similar to that in samples from patients treated with ACT and no difference in the proteoglycan secretion in chondrocytes isolated from affected and unaffected areas. All samples had the further ability to produce type II collagen in the pellet model. Possible explanations for this are that during the proliferation phase the cells are exposed to an environment and to growth factors, which 'revitalizes' the cells [10], or simply that there is a positive selection of potent cells during the monolayer culture.
Although the chondrocytes from OA patients analysed in this study produced a cartilage-specific matrix, the ability of the chondrocytes to redifferentiate seemed be different from that of chondrocytes isolated from ACT donors [15]. Whereas ACT chondrocytes, once placed in the 3D serum-free pellet model, stopped their DNA synthesis and started to differentiate, the OA chondrocytes continued to proliferate up to day 14. The proliferation was accompanied with significantly less collagen production in all OA chondrocytes than in ACT chondrocytes (Fig. 4b). A shift from a differentiated phenotype to a proliferative state has further been suggested as an explanation for the development of OA [9,28] and could possibly be reflected in the inability to redifferentiate seen in the pellet model.
Another important issue in cell-based cartilage repair, especially for large defects, is the positioning of cells in large defects. This can be done by delivering the cells to the patient within a vehicle or a scaffold. Within the scaffold, which is preferably biodegradable and has a controlled degradation time, the cells are able to attach and to start producing cartilage matrix. In our study we observed that, within the hyaluronic acid scaffold after 2 weeks of culture in serum-free medium, the OA chondrocytes formed cartilage matrix proteins. This result concurs with previous studies with human epiphyseal chondrocytes and chick embryonic sternal chondrocytes, in which an increased expression of cartilage typical genes was observed in cultures with Hyaff-11 (scaffold based on hyaluronic acid) [29].
The redifferentiation of the dedifferentiated cells was typically more obvious in the scaffolds seeded with the high density of cells (25 × 106 cells/cm3), indicating that the cell density is important for the restoration of the chondrogenic phenotype. The cell density and redifferentiation could also be important for matrix production, because in the scaffolds seeded with a low cell density (5 × 106 cells/cm3) we observed a threefold to fivefold lower secretion of proteoglycans compared with the pellet cultures (data not shown). Similar observations have been presented by Puelacher and colleagues [30], who showed that at least 20 × 106 cells/cm3 were needed for good matrix formation within the scaffold.
Further, it is of great importance that the scaffold, when implanted into the joint, has the ability to integrate with the surrounding cartilage and with the subchondral bone. Integration with the subchondral bone could possibly be increased by the induction of subchondral bleeding, for example by microfracture. However, the importance of an uninjured subchondral bone plate for the integrity of the articular cartilage and the ability to withstand mineralization has not been clarified.
The integration could also be altered by the grade of differentiation of the scaffold, as demonstrated in a study made by Obradovic and colleagues [31]. They showed that the integration of tissue-engineered cartilage to articular cartilage explants was better with immature (redifferentiated for 5 days) than mature (redifferentiated for 5 weeks) cartilaginous explants. The positive immunostaining of type II collagen seen in our scaffold seeded with the higher density of cells could indicate that the cells had redifferentiated too far and that the implant would therefore be less integrative. In contrast, in the treatment of large injuries, the scaffold needs to be able to withstand mechanical load and shear forces from the time of implantation. These forces can possibly be lowered by alignment of the mechanical axis (tibia osteotomy) to reduce the weight bearing of the implant, but the scaffold will in any case be subjected to mechanical stress and will have to be able to withstand this. A possible way of strengthening the scaffold without redifferentiation would be to distribute the chondrocytes more uniformly in the scaffold by improving the seeding method. Both spinner flask and perfusion culture techniques have been shown to be superior to static cultures [32].
Another reason that the implant has to withstand mechanical stress is that systematic redifferentiation signalling, as part of the disease condition, could be impaired within the OA joint. Redifferentiation and proteoglycan synthesis could instead be stimulated by dynamic mechanical compression of the implant. The mechanical load could possibly be gradually increased to adapt to the differentiation state of the implant through specifically developed physiotherapy programmes, which will therefore probably have an important role in the development of biological implants for OA.
Conclusion
We demonstrate in this paper that OA chondrocytes have the ability to proliferate, redifferentiate and secrete cartilage-specific matrix proteins. We also show that OA chondrocytes have an inability to shift definitely from a proliferative to a differentiating state. How to change the cells from a proliferative to a collagen-secreting phenotype needs to be explored further, especially when considering the importance of collagen in maintaining the cartilage structure.
We further showed that the OA chondrocytes are able to bind to a scaffold, but further studies will be needed to establish how far the cartilage in this scaffold should be differentiated to be able to integrate with the surrounding cartilage and subchondral bone and to withstand the mechanical forces applied within the joint.
The results in this paper give hopes for finding a cell-based autologous biological treatment for young active patients with OA, but we have to remember that there is no normal cartilage in OA and more research must be done before such a treatment can be put into clinical practice.
Abbreviations
3D = three-dimensional; ACT = autologous chondrocyte transplantation; BSA = bovine serum albumin; DMEM = Dulbecco's modified Eagle's medium; PCR = polymerase chain reaction; PBS = phosphate-buffered saline; OA = osteoarthritis.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
TT conceived of the study, coordinated the experiments, performed immunohistochemical staining, performed the statistical analysis and drafted the manuscript. C Bengtsson performed the cell culture and RNA preparations. C Brantsing performed immunohistochemical stainings and the quantitative PCR analysis. ESJ participated in the design of the study and gave clinical cell culturing input. LC isolated the biopsies and gave clinical feedback. LP and MB provided critical clinical input to the study design and to the manuscript. AL conceived of the study and gave critical comments on the manuscript. All authors read and approved the final manuscript.
Acknowledgements
We acknowledge assistance from Mrs Helena Barreto in the biochemical analysis of the pellets and scaffolds, and from Fidia Biopolymers in providing us with the Hyaff scaffold.
Figures and Tables
Figure 1 Histology of biopsies. The figure shows sections stained with Alcian blue/van Gieson. The biopsies originate from two representative autologous chondrocyte transplantation patients (ACT) (a, b) and from the unaffected and affected side from two patients with osteoarthritis (OA) undergoing total knee replacement (c–f). (c, e) Biopsy from one OA donor (female, aged 81 years) with a Mankin score of 1.5 on both sides. (d, f) Biopsy from another OA donor (male, aged 74 years) with Mankin scores of 5 (d) and 11 (f).
Figure 2 Gene expression of cells cultured in pellets. The graphs show the quantitative gene expression of typical cartilage gene expression markers (types I, II and X collagen) in the monolayer (ML) and in day 7 pellets. Results are means ± SD for separate donors (n = 4) from samples from autologous chondrocyte transplantation patients (black bars) and from affected (grey bars) and unaffected (white bars) areas.
Figure 3 Histology of pellets. The figure shows stained sections of pellets from the unaffected and affected side from same OA donors ((a–f) female, age 81 years and (g–l) male, age 74 years) as shown in Fig. 1. (a, d, g, f) Alcian blue–van Gieson staining indicating the accumulation of proteoglycans. (b, e, h, k) Immunohistochemical staining with anti-type II collagen antibody. (c, f, i, l) Immunohistochemical staining with anti-type I collagen antibody. Positive staining is indicated by a red colour in the extracellular matrix.
Figure 4 Biochemical analysis of pellets. Results are shown of the measurement of cartilage matrix protein accumulation in day 14 pellets normalized to cell number (DNA) in cells from the unaffected (filled bars) and affected (open bars) sides of four consecutive donors; the sex and age (in years) of the donors are indicated. Glycosaminoglycan (GAG) (a) and hydroxyproline (as a measure of total collagen content) (b) are shown as amounts per microgram of DNA. Results are means ± SD for three identical pellets. The amounts are compared with the mean value for four sequential autologous chondrocyte transplantation patients (ACT). An asterisk indicates a significant difference (P < 0.05) between ACT and osteoarthritis chondrocytes.
Figure 5 Histology of cell-seeded scaffolds. The figure shows scaffolds seeded with chondrocytes from one representative autologous chondrocyte transplantation (ACT) donor (a–c) and one osteoarthritis (OA) donor (unaffected) at two different cell seeding densities, 106 cells/cm2 (d–f) and 5 × 106 cells/cm2 (g–i). Accumulation of proteoglycans is shown with Alcian blue/van Gieson stain and the presence of type II collagen is indicated by red staining within the scaffold. Panels (b), (e) and (h) are higher magnifications of portions of (a), (d) and (g), respectively.
Table 1 Clinical diagnosis and histological scores of the seven donors
No. Cartilage Age (years) Sex Ahlbäck Varus Valgus HKA angle Diagnosis Mankin Cells/mg
1 Affected 73 Male III Yes 175° Prim OA n/a n/a
Unaffected n/a n/a
2 Affected 74 Male IV Yes 174° Prim OA n/a 2,912
Unaffected n/a 2,422
3 Affected 83 Female V Yes 184° Prim OA n/a 2,147
Unaffected n/a 2,917
4 Affected 81 Female II Yes 182° Sec OA 4.5 2,054
Unaffected 5 1,923
5 Affected 74 Male III Yes 176° Prim OA 11* 1,310
Unaffected 5.5 1,938
6 Affected 64 Female IV Yes 167° Prim OA 1.5 3,465
Unaffected 1.5 3,091
7 Affected 81 Female IV Yes 173° Prim OA 9* 2,649
Unaffected 6.5 2,410
An asterisk indicates a significant difference (P < 0.05) between unaffected and affected biopsies. HKA angle, hip–knee–ankle angle; n/a, not analysed; OA, osteoarthritis; Prim, primary; Sec, secondary.
==== Refs
Muir IHM Freeman MAR Biochemistry Adult Articular Cartilage 1979 2 Tunbridge Wells: Pitman Medical 1 68
Brittberg M Lindahl A Nilsson A Ohlsson C Isaksson O Peterson L Treatment of deep cartilage defects in the knee with autologous chondrocyte transplantation N Engl J Med 1994 331 889 895 8078550 10.1056/NEJM199410063311401
Cherubino P Grassi FA Bulgheroni P Ronga M Autologous chondrocyte implantation using a bilayer collagen membrane: a preliminary report J Orthop Surg (Hong Kong) 2003 11 10 15 12810965
Pavesio A Abatangelo G Borrione A Brocchetta D Hollander AP Kon E Torasso F Zanasi S Marcacci M Hyaluronan-based scaffolds (Hyalograft C) in the treatment of knee cartilage defects: preliminary clinical findings Novartis Found Symp 2003 249 203 217 12708658
Mooney DJ Mazzoni CL Breuer C McNamara K Hern D Vacanti JP Langer R Stabilized polyglycolic acid fibre-based tubes for tissue engineering Biomaterials 1996 17 115 124 8624388 10.1016/0142-9612(96)85756-5
Benedetti L Cortivo R Berti T Berti A Pea F Mazzo M Moras M Abatangelo G Biocompatibility and biodegradation of different hyaluronan derivatives (Hyaff) implanted in rats Biomaterials 1993 14 1154 1160 8130320 10.1016/0142-9612(93)90160-4
Aigner J Tegeler J Hutzler P Campoccia D Pavesio A Hammer C Kastenbauer E Naumann A Cartilage tissue engineering with novel nonwoven structured biomaterial based on hyaluronic acid benzyl ester J Biomed Mater Res 1998 42 172 181 9773813 10.1002/(SICI)1097-4636(199811)42:2<172::AID-JBM2>3.0.CO;2-M
Solchaga LA Dennis JE Goldberg VM Caplan AI Hyaluronic acid-based polymers as cell carriers for tissue-engineered repair of bone and cartilage J Orthop Res 1999 17 205 213 10221837 10.1002/jor.1100170209
Sandell LJ Aigner T Articular cartilage and changes in arthritis. An introduction: cell biology of osteoarthritis Arthritis Res 2001 3 107 113 11178118 10.1186/ar148
Teshima R Treadwell BV Trahan CA Mankin HJ Comparative rates of proteoglycan synthesis and size of proteoglycans in normal and osteoarthritic chondrocytes Arthritis Rheum 1983 26 1225 1230 6626281
Bulstra SK Buurman WA Walenkamp GH Van der Linden AJ Metabolic characteristics of in vitro cultured human chondrocytes in relation to the histopathologic grade of osteoarthritis Clin Orthop 1989 242 294 302 2706860
Schnabel M Marlovits S Eckhoff G Fichtel I Gotzen L Vecsei V Schlegel J Dedifferentiation-associated changes in morphology and gene expression in primary human articular chondrocytes in cell culture Osteoarthritis Cartilage 2002 10 62 70 11795984 10.1053/joca.2001.0482
Diaz-Romero J Gaillard JP Grogan SP Nesic D Trub T Mainil-Varlet P Immunophenotypic analysis of human articular chondrocytes: changes in surface markers associated with cell expansion in monolayer culture J Cell Physiol 2005 202 731 742 15389573 10.1002/jcp.20164
Tallheden T Dennis JE Lennon DP Sjogren-Jansson E Caplan AI Lindahl A Phenotypic plasticity of human articular chondrocytes J Bone Joint Surg Am 2003 85A Suppl 2 93 100
Tallheden T Karlsson C Brunner A Van Der Lee J Hagg R Tommasini R Lindahl A Gene expression during redifferentiation of human articular chondrocytes Osteoarthritis Cartilage 2004 12 525 535 15219567 10.1016/j.joca.2004.03.004
Ahlback S Osteoarthrosis of the knee. A radiographic investigation Acta Radiol Diagn (Stockh) 1968 7 72 5706059
Mankin HJ Dorfman H Lippiello L Zarins A Biochemical and metabolic abnormalities in articular cartilage from osteo-arthritic human hips. II. Correlation of morphology with biochemical and metabolic data J Bone Joint Surg Am 1971 53 523 537 5580011
Tallheden T Van der Lee J Brantsing C Sjögren Jansson E Lindahl A Human serum for articular chondrocyte culture Cell Transplantation 2005
Johnstone B Hering TM Caplan AI Goldberg VM Yoo JU In vitro chondrogenesis of bone marrow-derived mesenchymal progenitor cells Exp Cell Res 1998 238 265 272 9457080 10.1006/excr.1997.3858
Martin I Jakob M Schafer D Dick W Spagnoli G Heberer M Quantitative analysis of gene expression in human articular cartilage from normal and osteoarthritic joints Osteoarthritis Cartilage 2001 9 112 118 11237658 10.1053/joca.2000.0366
Sjögren Jansson E Tallheden T Lindahl A The Eurocell Experience of Autologous Chondrocyte Implantation 2004 Oswestry: The Institute of Orthopaedics (Oswestry) Publishing Group
Jakob M Demarteau O Schafer D Hintermann B Dick W Heberer M Martin I Specific growth factors during the expansion and redifferentiation of adult human articular chondrocytes enhance chondrogenesis and cartilaginous tissue formation in vitro J Cell Biochem 2001 81 368 377 11241676 10.1002/1097-4644(20010501)81:2<368::AID-JCB1051>3.0.CO;2-J
Dell'Accio F De Bari C Luyten FP Molecular markers predictive of the capacity of expanded human articular chondrocytes to form stable cartilage in vivo Arthritis Rheum 2001 44 1608 1619 11465712 10.1002/1529-0131(200107)44:7<1608::AID-ART284>3.0.CO;2-T
Haddo O Mahroof S Higgs D David L Pringle J Bayliss M Cannon SR Briggs TW The use of chondrogide membrane in autologous chondrocyte implantation Knee 2004 11 51 55 14967329 10.1016/S0968-0160(03)00041-3
Dell'Accio F Vanlauwe J Bellemans J Neys J De Bari C Luyten FP Expanded phenotypically stable chondrocytes persist in the repair tissue and contribute to cartilage matrix formation and structural integration in a goat model of autologous chondrocyte implantation J Orthop Res 2003 21 123 131 12507589 10.1016/S0736-0266(02)00090-6
Hunziker EB Quinn TM Hauselmann HJ Quantitative structural organization of normal adult human articular cartilage Osteoarthritis Cartilage 2002 10 564 572 12127837 10.1053/joca.2002.0814
Dozin B Malpeli M Camardella L Cancedda R Pietrangelo A Response of young, aged and osteoarthritic human articular chondrocytes to inflammatory cytokines: molecular and cellular aspects Matrix Biol 2002 21 449 459 12225810 10.1016/S0945-053X(02)00028-8
Pfander D Swoboda B Kirsch T Expression of early and late differentiation markers (proliferating cell nuclear antigen, syndecan-3, annexin VI, and alkaline phosphatase) by human osteoarthritic chondrocytes Am J Pathol 2001 159 1777 1783 11696438
Girotto D Urbani S Brun P Renier D Barbucci R Abatangelo G Tissue-specific gene expression in chondrocytes grown on three-dimensional hyaluronic acid scaffolds Biomaterials 2003 24 3265 3275 12763454 10.1016/S0142-9612(03)00160-1
Puelacher WC Kim SW Vacanti JP Schloo B Mooney D Vacanti CA Tissue-engineered growth of cartilage: the effect of varying the concentration of chondrocytes seeded onto synthetic polymer matrices Int J Oral Maxillofac Surg 1994 23 49 53 8163862
Obradovic B Martin I Padera RF Treppo S Freed LE Vunjak-Novakovic G Integration of engineered cartilage J Orthop Res 2001 19 1089 1097 11781010 10.1016/S0736-0266(01)00030-4
Wendt D Marsano A Jakob M Heberer M Martin I Oscillating perfusion of cell suspensions through three-dimensional scaffolds enhances cell seeding efficiency and uniformity Biotechnol Bioeng 2003 84 205 214 12966577 10.1002/bit.10759
| 15899043 | PMC1174951 | CC BY | 2021-01-04 16:02:36 | no | Arthritis Res Ther. 2005 Mar 3; 7(3):R560-R568 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1709 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17101589904210.1186/ar1710Research ArticleQuantitative ultrasound can assess the regeneration process of tissue-engineered cartilage using a complex between adherent bone marrow cells and a three-dimensional scaffold Hattori Koji [email protected] Yoshinori [email protected] Hajime [email protected] Takashi [email protected] Kota [email protected] Jun [email protected] Kenji [email protected] Takashi [email protected] Masao [email protected] Ken [email protected] Department of Orthopaedic Surgery, Nara Medical University, Nara, Japan2 National Institute of Advanced Industrial Science and Technology, Amagasaki Site, Hyogo, Japan3 Life Science Laboratories, Life Science RD Center, Kaneka Corporation, Takasago, Hyogo, Japan4 Institute for Frontier Medical Sciences, Kyoto University, Kyoto, Japan2005 1 3 2005 7 3 R552 R559 10 1 2005 25 1 2005 1 2 2005 8 2 2005 Copyright © 2005 Hattori et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Articular cartilage (hyaline cartilage) defects resulting from traumatic injury or degenerative joint disease do not repair themselves spontaneously. Therefore, such defects may require novel regenerative strategies to restore biologically and biomechanically functional tissue. Recently, tissue engineering using a complex of cells and scaffold has emerged as a new approach for repairing cartilage defects and restoring cartilage function. With the advent of this new technology, accurate methods for evaluating articular cartilage have become important. In particular, in vivo evaluation is essential for determining the best treatment. However, without a biopsy, which causes damage, articular cartilage cannot be accurately evaluated in a clinical context. We have developed a novel system for evaluating articular cartilage, in which the acoustic properties of the cartilage are measured by introducing an ultrasonic probe during arthroscopy of the knee joint. The purpose of the current study was to determine the efficacy of this ultrasound system for evaluating tissue-engineered cartilage in an experimental model involving implantation of a cell/scaffold complex into rabbit knee joint defects. Ultrasonic echoes from the articular cartilage were converted into a wavelet map by wavelet transformation. On the wavelet map, the percentage maximum magnitude (the maximum magnitude of the measurement area of the operated knee divided by that of the intact cartilage of the opposite, nonoperated knee; %MM) was used as a quantitative index of cartilage regeneration. Using this index, the tissue-engineered cartilage was examined to elucidate the relations between ultrasonic analysis and biochemical and histological analyses. The %MM increased over the time course of the implant and all the hyaline-like cartilage samples from the histological findings had a high %MM. Correlations were observed between the %MM and the semiquantitative histologic grading scale scores from the histological findings. In the biochemical findings, the chondroitin sulfate content increased over the time course of the implant, whereas the hydroxyproline content remained constant. The chondroitin sulfate content showed a similarity to the results of the %MM values. Ultrasonic measurements were found to predict the regeneration process of the tissue-engineered cartilage as a minimally invasive method. Therefore, ultrasonic evaluation using a wavelet map can support the evaluation of tissue-engineered cartilage using cell/scaffold complexes.
==== Body
Introduction
Defects in articular cartilage (hyaline cartilage) resulting from traumatic injury or degenerative joint disease do not repair themselves spontaneously, because of the low mitotic activity of chondrocytes and the avascular nature of this type of cartilage [1,2]. Therefore, defects may require novel regenerative strategies to restore the biological and biomechanical function of the tissue. Recently, tissue engineering using cell/scaffold complexes has emerged as an approach for repairing cartilage defects and restoring cartilage function [3-5]. However, little is known about which scaffolds and which cells (chondrocytes or cells derived from bone marrow) are effective for the treatment of cartilage defects. Furthermore, the length of time required for chondrocyte maturation or stem cell differentiation into hyaline cartilage is unknown.
With the advent of new technologies in scaffold processing and cell biology, accurate methods for evaluating articular cartilage have become important. In particular, in vivo evaluation is essential for determining the best treatment. However, without a biopsy, which causes damage, articular cartilage cannot be accurately evaluated in a clinical context.
We therefore developed a new ultrasonic evaluation system for articular cartilage and showed that this system can quantitatively evaluate cartilage degeneration clinically [6,7]. The analysis system is based on wavelet transformation of the reflex echogram from articular cartilage. Our previous study revealed that the system could predict the histological findings for tissue-engineered cartilage [8,9]. However, it remained to be seen whether this system could accurately evaluate tissue-engineered cartilage from cell/scaffold complexes, especially the regeneration process. The purpose of the present study was to find out. Therefore, we fabricated three-dimensional scaffolds using a biodegradable polymer to engineer hyaline-cartilage-like tissue derived from adherent bone marrow cells and evaluated the tissue-engineered cartilage after implantation in rabbit cartilage defects. We investigated whether ultrasound could evaluate the regeneration process at 4 and 12 weeks after the implantation of a cell/scaffold complex. The relations between the ultrasonic examination and histological or biochemical examinations were analyzed.
Materials and methods
Three-dimensional PLGA scaffold
The biodegradable scaffolds (GC Corporation, Tokyo, Japan) used in this study were described previously [10-12]. The scaffolds (5 mm in diameter, 1.5 mm thick) were composed of poly(lactic-glycolic acid) (PLGA) with a molecular mass of approximately 100,000. The outline of the scaffold construction is described below. Poly(DL-lactic-co-glycolic acid) was dissolved in dioxane added to sodium citrate particles and then frozen. The PLGA scaffold was created by a series of processes involving evaporating the solvent, washing with water to remove salts, and drying the frozen PLGA/sodium citrate. The pores at the top of the scaffold were created by the salt leaching and those at the bottom were made by the solvent evaporation. Therefore, the scaffold had micropores on the top surface and had numerous cylindrical boreholes (Fig. 1), and within the scaffold the cells lay in a uniform array at the palisade. The average pore size in the unit area on the top surface of the scaffold was 300μm. Since the micropores were present only on the top surface, the cultured cells infiltrated the scaffold after instillation of the cell suspension and did not leak out.
Culture of adherent bone marrow cells
Twenty adult male Japanese white rabbits (3 to 4 kg) were used in this study; they were individually maintained in stainless-steel cages. The rabbits were anesthetized with a mixture of ketamine (50 mg/ml) and xylazine (20 mg/ml) at a ratio of 2:1, via a dose of 1 ml/kg injected intramuscularly into the gluteal muscle. Bone marrow was then isolated from the humeral head using an 18-gauge bone marrow needle, and 5 ml of the marrow was drawn into a 10-ml syringe containing 0.1 ml heparin (3,000 U/ml). The released cells were transferred to a T-75 flask (Costar, Cambridge, MA, USA) containing 15 ml of medium. The medium used was Eagle's minimal essential medium (MEM) containing 10% fetal bovine serum and antibiotics (penicillin, 100 U/ml; streptomycin, 0.1 mg/ml; and amphotericin B (Fungizone), 0.25 g/ml; all from Sigma Chemicals, St Louis, MO, USA). The cells were grown in a humidified atmosphere of 5% carbon dioxide at 37°C and the medium was replaced with fresh medium every 2 days. No growth factors were added. The cell culture was maintained for 2 weeks until the cells reached confluence, and then the cultured adherent bone marrow cells were released from the substratum using 0.25% trypsin and counted in a hemocytometer. The cultured cells obtained from each rabbit were reseeded onto three-dimensional PLGA scaffolds by simply dropping the cell suspension onto the scaffolds. The density of the cultured cells in a scaffold was 1 × 107 cells/cm3. To these composites in 35-mm tissue-culture plates we added 2 ml of fresh medium for subculture and the cultures were left to stand overnight at 37°C in 5% carbon dioxide atmosphere. During this static overnight culture, the cultured cells in the scaffold lay in uniform arrays in the palisades. The next day, the composites of adherent bone marrow cells with the three-dimensional PLGA scaffold were implanted into osteochondral defects in rabbit knee joints.
Implantation
Under general anesthesia as described above, an anteromedial arthrotomy was performed in one knee with the joint flexed maximally. The patella was dislocated laterally and the surface of the femoropatellar groove was exposed. A full-thickness cylindrical cartilage defect (5 mm in diameter, 1.5 mm deep) was created in the patellar groove of the knee using a chisel and a disposable stainless-steel punch. After washing the knee with saline solution and drying with a swab to remove any debris, in some rabbits the defect in one knee was covered with a cell/PLGA scaffold, with the surface bearing the micropores facing the subchondral bone; this was the tissue-engineered-cartilage group (group T; n = 14). In a control group (group C; n = 6), defects were washed with saline solution and dried in the same way but were left without any further treatment. Finally, fibrin sealant (Tisseel®; Baxter AG, Vienna, Austria) was applied between the scaffold and the edge of the defect in group T and to the edge of the defect in group C. The wound was then closed in layers with 2-0 vicryl sutures.
The rabbits were returned to their cages and allowed to move freely without joint immobilization. The rabbits were humanely killed with an overdose of phenobarbital sodium salt at 4 and 12 weeks in group T (groups T-4 (n = 8) and T-12 (n = 6), respectively) and at 12 weeks in group C (n = 6). All the knee joints were opened and the cartilage surfaces were observed with the naked eye and photographed. The knee joint was dissected free from all the soft tissues and the tibia was removed. The distal femur was cut proximal to the patellofemoral joint and cartilage samples were taken. All the animals were operated on in accordance with the guidelines for animal experiments of the Nara Medical University Ethics Committee.
Ultrasound measurements
The ultrasonic evaluation method has been described previously (Fig. 2) [6,7,13]. Briefly, the examination was carried out in saline solution, using a transducer and pulser receiver (Panametrics Japan, Tokyo, Japan). The transducer (3 mm in diameter, 3 mm long) sent and received a flat ultrasonic wave of 10 MHz center frequency. The reflex echogram from the cartilage was transformed into a wavelet map using wavelet transformation. The wavelet transformation (W(a,b)) of the reflex echogram (f(t)) is expressed by:
where Ψ(t) is the mother wavelet function.
For the mother wavelet function, Gabor's function was selected. As a quantitative index of the wavelet map, the maximum magnitude was selected. This index was calculated automatically with a personal computer. The results obtained for the ultrasonic evaluation were the averages of five measurements. For the cartilage defect area, the measurement points were the center and four points at 1 mm above, below, left, and right of the center. The percentage maximum magnitude (the maximum magnitude of the measurement area of the operated knee divided by that of the intact cartilage of the opposite, nonoperated knee; %MM) was used as a quantitative index of the cartilage regeneration.
Histological analysis
After ultrasonic evaluation, each cartilage sample was divided in two along a sagittal plane using a diamond band saw (EXAKT BS300CL; Meiwa, Tokyo, Japan). One part was used for histological analysis and the other for biochemical analysis. Histological samples were fixed in 10% formalin, decalcified in EDTA, and embedded in paraffin. Sagittal sections (5 μm thick) were prepared from the center of the defect area and stained with hematoxylin and eosin, alcian blue, and safranin-O–fast green. Sections stained with safranin-O–fast green were scored by an orthopedic surgeon under blinded conditions according to the semiquantitative histologic grading scale composed of six categories described by Caplan and colleagues [14] and were assigned a score ranging from 0 to 16 points. A high total score represented good cartilage regeneration.
Biochemical analysis
The chondroitin 4-sulfate, chondroitin 6-sulfate, and dermatan sulfate contents were evaluated to quantify the proteoglycan content using high-performance liquid chromatography analysis [15]. The hydroxyproline content was evaluated to quantify the collagen content [16].
Statistic analysis
All data in this study are reported as means ± standard deviations. Differences were analyzed using the nonparametric Mann–Whitney U test. Pearson correlations were performed to determine the associations between the ultrasonic data and the histological data. The significance level was set at P < 0.05.
Results
Ultrasonic analysis
The %MM values were 29.8 ± 15.1% in group C, 38.8 ± 16.9% in group T-4, and 76.5 ± 18.7% in group T-12 (Fig. 3). Significant differences in the %MM were seen between C and T-12 (P = 0.007) and between T-4 and T-12 (P = 0.007). The average %MM in group T-4 was higher than that in group C, but the difference was not significant (P = 0.32).
Histological findings
In the histological findings, the defect area in group C was filled with fibrous tissue. None of the defects from group C contained any fibrocartilage or hyaline-like cartilage (Fig. 4a). The defect area in group T-4 was filled with scattered cartilage-like tissue in the scaffold. Chondroid cells with round nuclei were observed in an extracellular matrix showing normal or nearly normal safranin-O staining (Fig. 4b). The defect area in group T-12 was filled with hyaline-like cartilage, and chondroid cells lay in a uniform array in the palisades (Fig. 4c). The semiquantitative histologic grading scale scores were 4.17 ± 1.72 for group C, 4.88 ± 1.81 for group T-4, and 10.7 ± 0.82 for group T-12 (Fig. 5). Significant differences in the scores were found between C and T-12 (P = 0.004), and between T-4 and T-12 (P = 0.003). There was a correlation between the %MM from the ultrasonic examinations and the semiquantitative histologic grading scale scores for the overall results of all the measurements (R2 = 0.66) (Fig. 6). The histological scoring showed a strong similarity to the results of the %MM values.
Biochemical analyses
The mean chondroitin sulfate contents were 22.9 nmol/mg in group C, 59.9 nmol/mg in group T-4, and 112.1 nmol/mg in group T-12 (Fig. 7). Significant differences in the chondroitin sulfate contents were found between group T-12 and group C (P = 0.006). The mean hydroxyproline contents were 28.5 μg/mg in group C, 25.0 μg/mg in group T-4, and 26.6 μg/mg in group T-12. There were no significant differences among the three groups. In the biochemical findings, the chondroitin sulfate content showed a similarity to the results of the %MM values.
Discussion
In this study, ultrasonic measurements were found to predict the process of cartilage regeneration using tissue-engineered cartilage as a minimally invasive method. The main finding of the study is that the ultrasonic results were able to judge cartilage regeneration on the basis of objective data such as the %MM, since all the hyaline-like cartilage had a high %MM and the %MM increased with increasing cartilage regeneration. Therefore, ultrasound could be used to examine the microstructure of tissue-engineered cartilage using cell/scaffold complexes and investigate the length of time required for stem cells in a scaffold to differentiate into hyaline cartilage without a biopsy.
A three-dimensional porous scaffold is thought to be necessary for cartilage tissue engineering, in order to prevent the seeded cells from diffusing out of the defect site and to provide the cells with an optimal environment for cartilage differentiation [17-20]. Almost all of the scaffolds investigated have been fabricated using biodegradable polymers that have received approval for use from the US Food and Drug Administration. These polymers are favorable for the synthesis and secretion of a cartilaginous matrix, such as proteoglycans and type II collagen, and act as a physical and mechanical support for the seeded cells and their developing matrix until the polymer is remodeled by the host tissue [21]. Therefore, the clinical application of cell/scaffold complexes for cartilage regeneration is anticipated.
There are numerous clinical methods of grading regenerated cartilage at the time of surgery or arthroscopy by direct observation of the cartilage surface [22-24]. However, accurate evaluation of cartilage regeneration from cell/scaffold complexes is difficult by macroscopic observation alone. In addition, it is well established that probing cannot evaluate the cartilage condition quantitatively. As a quantitative method that could replace probing, attempts have been made to evaluate cartilage using MRI, but such in situ evaluation has been performed only in experimental trials [25-27]. Cartilage biopsy and histological examination have been performed to evaluate articular cartilage clinically. However, the histological score is defined by the subjectivity of the examiner, and it is still difficult to measure the degree of cartilage regeneration nondestructively. Therefore, ultrasonic evaluation using a wavelet map will be useful for supporting the evaluation of tissue-engineered cartilage using cell/scaffold complexes.
Recently, high-frequency ultrasonography was used to assess cartilage degeneration quantitatively. Chérin and colleagues [28] revealed a relation between quantitative ultrasound and maturation-related changes in rat cartilage. Jaffré and colleagues [29] reported that quantitative 55 MHz ultrasound allowed detection of early cartilage lesions due to experimental arthritis and could also detect the effects of anti-inflammatory drugs. Therefore, high-frequency ultrasonography could be useful for investigating structural changes in the cartilage matrix and evaluating the efficacy of specific therapeutic agents. However, no studies have focused on assessing tissue-engineered cartilage using high-frequency ultrasonography. In our previous work, we found that ultrasound assessment using wavelet transformation could predict the histological findings of tissue-engineered cartilage [8,9]. Using the same method, Kuroki and colleagues successfully assessed the cartilage condition of osteochondral plugs when articular cartilage defects were treated with an autologous osteochondral graft [30]. Moreover, this method has been used to assess living human cartilage under arthroscopy [7]. Therefore, ultrasound assessment using wavelet transformation should contribute to novel therapies for cartilage regeneration.
Although, the %MM was used as a quantitative index of the regenerated cartilage, what the %MM is closely related to remains unknown. Töyräs and colleagues [31] reported that ultrasound reflection could detect structural changes in the superficial collagen network and that tangential collagen fibrils act as ultrasound reflectors at the cartilage surface. Pellaumail and colleagues [32] stated that changes in high-frequency ultrasound back scatter were related to changes in the extracellular matrix collagen and most likely in its fibrillar network organization. However, these observations apparently contradict our results that the collagen content did not differ between the three groups. One explanation for this inconsistency could be differences between the reflex echoes from flat ultrasound and focal ultrasound. Another explanation could be differences in the ultrasonic frequency level (10 MHz vs 20 to 55 MHz). From an acoustic point of view, differences in the surface reflection indicate significant alterations in the acoustic impedance among regenerated cartilage samples. Therefore, the extracellular matrix, which includes not only collagen but also proteoglycans and water in the intrafibrillar space and molecular pore spaces of the extracellular matrix as hydrophilic proteoglycan aggregates, should be related to the %MM. The %MM reveals the microstructural changes in regenerated cartilage and can provide diagnostically important information about the regenerated cartilage.
Two limitations of our study should be considered. First, the maximum magnitude in our evaluation system could detect microstructural changes in a layer to a depth of 500 μm [13]. Therefore, the maximum magnitude could only evaluate the surface layer in human cartilage. However, it is of great significance to evaluate the surface layer of tissue-engineered cartilage, since this layer plays an important role in the biomechanical function of the joint. Therefore, ultrasound represents a sensitive tool for detecting regeneration of the cartilage surface in tissue engineering. Further studies using low-frequency ultrasound may provide a better assessment of the deeper layers in tissue-engineered cartilage. Second, we did not detect cartilage regeneration in living humans. However, we have previously reported relevant clinical acoustic data from human cartilage in situ under arthroscopy [7]. Therefore, further studies are needed to determine whether this evaluation system will prove beneficial for tissue-engineered cartilage using cell/scaffold complexes.
Conclusion
This study reports the first results regarding the relation between quantitative ultrasound and the regeneration process of tissue-engineered cartilage. Ultrasonic evaluation using a wavelet map can support the evaluation of tissue-engineered cartilage using cell/scaffold complexes. Ultrasonic assessment using a wavelet map may contribute to the progress of tissue engineering in the musculoskeletal field, and the %MM obtained from this ultrasonic assessment can be expected to become one of the quantitative indexes of cartilage regeneration therapy.
Abbreviations
%MM = maximum magnitude of the measurement area divided by that of the intact cartilage of the nonoperated knee; PLGA = poly(lactic-glycolic acid).
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
KH conceived the study, participated in its design, and performed all the experiments. YT and HO participated in the design of the animal study. TH, KU, and JY performed the animal study. KY, TF, and MS fabricated the three-dimensional PLGA scaffold and performed the cell culture. KI participated in the design of the ultrasound analysis and performed the ultrasonic assessment. All authors read and approved the final manuscript.
Acknowledgements
We appreciate the advice and expertise of Dr Koji Mori and Dr Yusuke Morita. We are indebted to Kaneka Corporation for their generous donation of the three-dimensional PLGA scaffolds. We thank Kyoto University and Nara Medical University for financial support. There were no other funding sources for this study. The study sponsors had no role in the study design, data collection, data analysis, or data interpretation, or in the writing of the report.
Figures and Tables
Figure 1 The three-dimensional poly(lactic-glycolic acid) (3D-PLGA) scaffold. The micropore side (cell seeding side) (a) and a cross section (b) of the scaffold. Schematic illustration of cell seeding (left) and scanning electron photomicrograph of cross section of cells seeded in the 3D-PLGA scaffold (right) (c). The cells lie in a uniform array at the palisades, similar to hyaline cartilage. Gross appearance of a cartilage defect in the patella groove implanted with a complex between adherent bone marrow cells and 3D-PLGA scaffold (d). The arrows indicate cell/PLGA scaffold.
Figure 2 Schematic illustration of articular cartilage analysis and measurement methods of cartilage samples used in [13]. A reflex echo of articular cartilage (upper) and a wavelet map (lower) are shown on the right. The maximum magnitude is indicated by the gray scale and the percentage maximum magnitude (the maximum magnitude of the measurement area divided by that of the intact cartilage of the nonoperated knee; %MM) is used as a quantitative index of cartilage regeneration.
Figure 3 Bar graph representing ultrasonic findings in rabbits with cartilage defects treated with cell/scaffold implants. Group C, control defect group; Group T-4, tissue-engineered-cartilage group at 4 weeks after implantation; Group T-12, tissue-engineered-cartilage group at 12 weeks after implantation. The error bars represent the standard deviation of each group. The percentage maximum magnitude is expressed as the maximum magnitude of the measurement area in the operated knee, divided by that of the intact cartilage of the opposite, nonoperated knee. *P < 0.05 on the nonparametric Mann–Whitney U test.
Figure 4 Photomicrographs of cartilage defect lesions in rabbits. (a) Group with control (untreated) defects (group C); and groups given tissue-engineered cartilage implants at (b) 4 weeks after implantation (group T-4) and (c) 12 weeks after implantation (group T-12). Safranin-O–fast-green staining; original magnification × 2.5.
Figure 5 Bar graph representing semiquantitative histologic gradings for the three groups of rabbits with cartilage defects. Group with control (untreated) defects (group C); and groups given tissue-engineered cartilage implants at 4 weeks after implantation (group T-4) and 12 weeks after implantation (group T-12). Error bars represent standard deviations. *P < 0.05 on the nonparametric Mann–Whitney U test.
Figure 6 Scatter plot of ultrasound findings in rabbits with cartilage defects treated with cell/scaffold implants. The percentage maximum magnitude is expressed as the maximum magnitude of the measurement area in the operated knee, divided by that of the intact cartilage of the opposite, nonoperated knee.
Figure 7 Bar graphs representing cartilage constituents in rabbits with cartilage defects given cell/scaffold implants. Group with control (untreated) defects (group C); and groups given tissue-engineered cartilage implants at 4 weeks after implantation (group T-4) and 12 weeks after implantation (group T-12). Error bars represent standard deviations. *P < 0.05 on the nonparametric Mann–Whitney U test.
==== Refs
Mankin HJ Maw VC Buckwalter JA Buckwalter JA, Einhorn TA, Simon SR Articular cartilage repair and osteoarthritis Orthopaedic Basic Science 2000 2 Rosemont, IL: American Academy of Orthopaedic Surgeons 472 488
Dewire P Einhorn TA Moskowitz RW, Howell DS, Altman RD, Buckwalter JA, Goldberg VM The joint as an organ Osteoarthritis 2001 3 Philadelphia: WB Saunders 49 68
Brittberg M Lindahl A Nilsson A Ohlsson C Isaksson O Peterson L Treatment of deep cartilage defects in the knee with autologous chondrocyte transplantation N Engl J Med 1994 331 889 895 8078550 10.1056/NEJM199410063311401
Ochi M Uchio Y Kawasaki K Wakitani S Iwasa J Transplantation of cartilage-like tissue made by tissue engineering in the treatment of cartilage defects of the knee J Bone Joint Surgr 2002 84 571 578 10.1302/0301-620X.84B4.11947
Shieh SJ Terada S Vacanti JP Tissue engineering auricular reconstruction: in vitro and in vivo studies Biomaterials 2004 25 1545 1557 14697857 10.1016/S0142-9612(03)00501-5
Hattori K Mori K Habata T Takakura Y Ikeuch K Measurement of the mechanical condition of articular cartilage with an ultrasonic probe: quantitative evaluation using wavelet transformation Clin Biomech (Bristol, Avon) 2003 18 553 557 12828905 10.1016/S0268-0033(03)00048-2
Hattori K Takakura Y Ishimura M Habata T Uematsu K Ikeuch K Quantitative arthroscopic ultrasound evaluation of living human cartilage Clin Biomech (Bristol, Avon) 2004 19 213 216 14967587 10.1016/j.clinbiomech.2003.11.005
Hattori K Takakura Y Morita Y Takenaka M Uematsu K Ikeuchi K Can ultrasound predict histological findings in regenerated cartilage? Rheumatology 2004 43 302 305 14585924 10.1093/rheumatology/keh036
Hattori K Takakura Y Ohgushi H Habata T Uematsu K Takenaka M Ikeuchi K Which cartilage is regenerated, hyaline cartilage or fibrocartilage? Non-invasive ultrasonic evaluation of tissue-engineered cartilage Rheumatology 2004 43 1106 1108 15199220 10.1093/rheumatology/keh256
Uematsu K Hattori K Ishimoto Y Yamauchi J Habata T Takakura Y Ohgushi H Fukuchi T Satoh M Cartilage repair with a specially three-dimensional poly-lactic-glycolic acid scaffold [abstract] Trans Orthop Research Soc 2004 29 676
Kawanishi M Ushida T Kaneko T Niwa H Fukubayashi T Nakamura K Oda H Tanaka S Tateishi T New type of biodegradable porous scaffolds for tissue-engineered articular cartilage Mater Sci Eng C 2004 24 431 435 10.1016/j.msec.2003.11.008
Uematsu K Hattori K Ishimoto Y Yamauchi J Habata T Takakura Y Ohgushi H Fukuchi T Sato M Cartilage regeneration using mesenchymal stem cells and a three-dimensional poly-lactic-glycolic acid (PLGA) scaffold Biomaterials 2005 26 4273 4279 15683651 10.1016/j.biomaterials.2004.10.037
Hattori K Ikeuchi K Morita Y Takakura Y Quantitative ultrasonic assessment for detecting microscopic cartilage damage in osteoarthritis Arthritis Res Ther 2005 7 R38 46 15642141 10.1186/ar1463
Caplan AI Elyaderani M Mochizuki Y Wakitani S Goldberg VM Principles of cartilage repair and regeneration Clin Orthop 1997 342 254 269 9308548 10.1097/00003086-199709000-00033
Shinmei M Miyauchi S Machida A Miyazaki K Quantification of chondroitin 4-sulfate and chondroitin 6-sulfate in pathologic joint fluid Arthritis Rheum 1992 35 1304 1308 1445446
Woessner J The determination of hydroxyproline in tissue and protein samples containing small proportions of this imino acid Arch Biochem Biophys 1961 93 440 447 13786180 10.1016/0003-9861(61)90291-0
Chen G Sato T Ushida T Ochiai N Tateishi T Tissue engineering of cartilage using a hybrid scaffold of synthetic polymer and collagen Tissue Eng 2004 10 323 330 15165449 10.1089/107632704323061681
Woodfield TB Malda J de Wijn J Peters F Riesle J van Blitterswijk CA Design of porous scaffolds for cartilage tissue engineering using a three-dimensional fiber-deposition technique Biomaterials 2004 25 4149 4161 15046905 10.1016/j.biomaterials.2003.10.056
Peter SJ Miller MJ Yasko AW Yaszemski MJ Mikos AG Polymer concepts in tissue engineering J Biomed Mater Res 1998 43 422 427 9855200 10.1002/(SICI)1097-4636(199824)43:4<422::AID-JBM9>3.0.CO;2-1
Isogai N Landis W Kim TH Gerstenfeld LC Upton J Vacanti JP Formation of phalanges and small joints by tissue-engineering J Bone Joint Surg Am 1999 81 306 316 10199268 10.1302/0301-620X.81B2.9356
Isogai N Asamura S Higashi T Ikada Y Morita S Hillyer J Jacquet R Landis WJ Tissue engineering of an auricular cartilage model utilizing cultured chondrocyte-poly(L-lactide-epsilon-caprolactone) scaffolds Tissue Eng 2004 10 673 687 15265285 10.1089/1076327041348527
Outerbridge RE The etiology of chondromalacia patellae J Bone Joint Surg 1961 43B 752 757
Dougados M Ayral X Listrat V Gueguen A Bahuaud J Beaufils P Beguin JA Bonvarlet JP Boyer T Coudane H The SFA system for assessing articular cartilage lesions at arthroscopy of the knee Arthroscopy 1994 10 69 77 8166905
Brismar BH Wredmark T Movin T Leandersson J Svensson O Observation reliability in the arthroscopic classification of osteoarthritis of the knee J Bone Joint Surg 2002 84B 42 47 10.1302/0301-620X.84B1.11660
Nissi MJ Toyras J Laasanen MS Rieppo J Saarakkala S Lappalainen R Jurvelin JS Nieminen MT Proteoglycan and collagen sensitive MRI evaluation of normal and degenerated articular cartilage J Orthop Res 2004 22 557 564 15099635 10.1016/j.orthres.2003.09.008
Graichen H von Eisenhart-Rothe R Vogl T Englmeier KH Eckstein F Quantitative assessment of cartilage status in osteoarthritis by quantitative magnetic resonance imaging: technical validation for use in analysis of cartilage volume and further morphologic parameters Arthritis Rheum 2004 50 811 816 15022323 10.1002/art.20191
Watrin-Pinzano A Ruaud JP Cheli Y Gonord P Grossin L Gillet P Blum A Payan E Olivier P Guillot G T2 mapping: an efficient MR quantitative technique to evaluate spontaneous cartilage repair in rat patella Osteoarthritis Cartilage 2004 12 191 200 14972336 10.1016/j.joca.2003.10.010
Chérin E Saïed A Pellaumail B Loeuille D Laugier P Gillet P Netter P Berger G Assessment of rat articular cartilage maturation using 50-MHz quantitative ultrasonography Osteoarthritis Cartilage 2001 9 178 186 11237665 10.1053/joca.2000.0374
Jaffré B Watrin A Loeuille D Gillet P Netter P Laugier P Saïed A Effects of antiinflammatory drugs on arthritic cartilage: a high-frequency quantitative ultrasound study in rats Arthritis Rheum 2003 48 1594 1601 12794827 10.1002/art.11023
Kuroki H Nakagawa Y Mori K Ohba M Suzuki T Mizuno Y Ando K Takenaka M Ikeuchi K Nakamura T Acoustic stiffness and change in plug cartilage over time after autologous osteochondral grafting: correlation between ultrasound signal intensity and histological score in a rabbit model Arthritis Res Ther 2004 6 R492 504 15535827 10.1186/ar1219
Töyräs J Rieppo J Nieminen MT Helminen HJ Jurvelin JS Characterization of enzymatically induced degradation of articular cartilage using high frequency ultrasound Phys Med Biol 1999 44 2723 2733 10588280 10.1088/0031-9155/44/11/303
Pellaumail B Watrin A Loeuille D Netter P Berger G Laugier P Saïed A Effect of articular cartilage proteoglycan depletion on high frequency ultrasound backscatter Osteoarthritis Cartilage 2002 10 535 541 12127833 10.1053/joca.2002.0790
| 15899042 | PMC1174952 | CC BY | 2021-01-04 16:02:36 | no | Arthritis Res Ther. 2005 Mar 1; 7(3):R552-R559 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1710 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17151589904110.1186/ar1715Research ArticleSystemic lupus erythematosus induced by anti-tumour necrosis factor alpha therapy: a French national survey De Bandt Michel [email protected] Jean 2Le Loët Xavier 3Prouzeau Sebastian 4Fautrel Bruno 5Marcelli Christian 6Boucquillard Eric 7Siame Jean Louis 8Mariette Xavier 9the Club Rhumatismes et Inflammation1 Rheumatology Department, Hôpital Robert Ballanger, Aulnay sous Bois, France2 CHU Hautepierre, Strasbourg, France3 Hôpital Bois-Guillaume, Rouen, France4 Hôpital de Saint Lo, Saint Lo, France5 Hôpital Pitié Salpétrière, Assistance Publique-Hôpitaux de Paris, Paris, France6 CHU côte de Nacre, Caen, France7 Hôpital de Saint Pierre de la Réunion, St Pierre France8 Polyclinique de Riaumont, Liévin, France9 Hôpital de Bicêtre, Assistance Publique-Hôpitaux de Paris, Le Kremlin Bicêtre, France2005 1 3 2005 7 3 R545 R551 29 11 2004 29 12 2004 7 2 2005 10 2 2005 Copyright © 2005 De Bandt et al., licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The development of drug-induced lupus remains a matter of concern in patients treated with anti-tumour necrosis factor (TNF) alpha. The incidence of such adverse effects is unknown. We undertook a retrospective national study to analyse such patients.
Between June and October 2003, 866 rheumatology and internal medicine practitioners from all French hospital centres prescribing anti-TNF in rheumatic diseases registered on the website of the 'Club Rhumatismes et Inflammation' were contacted by email to obtain the files of patients with TNF-induced systemic lupus erythematosus. Twenty-two cases were collected, revealing two aspects of these manifestations. Ten patients (six patients receiving infliximab, four patients receiving etanercept) only had anti-DNA antibodies and skin manifestations one could classify as 'limited skin lupus' or 'toxidermia' in a context of autoimmunity, whereas 12 patients (nine patients receiving infliximab, three patients receiving etanercept) had more complete drug-induced lupus with systemic manifestations and at least four American Congress of Rheumatology criteria. One patient had central nervous system manifestations. No patients had lupus nephritis. The signs of lupus occurred within a mean of 9 months (range 3–16 months) in patients treated with infliximab and within a mean of 4 months (range 2–5 months) in patients treated with etanercept. In all cases after diagnosis was determined, anti-TNF was stopped and specific treatment introduced in eight patients: two patients received intravenous methylprednisolone, four patients received oral steroids (15–35 mg/day), and two patients received topical steroids. Lupus manifestations abated within a few weeks (median 8 weeks, standard deviation 3–16) in all patients except one with longer-lasting evolution (6 months). At that time, cautious estimations (unpublished data from Schering Plough Inc. and Wyeth Inc.) indicated that about 7700 patients had been exposed to infliximab and 3000 to etanercept for inflammatory arthritides in France. It thus appears that no drug was more implicated than the other in lupus syndromes, whose incidence was 15/7700 = 0.19% with infliximab and 7/3800 = 0.18% with etanercept.
Clinicians should be aware that lupus syndromes with systemic manifestations may occur in patients under anti-TNF alpha treatment.
==== Body
Introduction
Therapy with anti-tumour necrosis factor (TNF) alpha is effective for rheumatoid arthritis (RA) [1,2], with an estimated 500,000 patients being treated worldwide. The possible occurrence of drug-induced autoimmune disorders remains a matter of concern [3]because induction of autoantibodies is frequently observed in patients treated with TNF alpha inhibitors [4]. Of concern is the possible induction of lupus-like (or drug-induced lupus) syndromes, but few cases have been reported [5-7]. In all reported cases, the signs disappeared after treatment was stopped. The incidence of cases is unknown.
We report here the results of a French national survey revealing 22 cases of drug-induced lupus erythematosus (systemic lupus erythematosus [SLE]) in French patients being treated with anti-TNF alpha for inflammatory arthritides.
Methods
Between June and October 2003, the 'Club Rhumatismes et Inflammation', a section of the French Society of Rheumatology, carried out a retrospective survey among all French rheumatologists and specialists in internal medicine to uncover cases of SLE with anti-TNF alpha treatment (infliximab or etanercept at that time). Eight hundred and sixty-six rheumatology and internal medicine practitioners from all the French hospital centres prescribing anti-TNF in rheumatic diseases, registered on the website of the Club Rhumatismes et Inflammation , were contacted four times by email at 1-month intervals to obtain the files of patients with TNF-induced SLE. The study included all the patients ever known to have developed an SLE-like illness during anti-TNF treatment and not only those who developed an SLE-like illness during the 3-month study period.
As the prescription for anti-TNF alpha is limited to the hospitals in France, all the units of rheumatology using biologics were contacted. Eighteen units gave positive results, 22 gave negative results and very few (<10) did not participate. As all the units of rheumatology using biologics were contacted and most of them participated in the study, we can estimate that the survey involved almost all of the French patients treated with anti-TNF for arthritides. At that time, cautious estimations indicated that about 7700 patients had been exposed to infliximab and 3800 had been exposed to etanercept for inflammatory arthritides in France (unpublished data from Schering Plough Inc. and Wyeth Inc.).
As there are no recognized criteria for drug-induced lupus [8], we considered diagnosis in the case of: a patient with anti-TNF alpha treatment for inflammatory arthritides; a temporal relationship between clinical manifestations and anti-TNF alpha treatment; the presence of at least four American Congress of Rheumatology (ACR) criteria of SLE [9]. Musculoskeletal symptoms were taken into account only if they reappeared with other lupus symptoms in a patient in whom they had previously disappeared under anti-TNF therapy, and isolated positive results for antinuclear antibodies (ANA) or anti-dsDNA antibodies were not considered for diagnosis, given their high frequency in patients under this therapy. Telephone calls were made to collect information in the case of missing data. Physicians were asked to provide information about the clinical status of the patients and the presence of lupus criteria. Information about the immunological status of the patients was requested (before and after the onset of the manifestations as well as after drug discontinuation).
The biological tests used for the detection of autoantibodies were an indirect immunofluorescent assay for ANA, an ELISA or Farr assay for anti-DNA antibodies, Ouchterlony's method for anti-extractable nuclear antigens (anti-ENA), and an ELISA for anti-histone, anti-Ro, anti-La, anti-SM, anti-RNP, anti-JO1, anti-Topo 1, and anticardiolipin antibodies (ACL).
Results
A total of 32 patients was collected, three of which had been previously described [5]. Ten patients were ruled out due to improper diagnosis of lupus syndrome, due to pre-existing lupus syndrome or due to mixed connective tissue disease before introduction of anti-TNF alpha therapies. We observed two types of manifestations among the remaining patients.
Ten patients (six patients treated with infliximab, four patients treated with etanercept) had a diagnosis of 'anti-TNF-induced SLE' based on three ACR criteria (Table 1). None of these patients had previous signs of lupus before treatment except one with isolated positive ANA. All of them had RA with joint erosions. The mean age at onset of RA was 39 years (range 24–57 years), and the mean disease duration before onset of 'SLE' was 13 years (range 6–31 years). All patients had been treated with a mean of five disease-modifying antirheumatic drugs, including methotrexate in all cases. Before anti-TNF therapy, no patients had clinical sign of lupus, one had positive isolated ANA (1/160) without any other lupus criteria, and no patients had anti-DNA or low complement. At the time of treatment, all patients were being treated with steroids (mean 8 mg/day, range 4–16 mg/day) and methotrexate. No patient had any other drug known as a lupus-inducing drug.
The only signs were isolated skin lesions (Table 2): pruritic rash (two cases), butterfly rash (three cases), photosensitivity (two cases), purpura (two cases), chilblains (one case), in a context of autoimmunity with positive ANA and anti-dsDNA antibodies. In all cases the clinical manifestations led to ceasing the treatment with anti-TNF alpha, and the signs abated quickly thereafter (<1 month). Despite the presence of three manifestations or criteria for systemic lupus we did not consider that these patients had drug-induced SLE, but rather that they presented toxidermia associated with ANA. Moreover, all of these clinical manifestations are not specific for lupus. Unfortunately no skin biopsy was performed.
Biological signs were positive results for ANA in all patients (new onset or rise in titre, range 1/160–1/250°; three patients with a speckled pattern, seven patients with a diffuse pattern). A new onset of positive results for anti-dsDNA antibodies in the 10 patients was noted by ELISA. None had any other biological and/or immunological manifestations of lupus. No confounding agent was involved in the occurrence of ANA and anti-dsDNA.
Twelve other patients (10 females, two males; nine patients receiving infliximab, three patients receiving etanercept) had a diagnosis of drug-induced systemic lupus supported by the presence of at least four ACR criteria (Tables 2 and 3). Eleven patients had erosive and destructive RA and one patient had severe psoriatic arthritis. The mean age at onset of RA was 36 years (range 14–54 years), and the mean disease duration before onset of SLE was 16 years (range 3–40 years). All patients had been treated with a mean of five disease-modifying antirheumatic drugs (range 2–8), including methotrexate in all cases.
Before anti-TNF therapy, no patients had clinical sign of lupus, three had positive ANA (range 1/160–1/1280), one of these (the patient with the highest level of ANA) had one time a limit positive anti-dsDNA titre (ELISA test, 46 UI; normal value <40), and nine had negative results. The two other patients with positive ANA had positive anti-Ro antibodies, and a clinical history of secondary Sjögren syndrome. None of the three patients with positive ANA had any other sign or lupus criteria (Table 2). Eleven patients had a typical history of severe and erosive RA and one patient a history of severe psoriatic arthritis. At the time of treatment, all patients were being treated with steroids (mean 9 mg/day; range 5–15 mg/day) and methotrexate (except one patient on etanercept alone).
Clinical signs were skin manifestations in 11 patients (papules, alopecia, rash, butterfly rash, photosensitivity), general manifestations in nine patients (fever, weight loss, asthenia), reappearance of polyarthritis in six patients, inflammatory myalgias in four patients, serositis in three patients, deep vein thrombosis (twice) in one patient, lung disease (life-threatening pneumonitis) with pleuritis in one patient, and neuritis of the third cranial nerve in one patient. No case of nephritis was found. The mean ACR criteria number was 5.5 (range 4–7).
Skin lesions were largely symmetrical (arm, face, trunk) and not at injection sites (in the case of etanercept). Histological analysis (four patients) revealed atrophy of the epidermis, necrosis of some keratinocytes, and perifollicular and perivascular lymphocytic infiltration in the dermis without vasculitis. No indirect immunofluorescent test was performed. The patient with deep vein thrombosis also had positive ACL antibodies. Articular symptoms were taken into account only if they reappeared with other lupus symptoms in a patient in whom they had previously disappeared under anti-TNF therapy and/or were different from previous complaints.
The patient with optical neuritis had no previous sign of multiple sclerosis before treatment with infliximab; she had an isolated neuritis of the third cranial nerve, with malar rash and autoantibodies. The lumbar puncture was normal. Magnetic resonance imaging showed an isolated hyper signal of the third cranial nerve. Extensive research for other manifestations of multiple sclerosis was performed without success. The presence of the neurological manifestations with other clinical signs and autoimmunity led to the diagnosis of drug-induced lupus.
Biological signs were positive results for ANA in all patients (new onset or rise in titre, range 1/160–1/2560°; four with a speckled pattern, eight with a diffuse pattern), and positive results for anti-dsDNA antibodies (new onset) in 11 patients by ELISA. Among the 11 patients tested with ELISA: five had anti-IgM antibodies and six had a positive test with no more detail; among them, three patients were tested by Farr assay and were positive. The patient without anti-DNA had a high titre of ANA, positive anti-ENA and anti-histone antibodies. Positive anti-ENA antibodies were present in five patients (two patients with previously known anti-SS-A/Ro antibodies, three patients with newly detected anti-ENA antibodies with an unidentified aspect), anti-histone in two patients and anti-cardiolipin in six patients. Leucopenia (blood count <4000/mm3), thrombopenia (blood count <100,000/mm3), lymphopenia (blood count <1500/mm3) and positive Coombs test (without haemolytic anaemia) were present in five patients, four patients, two patients and one patient, respectively. Elevated muscle enzymes were present in three of four patients with inflammatory myalgias. One patient had isolated elevated creatinin phosphokinase. No patient had muscle weakness. Transient low levels of C4 were detected in four cases (nine tested).
The signs of SLE occurred in a mean of 9 months in patients treated with infliximab and 4 months in patients treated with etanercept. In all cases, after the diagnosis was determined, the treatment was stopped and the manifestations then abated within a few weeks (median 8 weeks, range 3–16 weeks), except one (patient 12, Table 2) with a longer lasting evolution (6 months) before resolution. She had persistent asthenia, immunological and haematological abnormalities before resolution. But after 6 months all the signs abated. Biological signs normalized within a few months: in eight patients the ANA results were negative, and in four they were reduced; in nine patients, anti-dsDNA results were negative, and in three patients they were reduced.
Recovery was spontaneous without treatment in four cases. Steroids were needed in the eight other patients: two patients received topical steroids for skin lesions, two patients received intravenous methylprednisolone, and four patients received oral steroids (15–30 mg/day) for intense general signs. In no patient did SLE signs reappear.
Cautious estimations at that time (unpublished data from Schering Plough Inc. and Wyeth Inc.) indicated that about 7700 patients had been exposed to infliximab and 3800 patients had been exposed to etanercept for inflammatory arthritides in France. The incidence of lupus syndromes was thus the same with infliximab (15/7700 = 0.19%) and with etanercept (7/3800 = 0.18%).
Discussion
We report on 22 patients treated with anti-TNF alpha for severe RA or psoriatic arthritis (15 patients receiving infliximab and seven patients receiving etanercept) with no previous sign of lupus disease who developed clinical and biological manifestations of drug-induced lupus.
We are aware that the scientific interest of a retrospective analysis is of limited value compared with a prospective study. However, at that time only isolated case reports were available. As far as we know, this survey is the only one providing further information about the clinical problem of drug-induced lupus. We hope that the national observatories and registers that have been settled in various countries around the world will precisely and prospectively answer the question of anti-TNF-induced lupus.
Analysis of the cases revealed two subgroups of patients. The first group of patients was considered by the referring physician as 'drug-induced lupus'. In our opinion these patients had what we term 'toxidermia' – that is, isolated skin manifestations in a context of autoimmunity and the absence of systemic manifestations. We are aware that some colleagues will feel uneasy with the term 'toxidermia' and would prefer to qualify these patients as 'incomplete lupus with isolated skin manifestations'. We do understand the reserve about the expression 'toxidermia' rather than 'drug-induced lupus erythematosus'. We preferred to be rigorous with the diagnosis of SLE and to use a more stringent definition (at least four ACR criteria for SLE) to describe a core of patients. Indeed, patients treated with anti-TNF (mostly infliximab rather than etanercept) have frequent and isolated skin manifestations with positive autoantibodies. The frequency of these clinical pictures is unknown but seems important regarding the frequency of autoantibodies (up to 50% for ANA, 25% for ACL and 15% for anti-DNA with infliximab) and of skin manifestations [3-7,10-12]. Do all these patients only have toxidermia in a context of autoimmunity or 'limited drug-induced skin lupus'?
The second group of 12 patients had what we considered true 'drug-induced SLE' with at least four ACR criteria and systemic manifestations, with a very acute and complete syndrome, associating general manifestations and clinical and biological signs of lupus. All 12 patients fulfilled the ACR criteria for SLE [9], and did not simply have drug-induced skin toxidermia in a context of autoimmunity. Clinicians should thus be aware that lupus syndromes may occur in patients under anti-TNF alpha treatment and may be complicated by central nervous system signs. However, withdrawal of the drug leads to an abatement of the signs.
ACL antibodies were detected in six patients whereas only one patient had developed thrombosis. The occurrence of ACL antibodies in anti-TNF alpha-treated patient is well documented [13]: up to 25% of RA patients with anti-TNF develop IgG or IgM ACL, but thrombosis is observed in much fewer patients (about 4%). It is also known that TNF has potent anti-thrombotic properties [14]. It is therefore conceivable that the association of ACL antibodies and inhibition of TNF could lead to an increase number of thrombosis. Should we routinely look for ACL antibodies in our patients?
The imputability of anti-TNF therapy in inducing lupus syndrome is probable, given the temporal relationship between the onset of signs with treatment and the resolution following withdrawal of the drug in all cases. No confounding agent (such as statins) was involved in patients with myositis, myalgias or elevated creatinin phosphokinase. No other drug, known as a lupus-inducing-drug, was present in any patient.
The incidence of anti-TNF-induced lupus is difficult to evaluate. We estimated that we covered most, if not all, cases in France as of October 2003. It is possible we have missed some cases, as in all retrospective studies. However, with the opportunity to use the quite unique organized system that is the Club Rhumatismes et Inflammation website, which encompasses most if not all the physicians interested in biologics and systemic diseases, we think these missing cases are scarce. Moreover, we sent four email recall letters at 1-week intervals to detect the cases. The estimation of the number of exposed patients to the drugs is always difficult, even by the pharmaceutical companies themselves. At the time, cautious estimations from Schering Plough Inc. and Wyeth Inc. allowed one to determine the number of exposed patients to each of the drugs from the beginning of the clinical trials until the time of the study, but did not allow one to determine the length of exposition in terms of the number of patients-years. Therefore, with these estimations, it appears that no drug was more implicated than the other in lupus syndromes.
Interestingly, no case of lupus nephritis was observed in this survey. However, one case of etanercept-associated renal disease has been recently described (active urine sediment, new onset of anti-Ro, anti-Sm and anti-RNP antibodies) but no biopsy has been performed. In that case signs abated shortly after drug discontinuation [15].
The mechanism of induction remains unclear. One hypothesis could be an increase of apoptotic particles and antigens from apoptotic cells. It has been shown that RA patients had no circulating nucleosomes at the steady state and some of them had significantly higher levels of plasma nucleosomes after receiving infliximab [16]. The accumulation of nucleosomes could possibly enhance the development of autoantibodies in subjects with appropriate genetic backgrounds.
Another hypothesis is that the suppression of the T-helper type 1 response by TNF blockers could favour a T-helper type 2 response leading to SLE, but this hypothesis needs to be tested in man. Neutralization of TNF alpha was tested in mice undergoing acute graft versus host disease using the parent-into-F1 model [17]. Monoclonal antibody against TNF alpha blocked the lymphocytopenic features characteristic of acute graft versus host disease and induced a lupus-like chronic graft versus host disease phenotype (lymphoproliferation and autoantibody production). These effects resulted from complete inhibition of detectable anti-host cytotoxic T lymphocytes. In that model the authors showed that in vivo blockade of TNF alpha preferentially inhibited the production of interferon gamma and blocked interferon-gamma-dependent upregulation of Fas; and that cytokines such as IL-10, IL-6, or IL-4 were not inhibited. These results suggest that a therapeutic TNF alpha blockade may promote humoral autoimmunity by selectively inhibiting the induction of a cytotoxic T lymphocyte response that would normally suppress autoreactive B cells.
A final hypothesis is the role of bacterial infections. They are increased with TNF blockers and are also powerful stimulants leading to polyclonal B-lymphocyte activation and autoantibody production. Some cases of positive anti-DNA following infection after etanercept have been reported [18]. Interestingly, the titre returned to normal values after antibiotic treatment.
In conclusion, we collected 22 cases of 'anti-TNF alpha-induced lupus' based on the ACR lupus criteria in a retrospective national survey, allowing us to better define the clinical aspects of these manifestations. Given the frequency of autoantibodies in anti-TNF alpha-treated patients, we proposed to identify two subsets of patients. The first group only had skin manifestations and anti-DNA antibodies. Do these patients have 'toxidermia' in a context of autoimmunity or true 'limited drug-induced skin lupus'? Should they stop or continue treatment with anti-TNF alpha? We do not have the answer, and we let the reader decide. Whereas the second group has true drug-induced SLE (with at least four ACR criteria) and systemic manifestations (serositis, cranial neuritis). In all cases anti-TNF was stopped after diagnosis was determined, and specific manifestations abated within a few weeks. Clinicians should be aware that lupus syndromes with systemic manifestations may occur in patients receiving anti-TNF alpha treatment.
Abbreviations
ACL = anticardiolipin antibodies; ACR = American Congress of Rheumatology; ANA = antinuclear antibodies; ELISA = enzyme-linked immunosorbent assay; ENA = extractable nuclear antigens; IL = interleukin; RA = rheumatoid arthritis; SLE = systemic lupus erythematosus; TNF = tumour necrosis factor.
Competing interests
The author(s) declare that there are no competing interests.
Authors' contributions
MDB had the original idea for the work, collected all the data, analysed the data and wrote the paper. JS, XLL, SP, BF, ChM, EB and JLS were all in charge of the patients, reviewed the charts, and gathered and collected the data. XM participated in the data analysis and the writing of the work.
Acknowledgements
Ethic approval was given for the study by the advisory board of the Club Rhumatismes et Inflammation and the 'Société Française de Rhumatologie'. The following physicians referred cases of patients with only three signs of lupus: F Simon, T Pham, J Dupuis, I Chary-Valckenaere, M Longy-Boursier, T Schaeverbecke, V Goeb, Ch Richet, B Moura and J Bonidan.
Figures and Tables
Table 1 General presentation of the 10 patients with 'limited skin lupus' or toxiderma in a context of autoimmunity
Patient Before anti-TNF alpha treatment During onset of symptoms
Disease Autoantibody Clinical signs of lupus Treatment Duration of treatment (months) Clinical signs of lupus Biological signs of lupus
1 RA, erosive RF+ None ETA 36 Skin ANA+, dsDNA+
2 RA, erosive RF+ None INF 6 Skin ANA+, dsDNA+
3 RA, erosive RF+, ANA+ 1:160° None INF 18 Skin ANA+, dsDNA+
4 RA, erosive RF- None ETA 5 Skin ANA+, dsDNA+
5 RA, erosive RF+ None INF 7 Skin ANA+, dsDNA+
6 RA erosive RF+ None ETA 5 Skin ANA+, dsDNA+
7 RA, erosive RF+ None INF 11 Skin ANA+, dsDNA+
8 RA, erosive RF+ None INF 3 Skin ANA+, dsDNA+
9 RA, erosive RF+ None ETA 12 Skin ANA+, dsDNA+
10 RA, erosive RF+ None INF 13 Skin ANA+, dsDNA+
ANA, antinuclear antibodies; dsDNA, double-strand DNA; ETA, etanercept; INF, infliximab; RA, rheumatoid arthritis; RF, rheumatoid factor; TNF, tumour necrosis factor; +, positive; -, negative.
Table 2 General presentation of the 12 patients with 'complete lupus'
Patient Before anti-TNF alpha treatment During onset of symptoms
Disease Autoantibody Clinical signs of lupus Treatment Duration of treatment (months) Clinical signs of lupusa Biological signs of lupus
1 RA, RF+, erosive None None INF 27 General, skin, serositis, lung ANA+, dsDNA+, ACL+, leucopenia, thrombopenia, ENA+
2 RA, RF+, erosive ANA+, Ro+ None INF 4 General, skin (3), arthritis ANA+, dsDNA+, histone +
3 RA, RF+, erosive ANA+, Ro+ None INF 2 Skin, myalgias, arthritis ANA+, dsDNA+, ENA+,
4 RA, RF-, erosive None None ETA 4 General, skin (2), myositis ANA+, dsDNA+, ACL+, low C4, ENA+
5 RA, RF+, erosive None None INF 4 Skin (3), myositis, arthritis, pericarditis ANA+, dsDNA+, ACL+, CPK, lymphopenia
6 RA, RF+, erosive ANA+, dsDNA+ limit value None ETA 5 General, skin ANA+, dsDNA+, thrombopenia, leucopenia
7 RA, RF+, erosive None None INF 10 General, serositis, myositis ANA+, dsDNA+, ACL+, leucopenia, thrombopenia, CPK,
8 RA, RF+ None None ETA 2 Phlebitis, skin ANA+, dsDNA+, leucopenia, ACL+, thrombopenia
9 Psoriatic arthritis None None INF 14 General, skin, neurological ANA+, dsDNA+
10 RA, RF+, erosive None None INF 16 General, Skin, arthritis ANA+, dsDNA+
11 RA, RF+, erosive None None INF 12 General, skin (3), arthritis, myositis ANA+, dsDNA+, ENA+, CPK
12 RA, RF+, erosive None None INF 10 General, skin, arthritis ANA+, dsDNA+, CPK, ENA+, ACL+, low C4, histone+, leucopenia, lymphopenia, Coombs test+
ACL, positive anticardiolipin antibodies; ANA, antinuclear antibodies; CPK, creatin phospokinase or muscle enyme, elevated muscle enzymes; dsDNA, double-strand DNA; ENA, positive anti-extractable nuclear antigens antibodies; ETA, etanercept; general, general manifestations (fever, weight loss, asthenia); histone, positive anti-histone antibodies; INF, infliximab; RA, rheumatoid arthritis; RF, rheumatoid factor; TNF, tumour necrosis factor; +, positive; -, negative.
aThe skin manifestations were as follows (number in parentheses indicates number of skin manifestations observed in one patient): maculo papular rash observed in six patients, butterfly rash observed in five patients, alopecia present in one patient, photosensitivity observed in five patients, purpuric lesions observed in two patients. One patient had no skin manifestations, seven patients had one sign, one patient had two different signs, and three patients had three skin manifestations.
Table 3 Signs of systemic lupus erythematosus (SLE) in 12 patients under anti-tumour necrosis factor alpha treatment
Sign of SLE Number of patients
Skin (≥ 1 SLE skin criteria) 11
General (fever, weight loss, asthenia) 9
Musculoskeletal (arthritis, myositis, myalgias) 10 (6 arthritis, 3 myositis, 1 myalgia)
Serositis, lung, neuritis, phlebitis; haematological abnormalities, elevated muscle enzymes 3, 1, 1 and 1, respectively; 12 and 3, respectively
Antibody testing
Positive for ANA (>1/160) 12
Positive for anti-dsDNA antibodies (>40 UI) 11
Positive anti-ENA, anti-histone, anti-cardiolipid 5, 2 and 6, respectively
Recovery
After treatment withdrawn (with/without steroids) 12 (4 and 8)
Delay in recovery (median and range) 8 weeks (3–16 weeks)
Ten females, two males; nine patients treated with infliximab and three patients treated with etanercept. ANA, antinuclear antibodies; dsDNA, double-strand DNA; ENA, positive anti-extractable nuclear antigensantibodies.
==== Refs
Garrison L McDonnell N Etanercept: therapeutic use in patients with rheumatoid arthritis Ann Rheum Dis 1999 58 suppl 1 I65 I69 10577976
Maini R St Clair EW Breedveld F Furst D Kalden J Weisman M Smolen J Emery P Harriman G Feldmann M Lipsky P Infliximab (chimeric anti-tumour necrosis factor alpha monoclonal antibody) versus placebo in rheumatoid arthritis patients receiving concomitant methotrexate: a randomised phase III trial. ATTRACT Study Group Lancet 1999 354 1932 1939 10622295 10.1016/S0140-6736(99)05246-0
Pisetski DS Tumor necrosis factor alpha blockers and the induction of anti-DNA antibodies Arthritis Rheum 2000 43 2381 2382 11083257 10.1002/1529-0131(200011)43:11<2381::AID-ANR1>3.0.CO;2-M
Louis M Rauch J Armstrong M Fitzcharles M-A Induction of autoantibodies during prolonged treatment with infliximab J Rheumatol 2003 30 2557 2562 14719194
De Bandt M Vittecoq O Descamps V Le Loet X Meyer O Anti-TNF alpha-induced systemic lupus erythematosus Clin Rheumatol 2003 22 56 61 12605321 10.1007/s10067-002-0654-5
Shakoor N Michalska M Harris CA Block JA Drug-induced systemic lupus erythematosus associated with etanercept therapy Lancet 2002 359 579 580 11867114 10.1016/S0140-6736(02)07714-0
De Rycke L Kruithof E Van Damme N Hoffman IEA Van den Bossche N Van den Bosch F Veys EM De Keyser F Antinuclear antibodies following infliximab treatment in patients with rheumatoid arthritis or spondylarthropathy Arthritis Rheum 2003 48 1015 1023 12687543 10.1002/art.10876
Mongey AB Hess E Lahita R Drug and environmental lupus: clinical manifestations and differences Systemic Lupus Erythematosus 1999 3 New York: Academic Press
Hochberg MC Updating the American College of Rheumatology revised criteria for the classification of systemic lupus erythematosus Arthritis Rheum 1997 40 1725
letter
9324032
Schaible TF Treatment of inflammatory diseases: safety of long-term use of infliximab Presse Med 2001 30 610 613 11317923
Devos SA Van Den Bossche N De Vos M Naeyaert JM Adverse skin reactions to anti-TNF-alpha monoclonal antibody therapy Dermatology 2003 206 388 390 12771494 10.1159/000069965
Flendrie M Creemers MC Welsing PM den Broeder AA van Riel PL Survival during treatment with tumour necrosis factor blocking agents in rheumatoid arthritis Ann Rheum Dis 2003 62 suppl 2 ii30 ii33 14532145
Jonsdottir T Forslid J van Vollenhoven A Harju A Brannemark S Klareskog L van Vollenhoven RF Treatment with tumour necrosis factor alpha antagonists in patients with rheumatoid arthritis induces anticardiolipin antibodies Ann Rheum Dis 2004 63 1075 1078 15066863 10.1136/ard.2003.018093
Cambien B Bergmeier W Saffaripour S Mitchell HA Wagner DD Antithrombotic activity of TNF-alpha J Clin Invest 2003 112 1589 1596 14617760 10.1172/JCI200319284
Carlson E Rothfield N Etanercept-induced like lupus syndrome in a patient with rheumatoid arthritis Arthritis Rheum 2003 48 1165 1166 12687569 10.1002/art.11033
D'Auria F Rovere-Querini P Giazzon M Ajello P Baldissera E Manfredi A Sabbadini M Accumulation of plasma nucleosomes upon treatment with anti-tumour necrosis factor-alpha antibodies J Intern Med 2004 255 409 418 14871466 10.1111/j.1365-2796.2003.01298.x
Via CS Shustov A Rus V Lang T Nguyen P Finkelman FD In vivo neutralization of TNF-alpha promotes humoral autoimmunity by preventing the induction of CTL J Immunol 2001 167 6821 6826 11739498
Ferraccioli G Mecchia F Di Poi E Fabris M Anticardiolipin antibodies in rheumatoid patients treated with etanercept or conventional combination therapy: direct and indirect evidence for a possible association with infections Ann Rheum Dis 2002 61 358 361 11874843 10.1136/ard.61.4.358
| 15899041 | PMC1174953 | CC BY | 2021-01-04 16:02:36 | no | Arthritis Res Ther. 2005 Mar 1; 7(3):R545-R551 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1715 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17161589904910.1186/ar1716Research ArticleEffect of risedronate on joint structure and symptoms of knee osteoarthritis: results of the BRISK randomized, controlled trial [ISRCTN01928173] Spector Tim D [email protected] Philip G [email protected] J Christopher [email protected] Patrick [email protected] Gary A [email protected] John F [email protected] David J [email protected] Joan M [email protected] Twin Research & Genetic Epidemiology Unit, St Thomas' Hospital, London, UK2 Academic Unit of Musculoskeletal Disease, University of Leeds, UK3 King's College London, School of Biomedical Sciences, London, UK4 SYNARC, Lyon, France5 Procter & Gamble Pharmaceuticals, Mason, OH, USA2005 24 3 2005 7 3 R625 R633 20 12 2004 20 1 2005 1 2 2005 15 2 2005 Copyright © 2005 Spector et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
To determine the efficacy and safety of risedronate in patients with knee osteoarthritis (OA), the British study of risedronate in structure and symptoms of knee OA (BRISK), a 1-year prospective, double-blind, placebo-controlled study, enrolled patients (40–80 years of age) with mild to moderate OA of the medial compartment of the knee. The primary aims were to detect differences in symptoms and function. Patients were randomized to once-daily risedronate (5 mg or 15 mg) or placebo. Radiographs were taken at baseline and 1 year for assessment of joint-space width using a standardized radiographic method with fluoroscopic positioning of the joint. Pain, function, and stiffness were assessed using the Western Ontario and McMaster Universities (WOMAC) OA index. The patient global assessment and use of walking aids were measured and bone and cartilage markers were assessed. The intention-to-treat population consisted of 284 patients. Those receiving risedronate at 15 mg showed improvement of the WOMAC index, particularly of physical function, significant improvement of the patient global assessment (P < 0.001), and decreased use of walking aids relative to patients receiving the placebo (P = 0.009). A trend towards attenuation of joint-space narrowing was observed in the group receiving 15 mg risedronate. Eight percent (n = 7) of patients receiving placebo and 4% (n = 4) of patients receiving 5 mg risedronate exhibited detectable progression of disease (joint-space width ≥ 25% or ≥ 0.75 mm) versus 1% (n = 1) of patients receiving 15 mg risedronate (P = 0.067). Risedronate (15 mg) significantly reduced markers of cartilage degradation and bone resorption. Both doses of risedronate were well tolerated. In this study, clear trends towards improvement were observed in both joint structure and symptoms in patients with primary knee OA treated with risedronate.
==== Body
Introduction
Osteoarthritis (OA) is a chronic, progressive disease that particularly affects weight-bearing joints such as hips and knees. The entire joint is affected by a complex combination of degradative and reparative processes, which alter the anatomy and function of articular cartilage, subchondral bone, and other joint tissues. Of the joints affected, knee OA in particular is a major cause of morbidity, often resulting in knee replacement [1-3]. Moreover, costs associated with OA are high – in the USA alone in 1991, the annual cost of knee replacements was estimated to be more than one billion dollars [4]. OA is normally the result of an interplay between systemic (e.g. age, obesity) and local (e.g. sports injury) factors that affect the joints of the body. Radiographic evidence has shown that joint-space narrowing (a surrogate marker for articular cartilage [5]), sclerosis of the subchondral bone, and the presence of osteophytes are typical structural features of OA.
Current therapies for OA are largely aimed at providing symptom relief, and include a wide range of analgesics (e.g. nonsteroidal anti-inflammatory drugs and cyclooxygenase-2 agents). In contrast, only limited data are available about therapies that modify the course of the disease or affect joint structure. Historically, OA has been considered a disease of the cartilage, but more recent evidence suggests that subchrondral bone is also involved in the pathogenesis, in both disease initiation and progression. For example, increased local bone turnover, decreased bone mineral content and stiffness, and decreased trabecular numbers have been observed in OA subchondral bone structure compared with normal bone [6-8]. There is an increased level of interest in subchondral bone as a therapeutic OA target, and, in particular, the possibility that drugs affecting bone metabolism might alter the progression of knee and hip OA.
The Duncan-Hartley guinea pig model is a widely used spontaneous model of OA progression [9]. Several recent OA studies have evaluated this model for the effects of the antiresorptive agents bisphosphonates. For example, a comparative analysis of multiple bisphosphonates showed that only the group of nitrogen-containing bisphosphonates with pyridinyl sidechains demonstrated significant effects on the cartilage, although not all of these proved effective [10]. In a separate study using the guinea pig OA model, the pyridinyl bisphosphonate risedronate was shown to slow disease progression, as measured by the size and severity of cartilage lesions and the size of osteophytes, by up to 40% [11]. Based upon these preclinical studies, a clinical trial was performed in order to evaluate the effects of risedronate in patients with mild to moderate knee OA. The primary end points were changes in symptoms and function, with secondary end points of changes in joint structure or in markers of joint structure.
Materials and methods
Study design and selection of patients
The British Study of Risedronate in Structure and Symptoms of Knee OA (BRISK) was a 1-year prospective, double-blind, placebo-controlled study conducted in 10 centres in the UK. Male and female subjects aged 40 to 80 years with mild to moderate medial-compartment knee OA, diagnosed according to the clinical and radiological criteria of the American College of Rheumatology [12], were recruited into the trial.
OA in at least one knee, designated the signal knee, was required to meet the following clinical and radiographic inclusion criteria. Clinical inclusion criteria were the presence of daily knee pain for at least 1 month out of the 3 months preceding the study, with at least one of the following: age >50 years, morning knee stiffness of <30 minutes, or knee crepitus. The radiographic criteria for inclusion were a joint-space width (JSW) of 2 to 4 mm in the medial tibiofemoral compartment in the semiflexed anterior–posterior (AP) view of the signal knee and a narrower width than in the lateral compartment of the same knee. Patients were also required to have at least one osteophyte in either the medial or the lateral compartment of the tibiofemoral joint. Major exclusion criteria were the presence of rheumatic diseases that could be responsible for secondary OA; use of intra-articular hyaluronic acid in the signal knee; knee injury or diagnostic arthroscopy of the signal knee in the 6 months preceding enrollment; a history of knee surgery (including arthroscopy requiring an incision of internal joint components) in the signal knee at any time; intra-articular corticosteroids in the 3 months preceding enrollment; the presence of non-OA causes of knee pain in the signal knee (e.g. anserine bursitis, fibromyalgia, or osteonecrosis); and the use of bisphosphonates within the 12 months preceding enrollment.
The subjects gave their written, informed consent before entering the study, which was conducted in accordance with the International Conference on Harmonization (ICH) guidelines for Good Clinical Practice (GCP) and was approved by the UK Multicentre Research Ethical Committee (MREC).
Treatment assignment
The subjects were randomly assigned in a 1:1:1 ratio to one of three arms to receive risedronate at 5 mg or 15 mg or placebo once daily for 1 year. Before randomization, patients were stratified according to their current use of oestrogen or a selective oestrogen receptor modulator.
The subjects were instructed to take their study medication with at least 120 mL of water, 30 minutes before breakfast, or, if the medication was taken later in the day, at least 2 hours before or after food intake and at least 30 minutes before bedtime. They were instructed to take their study medication while they were upright and not to lie down for at least 30 minutes afterwards.
Symptom outcome measures
The outcome instrument for evaluation of risedronate efficacy on symptoms of OA was the Western Ontario and McMaster Universities (WOMAC) OA index [13]. The visual analogue scale (VAS) of the index was used, in which patients assessed each question using a 100-mm scale, with a higher score representing greater symptom severity. The total index score for the signal knee corresponded to the weighted composite of the 24 question scores standardized to a 100-point scale. Scores were also determined for the subscales of pain (5 questions), stiffness (2 questions), and physical function (17 questions). Other symptom outcome measures included the patient global assessment (PGA) of disease, consumption of pain medication, and the use of walking aids. For the PGA, patients answered the following question using a VAS: "Considering all the ways your OA affects you, how have you been in the last 48 hours?" Patients marked values on a scale from 0 to 100 mm.
A step-down reduction in the use of pain medication was effected 5 days before all symptom evaluations. Patients were provided with approximately 30 tablets each of paracetamol (500 mg) and diclofenac (50 mg) to be used as the only pain medication 3 to 5 days before the baseline assessment and at visits at months 3, 6, and 12 (the exit visit). No pain medications were to be used 2 days before the scheduled evaluation date or on the day itself, with the exception of low-dose acetylsalicylic acid (<350 mg/day) for cardiac protection. Rescue analgesia was permitted during the study except for the 2-day washout period before each visit.
Structure outcome measures
The outcome measure for assessment of joint structural changes was the mean change from baseline values in minimum JSW of the medial compartment of the knee. Radiographs of the knee were taken at baseline and at 1 year using a standardized radiographic method with fluoroscopic positioning of the joint in a semiflexed position [14,15]. By the use of this technique, the anterior and posterior rims of the tibia were aligned (to within 1 mm) for reproducible positioning. Radiographs were subjected to extensive quality control at the radiographic facility before dispatch to the Central Analysis Center [15]. Radiographs were read centrally and their quality control was rechecked before computer software was used to obtain the radiographic magnification. This was determined from measurement of a metal ball placed at the head of the fibula at the time of radiography and was used to adjust the computerized measurement obtained of the minimum medial compartment JSW [15]. The test–retest standard deviation of the difference between radiographs taken 2 days apart for this technique was approximately 0.2 mm, based upon repeat measurements in 199 subjects [15]. A retrospective analysis was performed taking into account the precision of the instrument. Retrospectively, clinically meaningful disease progression was defined as joint-space narrowing of ≥ 0.75 mm or a ≥ 25% loss from baseline values. The ≥ 0.75-mm value is almost four times the 0.2-mm standard deviation observed for the x-ray method.
Structure–symptom relation
The relation between knee OA symptoms and radiographic joint-space narrowing was assessed retrospectively; the mean change in symptom scores between baseline to month 12 of the total WOMAC score and pain and function subscales was compared with the magnitude of change in JSW over the study period.
Bone and cartilage markers
Early-morning fasting urine and serum samples were collected at baseline and at months 3, 6, and 12 for assessment of markers of bone and cartilage turnover. Bone resorption was assessed by measurement of urinary levels of the N-terminal crosslinking telopeptide of type I collagen (NTX-I, Osteomark; OrthoClinical Diagnostics, High Wycombe, Bucks, UK) [16]. Bone formation was assessed by measurement of bone-specific serum alkaline phosphatase (Ostase, Beckman-Coulter, San Diego, CA, USA) [17] and cartilage degradation was assessed by measurement of urinary levels of C-terminal crosslinking telopeptide of type II collagen (CTX-II, Cartilaps, Nordic Bioscience, Herlev, Denmark) [18]. The intra-assay and interassay coefficients of variation were lower than 10%.
Evaluation of safety
Patient-reported adverse events (AEs) were recorded throughout the study. AEs were categorized using the Coding Symbols for Thesaurus of Adverse Reaction Terms (COSTART®) coding dictionary. Clinical laboratory measurements for safety monitoring were made throughout the study. Serious AEs were defined as any that resulted in death; were life threatening; resulted in hospitalization; resulted in persistent or significant disability or incapacity; or were judged to be medically significant.
Upper-GI AEs included the following symptoms and conditions: substernal chest pain; duodenitis; dyspepsia; dysphagia; oesophagitis; gastritis; bleeding gastritis; gastro-oesophageal reflux; oesophageal bleeding; GI bleeding; haematemesis; melena; abdominal pain; ulcers (duodenal, oesophageal, peptic, gastric); bleeding ulcers (duodenal, peptic, gastric); perforated ulcers (duodenal, peptic, gastric); perforated and bleeding ulcers (duodenal, peptic, gastric); and reactivated ulcers (duodenal, peptic, gastric).
Statistical analysis
To ensure 80% power to detect a 20% effect of risedronate treatment versus placebo with respect to pain modification (quantified according to the WOMAC pain subscale, assuming a standard deviation of 70 mm on a 0- to 500-mm scale), a 1-year dropout rate of 20%, and a type I error rate of 5% without adjustment for two comparisons with placebo control, the sample size requirement was 100 patients per treatment group.
Analyses were undertaken on the intention-to-treat (ITT) population. This was defined as all randomized patients who received at least one dose of study medication. All statistical analyses were performed using a two-sided statistical test with a type-I-error rate of 0.05. Baseline characteristics were compared using Fisher's exact test for categorical variables and the Kruskal–Wallis test for continuous variables. Extended Mantel–Haenszel tests with pooled centres as strata were used for end points with categorical responses. Analysis of variance (ANOVA) methods were used. Symptom analyses were adjusted for baseline value (PGA < WOMAC total or subscale value, as appropriate), pooled study centres, baseline use of oestrogen or selective oestrogen receptor modulators, gender, age, body mass index, and baseline JSW. Mean JSW analyses were adjusted for pooled study centres, baseline use of oestrogen or selective oestrogen receptor modulators, gender, age, body mass index, and baseline JSW as covariates. Each risedronate group was compared with the placebo group. For walking aids, the percentages were compared with placebo using the Cochran–Mantel–Haenszel test after adjusting for pooled centres. Individual AEs and the proportion of clinically meaningful JSW progressors were analysed using Fisher's exact test.
The WOMAC scores were calculated in accordance with the WOMAC User's Guide [19]. The total scores were composed of subscales weighted as follows: pain = 42%, stiffness = 21%, and function = 37% [19]. For each subscale, the reported response was the patient's average. If at least two pain items, both stiffness items, or more than four physical function items were omitted, or if the patient's response was unclear, the items were regarded as invalid and the relevant subscale was not included.
Results
Patients
Two hundred and eighty-five patients were considered eligible for the study and were randomized to treatment. Of these, 284 received at least one dose of study medication and were included in the ITT population, and 231 (81%) completed the study (placebo, n = 80; risedronate at 5 mg, n = 80; risedronate at 15 mg, n = 71) (Fig. 1). The number of patients who completed the study and the reasons for withdrawal were similar across treatment groups. Table 1 shows the baseline characteristics for the ITT population. These were similar between treatment groups; the average age of the patients was 63.3 years. There were no significant differences in the use of concomitant analgesics between treatment groups. Patients' compliance during study treatment, based on pill counts, was ≥ 83% and was comparable in the three treatment groups.
Symptom outcome measures
There was an improvement from baseline values in the symptom outcome measure of total WOMAC scores (weighted and unweighted) (unweighted not shown) and the subscales for all treatment groups (Fig. 2). The group given risedronate at 15 mg showed a trend towards improvement from baseline values, although the differences were not statistically significant (P values from 0.10 to 0.33).
Assessment of PGAs revealed a statistically significant improvement with risedronate at 15 mg compared with placebo at 1 year (-19.4 for risedronate at 15 mg versus -5.7 for placebo, P < 0.001) (Fig. 3). Although all treatment groups showed a significant improvement from baseline values at 3 months, the improvement in the group receiving risedronate at 15 mg continued to increase with time, whereas the level of improvement with placebo or risedronate at 5 mg did not show any further improvement after 6 months.
Analysis of the use of walking aids during study treatment showed a statistically significant difference in the proportion of patients who used walking aids in patients treated with risedronate at 15 mg (7 patients, 4% reduction) compared with placebo (21 patients, 8% increase) (P = 0.009) at 12 months compared with the proportion of patients that had reported using a walking aid during the previous year.
Structure outcome measures
Assessment of the mean change from baseline values in minimum JSW in the medial compartment of the tibiofemoral joint at 1 year showed that there was a trend for patients receiving risedronate at 5 mg (JSW -0.08 ± 0.44 mm) or 15 mg (JSW -0.06 ± 0.25 mm). The change was greater in patients receiving placebo (JSW -0.12 ± 0.42 mm) compared with baseline values. Overall, the difference between treatment groups in loss of JSW at 12 months was not statistically significant (P = 0.275). The loss in JSW from baseline values was statistically significant only in the placebo group (P < 0.05).
In terms of detectable progression (i.e. loss of JSW ≥ 25% or ≥ 0.75 mm), the analysis of the distribution of change from baseline values in JSW at 1 year showed a greater presence of detectable progression in the placebo (8%) and risedronate (5 mg) (4%) (P = 0.36) groups than in the risedronate (15 mg) group (1%) (P = 0.067). The patients with JSW loss of >0.75 mm included none of the patients treated with risedronate at 15 mg and 6% of the patients treated with placebo (P = 0.060). Similarly, only 1% of the patients treated with risedronate at 15 mg but 7% of the patients treated with placebo had >25% loss of JSW (P = 0.12).
Structure–symptom relation
Figure 4 shows the relation between structure and symptoms for this population of patients. The mean WOMAC total score and the scores on the pain and function subscales increased (i.e. symptom severity increased) with increasing loss of JSW. In the group of patients with any loss of JSW, the mean changes of the WOMAC total score and the scores on the pain and function subscales at 1 year were -5.9 mm, -4.6 mm, and -6.3 mm, respectively, indicating that these symptoms were not increasing overall. In contrast, for the subset of patients with a loss in JSW of 40% or more, the corresponding mean changes were +1.4 mm, +6.0 mm, and +2.3 mm, indicating increased symptom severity in these patients concurrent with narrowing of their knee-joint space.
Markers of biochemical turnover
Risedronate treatment significantly reduced markers of cartilage degradation (Fig. 5) and bone resorption compared with placebo. At 1 year, treatment with risedronate at 15 mg significantly decreased mean urinary CTX-II values, by -22.8% ± 5.35; urinary NTX-I was reduced by -32.9% ± 4.92 relative to baseline values (P < 0.05). Dose-dependent effects were also observed with the 5-mg dose compared with placebo, but to a lesser magnitude. This finding is consistent with the known pharmacologic effect of risedronate on bone turnover. At 1 year, CTX-II and NTX-I values in the placebo group were significantly higher than those in the risedronate 15-mg group (14.5% ± 5.4 and 17.2% ± 4.9 higher, respectively). Significant decreases in bone alkaline phosphatase were observed in the risedronate groups compared with placebo. At 1 year, the mean decreases in the groups receiving risedronate at 15 mg and 5 mg were 29.1% ± 2.6 and 19.5% ± 2.5, respectively, compared with a mean decrease of 2.7% ± 2.5 in the placebo group (P < 0.001).
Safety
The frequencies of AEs were similar in the two treatment groups (Table 2). There were no clinically meaningful differences between groups in the percentage of patients with AEs in any body system and there were no significant differences in routine clinical chemistry parameters between the risedronate groups and the placebo group. The numbers of patients who dropped out of the study because of AEs were similar. Overall, 34 patients reported a total of 53 serious AEs. Investigators considered four serious AEs as possibly related to study treatment; two of these (rash and diarrhea) were in patients treated with placebo and two (anaemia and increased general joint pain) were in patients treated with risedronate at 5 mg.
Table 2 provides a summary of adverse events for the ITT population and the frequency of the overall GI AEs and the most common upper-GI AEs. Forty-seven upper-GI events were reported in 38 patients, of which abdominal pain and dyspepsia were the most frequently reported. The majority of the upper-GI AEs occurred in patients with a history of GI disease; there were no significant differences between the groups given risedronate and the group given placebo in the incidence of upper-GI AEs in these patients.
Discussion
Increased evidence of the role of bone in both the initiation and progression of OA has resulted in an interest in drugs that affect bone metabolism and might slow or even halt the process of joint degeneration [6]. The early findings reported here suggest that the bisphosphonate risedronate may have disease-modifying effects in patients with knee OA. A recent cross-sectional study also suggests an association between antiresorptive treatments (oestrogen or bisphosphonate) and improved symptoms and/or decreased bone marrow abnormalities [20].
Positive trends were observed with risedronate treatment with regard to symptomatic improvement, as assessed by the WOMAC index. Treatment with risedronate at 15 mg resulted in a consistent trend in improvement in WOMAC scores, whereas the group receiving placebo showed less improvement. The group receiving risedronate at 15 mg showed a significant improvement in the PGA of OA compared with placebo. Similarly, the percentage of patients who used a walking aid during the study decreased in the group treated with risedronate at 15 mg, contrasting with an increase in the placebo group. While the differences in JSW among the groups were not significant, there was a trend for less loss in the risedronate (15 mg) than in the placebo group. When the data were analysed in a post hoc manner to identify patients with detectable progression (i.e. approximately four times the precision of the measurement), 8% of patients receiving placebo and 1% of patients receiving 15 mg risedronate were found to exhibit this degree of progression. Additionally, risedronate significantly reduced levels of bone resorption and cartilage degradation, as assessed by NTX-I and CTX-II markers, respectively.
Despite these encouraging results, they were not confirmed in multicentre studies of risedronate treatment for 2 years using a similar protocol [21]. The following provides perspectives comparing the two studies. In our study, there was a dose-dependent trend for improvement in the WOMAC score. In the 2-year North American study, which enrolled 1,232 patients with knee OA, the placebo effect was approximately twice that observed in our study, and was comparable to the magnitude of change observed in the group given 15 mg risedronate in our study. Both studies showed significant decreases in CTX-II, the marker of cartilage degradation, but these were not associated with an attenuation of joint-space loss This lack of association may be related to the length of observation in these studies. Reijman and colleagues [22] observed 1,235 men and women with OA, followed up over an average period of 6.6 years. The subjects with baseline CTX-II levels in the highest quartile had a sixfold risk of progression of knee OA, defined as a decrease in JSW ≥ 2 mm, in comparison with subjects in the lowest quartile [22]. This suggests that an enriched population of subjects with an elevated rate of cartilage loss combined with a longer study period may be required in order to observe significant treatment effects.
Our study is one of the first to suggest a correlation between symptoms and structure in OA, although preliminary results with doxycycline in the treatment of obese women with knee OA have reported a significant reduction in the proportion of follow-up visits in which a clinically significant increase in pain occurred, favouring treatment over placebo, and coinciding with a decrease in joint-space narrowing [23]. This finding is important, because it runs contrary to previous results, which have suggested a poor correlation between these disease features [24,25]. The limitations of our study include the small number of patients (n = 10) with the greatest loss of JSW (>40%), for observations of concurrent increase in joint-space narrowing and WOMAC symptom severity scores. One possible explanation for discrepancies between the current study and previous studies is the difference in radiographic methodologies used. Several radiographic techniques have been described for measuring JSW in the knee. We used a highly standardized, fluoroscopic technique in which the knee was semiflexed. Recent studies have compared different radiograph imaging methods [26,27]. The results highlighted the importance of medial tibial plateau alignment with the central x-ray beam and showed that the standard clinical view of a standing extended knee is subject to considerable variability. In contrast, the fluoroscopic technique is well validated and is less variable in test-retest performance [15,28]. Further studies are required to further explore the possible correlation between symptoms and structure observed in our study. If validated, this relation may allow physicians to use the assessment of pain, perhaps in combination with a biomarker such as CTX-II, as a surrogate for other measures of disease progression
Conclusion
This study is one of the first to show a correlation between symptoms and joint structure changes in knee OA. While our findings were suggestive of a beneficial effect of risedronate treatment on preservation of bone and cartilage, these trends seen in this study have not been observed in larger, multicountry cohorts.
Abbreviations
AE = adverse event; BRISK = British Study of Risedronate in Structure and Symptoms of Knee OA; CTX-II = C-terminal crosslinking telopeptide of type II collagen; GI = gastrointestinal; ITT = intention-to-treat; JSW = joint-space width; NTX-I = N-terminal crosslinking telopeptide of type I collagen; OA = osteoarthritis; PGA = patient global assessment; VAS = visual analogue scale; WOMAC = Western Ontario and McMaster Universities [OA index].
Competing interests
This manuscript was sponsored by a grant from Procter & Gamble.
Authors' contributions
TDS, JFB, DJV, and JMM planned the study and prepared the manuscript. GAC performed the statistical analysis. JCB-W supervised radiological measurements. PG performed the marker assays. PGC assisted with the manuscript and recruited patients. All authors read and approved the final manuscript.
Acknowledgements
We thank the following investigators who participated in the trial: Prof C Cooper, Osteoporosis Clinical Research Unit, Southampton General Hospital; Dr M Horne, Synexus Limited, Reading; Prof P Emery, Rheumatology Out Patients Clinic, Leeds; Dr J Fraser, Synexus Limited, Wrightington Hospital, Wigan; Dr E George, Arrowe Park Hospital, Department of Rheumatology, Merseyside; Dr R Hughes, Rheumatology Department, St Peter's Hospital, Surrey; Dr M Irani, Ashford St Peter's Hospitals NHS Trust, Rheumatolgy Department, Middlesex; Prof P Maddison, Rheumatology Department, Bangor, North Wales; Prof G Nuki, University of Edinburgh, Western General Hospital, Edinburgh; Dr R Price, Department of Rheumatology, Queen Elizabeth Hospital, Woolwich; Dr J Robinson, Synexus Limited, Crosby Clinical Research Centre, Liverpool, UK. The authors would also like to acknowledge Chad Melson and Ruby Xia for their invaluable programming support.
Figures and Tables
Figure 1 Disposition of patients with knee osteoarthritis in a controlled, randomized trial of risedronate.
Figure 2 Changes in mean values at 12 months from baseline measures in patients with osteoarthritis. Patients were given risodronate (Ris) or placebo. Scores were the weighted composite of the 24 question scores on the visual analogue scale (1 to 100 mm) of the Western Ontario and McMaster Universities (WOMAC) osteoarthritis index or its subscales for pain, stiffness, and physical function. Vertical lines represent standard errors of the mean. P values refer to risedronate (15 mg) vs baseline values.
Figure 3 Changes in mean patient global assessment after risedronate or placebo in osteoarthritis. Vertical lines represent standard errors of the mean. Ris, risedronate.
Figure 4 Relation between WOMAC scores and minimum percentage change in JSW after 1 year in patients with osteoarthritis. Scores were the weighted composite (1 to 100 mm) of the Western Ontario and McMaster Universities (WOMAC) osteoarthritis index or its subscales for pain and physical function. JSW, joint-space width.
Figure 5 Changes in urinary CTX-II levels after treatment of osteoarthritis patients with placebo or risedronate. Ris, risedronate.
Table 1 Baseline characteristics of the intention-to-treat (ITT) population with osteoarthritis of the knee
Risedronate
Characteristic Placebo (n = 98) 5 mg/day (n = 96) 15 mg/day (n = 90) P
Age (years) 63.2 (0.82) 62.9 (0.90) 63.8 (0.88) 0.652a
Height (cm) 164.3 (0.90) 165.3 (0.97) 165.1 (0.93) 0.614a
Body mass index (kg/m2) 29.2 (0.38) 29.0 (0.40) 29.2 (0.42) 0.799a
Sex (no. (%)) 0.307b
Male 34 (35%) 40 (42%) 41 (46%)
Female 64 (65%) 56 (58%) 49 (54%)
Post-menopausal (no. (%)) 54 (84%) 49 (88%) 43 (88%) 0.985b
Years since menopause 14.4 (1.22) 16.5 (1.57) 17.4 (1.42) 0.273a
Oestrogen or SERM use (no. (%)) 18 (28%) 15 (27%) 10 (20%) 0.380b
Race (no.) 0.385b
Asian Oriental 1 0 0
Asian Indian 1 0 2
Caucasian 96 96 88
Joint-space width (mm) 3.03 (0.05) 2.95 (0.05) 3.01 (0.06) 0.577a
WOMAC – weighted total score 50.3 (2.0) 46.1 (2.0) 49.4 (2.4) 0.281a
Use of walking aids (no. (%)) 16 (16) 22 (23) 12 (13) 0.224b
uCTX-II (ng/mmol creatinine) 312.5 (19.9) 328.9 (29.7) 340.1 (24.0) 0.748a
uNTX-I (nmol/mmol creatinine) 40.3 (2.8) 42.3 (4.5) 38.6 (2.2) 0.954a
Unless indicated otherwise, values are means (standard errors of the mean). aKruskal–Wallis test; bFisher exact test. SERM, selective oestrogen receptor modulator; uCTX-II, urinary C-terminal crosslinking telopeptide of type II collagen; uNTX-I, urinary N-terminal crosslinking telopeptide of type I collagen; WOMAC, Western Ontario and McMaster Universities osteoarthritis index.
Table 2 Summary of adverse events (no. (%)) in patients with osteoarthritis who received risedronate or placebo
Placebo Risedronate
(n = 98) 5 mg/day (n = 96) 15 mg/day (n = 90)
Patients with AEs 94 (96%) 95 (99%) 84 (93%)
Dropouts due to AEs 12 (12%) 7 (7%) 10 (11%)
Overall GI AEs 15 (15%) 16 (17%) 7 (8%)
Abdominal pain 6 (6%) 7 (7%) 3 (3%)
Dyspepsia 7 (7%) 7 (7%) 3 (3%)
Gl disorder 3 (3%) 1 (1%) 1 (1%)
AE, adverse event; GI, gastrointestinal.
==== Refs
Dixon T Shaw M Ebrahim S Dieppe P Trends in hip and knee joint replacement: socioeconomic inequalities and projections of need Ann Rheum Dis 2004 63 825 830 15194578 10.1136/ard.2003.012724
Millar WJ Hip and knee replacement Health Rep 2002 14 37 50 15069801
Melzer D Guralnik JM Brock D Prevalence and distribution of hip and knee joint replacements and hip implants in older Americans by the end of life Aging Clin Exp Res 2003 15 60 66 12841420
Quam JP Michet CJ JrWilson MG Rand JA Ilstrup DM Melton LJ 3rdWallrichs SL Total knee arthroplasty: a population-based study Mayo Clin Proc 1991 66 589 595 2046396
Buckland-Wright JC Macfarlane DG Lynch JA Jasani MK Bradshaw CR Joint space width measures cartilage thickness in osteoarthritis of the knee: high resolution plain film and double contrast macroradiographic investigation Ann Rheum Dis 1995 54 263 268 7763102
Spector TD Bisphosphonates: potential therapeutic agents for disease modification in osteoarthritis Aging Clin Exp Res 2003 15 413 418 14703007
Bettica P Cline G Hart DJ Meyer J Spector TD Evidence for increased bone resorption in patients with progressive knee osteoarthritis: longitudinal results from the Chingford study Arthritis Rheum 2002 46 3178 3184 12483721 10.1002/art.10630
Li B Aspden RM Mechanical and material properties of the subchrondral bone plate from the femoral neck of patients with osteoarthritis or osteoporosis Ann Rheum Dis 1997 56 247 254 9165997
Bendele AM Hulman JF Spontaneous cartilage degeneration in guinea pigs Arthritis Rheum 1988 31 561 565 3358814
Meyer JM Dansereau SM Farmer RW Jeans GL Prenger MC Bisphosphonates structurally similar to risedronate (actonel) slow disease progression in the guinea pig model of primary osteoarthritis [abstract] Arthritis Rheum 2001 5307 abstract 1527
Meyer J Farmer R Prenger MC Risedronate but not alendronate slows disease progression in the guinea pig model of primary osteoarthritis [abstract] J Bone Miner Res 2001 5305 abstract SA472
Altman R Asch E Bloch D Bole G Borenstein D Brandt K Christy W Cooke TD Greenwald R Hochberg M Development of criteria for the classification and reporting of osteoarthritis. Classification of osteoarthritis of the knee. Diagnostic and Therapeutic Criteria Committee of the American Rheumatism Association Arthritis Rheum 1986 29 1039 1049 3741515
Bellamy N Buchanan WW Goldsmith CH Campbell J Stitt LW Validation study of WOMAC: a health status instrument for measuring clinically important patient relevant outcomes to antirheumatic drug therapy in patients with osteoarthritis of the hip or knee J Rheumatol 1988 15 1833 1840 3068365
Buckland-Wright JC Protocols for precise radio-anatomical positioning of the tibiofemoral and patellofemoral compartments of the knee Osteoarthritis Cartilage 1995 71 80 8581753
Buckland-Wright JC Bird CF Ritter-Hrncirik CA Cline GA Tonkin C Hangartner TN Ward RJ Meyer JM Meredith MP X-ray technologists' reproducibility from automated measurements of the medial tibiofemoral joint space width in knee osteoarthritis for a multicenter, multinational clinical trial J Rheumatol 2003 30 329 338 12563691
Hanson DA Weis MA Bollen AM Maslan SL Singer FR Eyre DR A specific immunoassay for monitoring human bone resorption: quantitation of type I collagen cross-linked N-telopeptides in urine J Bone Miner Res 1992 7 1251 1258 1466251
Garnero P Delmas PD Assessment of the serum levels of bone alkaline phosphatase with a new immunoradiometric assay in patients with metabolic bone disease J Clin Endocrinol Metab 1993 77 1046 1053 8104954 10.1210/jc.77.4.1046
Christgau S Garnero P Fledelius C Moniz C Ensig M Gineyts E Rosenquist C Qvist P Collagen type II C-telopeptide fragments as an index of cartilage degradation Bone 2001 29 209 215 11557363 10.1016/S8756-3282(01)00504-X
Bellamy N WOMAC Osteoarthritis Index: A User's Guide 1995 2 London, Ontario: Victoria Hospital
Carbone LD Nevitt MC Wildy K Barrow KD Harris F Felson D Peterfy C Visser M Harris TB Wang BW The relationship of antiresorptive drug use to structural findings and symptoms of knee osteoarthritis Arthritis Rheum 2004 50 3516 3525 15529367 10.1002/art.20627
Bingham C Beary J Cohen S Clauw D Cline G Meyer J Clinically significant placebo improvement occurs by 6 months and is maintained through 24 months in a study of knee OA pain and function [abstract] American College of Rheumatology Annual Meeting: October 16–21 2004; San Antonio, TX 2004 Abstract 263
Reijman M Hazes JM Birma-Zeinstra SM Koes BW Christgau S Christiansen C Uitterlinden AG Pols HA A new marker for osteoarthritis: cross-sectional and longitudinal approach Arthritis Rheum 2004 50 2471 2478 15334460 10.1002/art.20332
Brandt KD Mazzuca SA Katz BP Lane KA Doxycycline Study Group The disease modifying effect of doxycycline (doxy) includes symptomatic benefit for patients with knee osteoarthritis [abstract] Arthritis Rheum 2003 48 3643 Abstract LB1 10.1002/art.11017
Reginster JY Deroisy R Rovati LC Lee RL Lejeune E Bruyere O Giacovelli G Henrotin Y Dacre JE Gossett C Long-term effects of glucosamine sulphate on osteoarthritis progression: a randomised, placebo-controlled clinical trial Lancet 2001 357 251 256 11214126 10.1016/S0140-6736(00)03610-2
Dieppe PA Cushnaghan J Shepstone L The Bristol 'OA500' study: progression of osteoarthritis (OA) over 3 years and the relationship between clinical and radiographic changes at the knee joint Osteoarthritis Cartilage 1997 5 87 97 9135820
Buckland-Wright JC Wolfe F Ward RJ Flowers N Hayne C Substantial superiority of semiflexed (MTP) views in knee osteoarthritis: a comparative radiographic study, without fluoroscopy, of standing extended, semiflexed (MTP), and schuss views J Rheumatol 1999 26 2664 2674 10606380
Mazzuca SA Brandt KD Dieppe PA Doherty M Katz BP Lane KA Effect of alignment of the medial tibial plateau and x-ray beam on apparent progression of osteoarthritis in the standing anteroposterior knee radiograph Arthritis Rheum 2001 44 1786 1794 11508430 10.1002/1529-0131(200108)44:8<1786::AID-ART315>3.0.CO;2-L
Mazzuca SA Brandt KD Buckland-Wright JC Buckwalter KA Katz BP Lynch JA Ward RJ Emsley CL Field test of the reproducibility of automated measurements of medial tibiofemoral joint space width derived from standard knee radiographs J Rheumatol 1999 26 1359 1365 10381056
| 15899049 | PMC1174954 | CC BY | 2021-01-04 16:02:36 | no | Arthritis Res Ther. 2005 Mar 24; 7(3):R625-R633 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1716 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17171589905010.1186/ar1717Research ArticleBiomarkers of endothelial dysfunction, cardiovascular risk factors and atherosclerosis in rheumatoid arthritis Dessein Patrick H [email protected] Barry I 2Singh Sham 31 Department of Rheumatology, Johannesburg Hospital and Milpark Hospital, Parktown, University of the Witwatersrand, Johannesburg, South Africa2 Centre for Diabetes and Endocrinology, Houghton, University of the Witwatersrand, Johannesburg, South Africa3 Lancet Laboratories, Richmond, Johannesburg, South Africa2005 24 3 2005 7 3 R634 R643 9 1 2005 24 1 2005 2 2 2005 15 2 2005 Copyright © 2005 Dessein et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Cardiovascular event rates are markedly increased in rheumatoid arthritis (RA), and RA atherogenesis remains poorly understood. The relative contributions of traditional and nontraditional risk factors to cardiovascular disease in RA await elucidation. The present study comprises three components. First, we compared biomarkers of endothelial dysfunction (vascular cell adhesion molecule [VCAM]-1, intercellular adhesion molecule [ICAM]-1 and endothelial leucocyte adhesion molecule [ELAM]-1) in 74 RA patients and 80 healthy control individuals before and after controlling for traditional and nontraditional cardiovascular risk factors, including high-sensitivity C-reactive protein (hs-CRP), IL-1, IL-6 and tumor necrosis factor-α. Second, we investigated the potential role of an extensive range of patient characteristics in endothelial dysfunction in the 74 RA patients. Finally, we assessed associations between biomarkers of endothelial dysfunction and ultrasonographically determined common carotid artery intima–media thickness and plaque in RA. The three biomarkers of endothelial dysfunction, as well as hs-CRP, IL-1, IL-6 and tumor necrosis factor-α, were higher in patients than in control individuals (P < 0.0001). Patients were also older, exercised less and had a greater waist circumference, blood pressure and triglyceride levels (P ≤ 0.04). Five patients had diabetes. Differences in endothelial function were no longer significant between patients and controls (P = 0.08) only after both traditional and nontraditional cardiovascular risk factors were controlled for. In the 74 RA patients, IL-6 predicted levels of all three biomarkers (P ≤ 0.03), and rheumatoid factor titres and low glomerular filtration rate (GFR) both predicted levels of VCAM-1 and ICAM-1, independent of traditional cardiovascular risk factors (P ≤ 0.02). VCAM-1 was associated with common carotid artery intima–media thickness (P = 0.02) and plaque (P = 0.04) in RA. Patients had impaired endothelial function, less favourable traditional cardiovascular risk factor profiles, and higher circulating concentrations of hs-CRP and cytokines compared with healthy control individuals. Both traditional and nontraditional cardiovascular risk factors contributed to the differences in endothelial function between RA patients and healthy control individuals. IL-6, rheumatoid factor titres and low GFR were independently predictive of endothelial dysfunction in RA. Disease-modifying agents that effectively suppress both cytokine and rheumatoid factor production, and interventions aimed at preserving renal function may attenuate cardiovascular risk in RA.
==== Body
Introduction
Cardiovascular disease is an increasingly recognized contributor to excess morbidity and mortality in rheumatoid arthritis (RA) [1-5]. Traditional cardiovascular risk factors do not adequately accunt for the extent of cardiovascular disease in RA [3,5]. Although hypertension and age are potential additional contributors to cardiovascular events in this disease [6], markers of current and cumulative inflammation (white cell counts and radiographic joint damage, respectively) are associated with ultrasonographically determined subclinical atherosclerosis [7,8] – a predictor of cardiovascular events [9].
Atherosclerosis often develops subclinically over prolonged periods of time; therefore, it may be too insensitive to show associations with recently acquired or temporarily active modifiable cardiovascular risk factors, such as systemic inflammation secondary to recent onset or uncontrolled RA. Clearly, other outcome variables that can identify patients at risk for cardiovascular disease at any point in time are needed in RA.
One such potential marker is endothelial dysfunction – an essential step in atherogenesis [10]. Most if not all risk factors that are related to cardiovascular disease are also associated with endothelial dysfunction, and the process is reversible with effective treatment of operative risk factors [10]. Endothelial status may be regarded as an integrated index of all atherogenic and atheroprotective factors present in an individual [10].
Several methods have been employed to assess endothelial function. Using ultrasonographically measured brachial artery flow mediated dilatation or vasodilatory responses to intrabrachial artery infusion of acetylcholine, some [11-14] but not all investigators [15] have shown impaired endothelial function in RA. Endothelial dysfunction was related to inflammation [12] and HLA-DR1 [11], and was found to improve with infliximab treatment [13]. An alternative method to assess endothelial function involves the measurement of biomarkers of endothelial activation and dysfunction (circulating vascular cell adhesion molecule [VCAM]-1, intercellular adhesion molecule [ICAM]-1, and endothelial leucocyte adhesion molecule [ELAM]-1 [or selectin]) [16-20]. Elevated circulating adhesion molecules are associated with cardiovascular risk factors [17] and predict atherosclerosis and cardiovascular events [18-20]. The measurement of circulating adhesion molecules may not add much predictive information to that provided by more established risk factors in the general population [21]. In contrast, it has been reported that such biomarkers play a more important role than traditional risk factors in cardiovascular disease in RA [22,23]. Important in this context is that high circulating adhesion molecule levels may not only reflect synovial inflammation but also indicate exposure of the systemic vascular endothelium to high circulating cytokine concentrations [24].
To our knowledge, the relative impact of traditional versus nontraditional cardiovascular risk factors on endothelial dysfunction as assessed using biomarkers has not been reported in RA. The present study comprises three components. First, we compared biomarkers of endothelial dysfunction in 74 RA patients and 80 healthy control individuals before and after controlling for potential explanatory variables, including both traditional and nontraditional cardiovascular risk factors. Second, we investigated in the 74 RA patients the potential role played by an extensive range of patient characteristics in endothelial dysfunction. Finally, we assessed the association between biomarkers of endothelial dysfunction and ultrasonographically determined common carotid artery (CCA) intima–media thickness (IMT) and plaque [9].
Materials and methods
Biomarkers of endothelial dysfunction in RA patients and healthy control individuals
Seventy-six consecutive patients who fulfilled the American College of Rheumatology criteria for RA [25] were invited to participate. Only two patients refused, and the remaining 74 were included. Patients receiving lipid-lowering and antidiabetic medications were excluded. The baseline clinical and routine laboratory characteristics of the included RA patients are reported elsewhere [26]. Eighty-three individuals with no known diseases and who were not taking any medication agreed to act as controls. These were friends, patient friends and laboratory staff. Three of them were excluded because they were found to have impaired fasting glucose (plasma glucose >5.5 mmol/l); the remaining 80 were included in the study. The study was approved by the Ethics Committees for Research on Human Subjects (Medical) of the University of the Witwatersrand and the Milpark Hospital.
We recorded traditional cardiovascular risk factors in both RA patients and control individuals using previously reported methods (Table 1) [26]. We also recorded the following nontraditional cardiovascular risk factors: circulating high-sensitivity C-reactive protein (hs-CRP), cytokines (IL-1, IL-6 and tumor necrosis factor [TNF]-α), cytokine suppressant therapy (disease-modifying antirheumatic drug [DMARD] and prednisone use) and biomarkers of endothelial dysfunction (Tables 1 and 2). Blood sampling was performed on the same day that clinical data were recorded. Fasting blood samples were taken between 08:00 and 10:00 hours. All laboratory analyses except for assessments of cytokines and biomarkers of endothelial dysfunction were performed within 2 hours of sampling. These comprised lipids, hs-CRP and other laboratory tests that were performed for the second component of the study, which was conducted in the RA patients only (see below). Blood samples drawn for determination of cytokines and biomarkers of endothelial dysfunction were stored at -70°C before laboratory testing. Cytokines and adhesion molecules were measured using enzyme-linked immunosorbent assays (Hiss Diagnostics GmbH, Freiburg, Switzerland). The intra-assay and inter-assay coefficients of variation (respectively) were 5.1% and 8.6% for IL-1, 3.4% and 5.2% for IL-6, 6.9% and 7.4% for TNF-α, 3.1% and 5.2% for VCAM-1, 4.1% and 7.7% for ICAM-1, and 5.4% and 6.0% for ELAM-1.
Statistical analysis
The traditional and nontraditional risk factors were compared using the Mann–Whitney U-test (continuous variables) and the χ2 test (dichotomous variables; Table 1). Apart from hs-CRP, cytokines and cytokine suppressant therapy (DMARD and glucocorticoid use), several traditional cardiovascular risk factors differed between RA patients and control individuals. The ability of traditional and nontraditional cardiovascular risk factors to account for differences in the levels of biomarkers of endothelial dysfunction between RA patients and control individuals was assessed in logistic regression models (Table 2), using RA status as the dependent variable (RA = 1, non-RA control = 0). Continuous variables that were not normally distributed were logarithmically transformed. P < 0.05 was considered statistically significant.
Relationship between patient characteristics and biomarkers of endothelial dysfunction in RA patients
For the second component of the study, recorded variables other than cytokines and biomarkers of endothelial dysfunction are summarized in Tables 3 and 4. A descriptive analysis of these variables in the current cohort was previously reported [26]. Although 10 patients were on thyroxine replacement therapy and eight had subclinical hypothyroidism (thyrotropin >4 μIU/ml and normal thyroxine levels) [27], none had clinical hypothyroidism (decreased thyroxine levels) at the time of the study. IgM rheumatoid factor was determined using an immunoturbidimetric test on the Olympus AU 600 analyzer. The intra-assay and inter-assay coefficients of variation for rheumatoid factor were 3.4% and 4.6%, respectively.
Statistical analysis
Associations between patient characteristics and biomarkers of endothelial dysfunction were first assessed by univariate analyses comprising the Spearman correlation coefficients (continuous variables; Table 3) and the Mann–Whitney U-test (dichotomous variables; Table 4). Because many univariate analyses were conducted in this component of the study, P < 0.01 was considered statistically significant. Because IL-6, rheumatoid factor and low glomerular filtration rate (GFR) were associated with biomarkers of endothelial dysfunction, their potential roles as predictors of endothelial dysfunction were further assessed after controlling for traditional cardiovascular risk factors in multivariable regression models (Table 5). Continuous variables that were not normally distributed were logarithmically transformed. In these multivariable models, P < 0.05 was considered statistically significant.
Associations between biomarkers of endothelial dysfunction and common carotid artery intima–media thickness and plaque in RA patients
The CCAs were evaluated using high resolution B-mode ultrasound. Details of this investigation in the present cohort were previously reported [26].
Statistical analysis
The associations between biomarkers of endothelial dysfunction and CCA IMT and plaque were determined using the Spearman correlation coefficient and the Mann–Whitney U test, respectively. P < 0.05 was considered statistically significant.
Results
Biomarkers of endothelial dysfunction in RA patients and healthy control individuals
The recorded baseline characteristics in the RA patients and control individuals comprised traditional cardiovascular risk factors, hs-CRP and cytokines (Table 1). RA patients were younger, exercised less, had a higher waist circumference, higher systolic and diastolic blood pressures, and higher triglyceride levels than did control individuals. Five RA patients had diabetes that was being treated with dietary intervention only. With regard to nontraditional cardiovascular risk factors, hs-CRP, IL-1, IL-6 and TNF-α concentrations were higher in patients than in control individuals, and DMARD and prednisone were used by 56 and 11 RA patients, respectively. No patient was being treated or had been treated with biological agents at the time of the study.
Results for biomarkers of endothelial dysfunction in patients and control individuals are shown in Table 2. In univariate analysis, all three biomarkers of endothelial dysfunction (VCAM-1, ICAM-1 and ECAM-1) were higher in patients than in control individuals. These differences remained significant after controlling for cytokine suppressant agent (DMARD and prednisone) use (model 1) or traditional cardiovascular risk factors (model 2). After controlling for nontraditional cardiovascular risk factors, the differences in ELAM-1 between patients and control individuals were no longer significant (model 3). When traditional and nontraditional cardiovascular risk factors were simultaneously controlled for, the differences in levels of all three biomarkers of endothelial dysfunction were no longer significant between patients and control individuals (model 4).
Relationship between patient characteristics and biomarkers of endothelial dysfunction in RA patients
In univariate analysis, VCAM-1 was related to rheumatoid factor titres and low GFR, ICAM-1 to rheumatoid factor titres and IL-6, and ELAM-1 to IL-6 (Table 4). None of the other recorded baseline characteristics in the 74 RA patients were associated with biomarkers of endothelial dysfunction (Tables 4 and 5).
After controlling for traditional cardiovascular risk factors in multivariable regression models, IL-6 was predictive of all three biomarkers of endothelial dysfunction, and rheumatoid factor titre and low GFR were both predictive of VCAM-1 and ICAM-1 (Table 5). Additional controlling for thyrotropin levels did not materially alter these models (data not shown).
Associations between biomarkers of endothelial dysfunction and common carotid artery intima–media thickness and plaque in RA patients
As previously reported, the median (range) CCA IMT was 0.654 mm (0.496–1.150 mm), and 23 (31%) patients had plaque [26]. Twenty-one (28%) patients had a CCA IMT under 0.600 mm and no plaque. The role of clinical and routine laboratory characteristics as predictors of common carotid atherosclerosis in the present cohort was also previously reported [26]. VCAM-1 concentrations correlated with the CCA IMT (rs = 0.280; P = 0.016) and were higher in patients with plaque than in those without plaque (769 pg/ml [391–2073 pg/ml] versus 703 pg/ml [445–2001 pg/ml]; P = 0.043). The associations between CCA findings and other biomarkers did not achieve statistical significance.
Discussion
Biomarkers of endothelial dysfunction in RA patients and healthy control individuals
We found that biomarkers of endothelial dysfunction were markedly higher in RA patients than in healthy control individuals. Increased circulating adhesion molecule concentrations have been reported in RA [23,24,28,29]. Our patients also had higher hs-CRP and IL-1, IL-6 and TNF-α concentrations than did control individuals, but these nontraditional cardiovascular risk factors did not fully account for the differences in biomarkers of endothelial dysfunction between patients and control individuals. Indeed, the RA patients also had generally less favourable traditional cardiovascular risk factor profiles than healthy control individuals.
The younger age of the healthy control individuals included reflects the difficulties in recruiting healthy aged persons who are not taking any medication. Hence, we controlled for age as well as other traditional cardiovascular risk factors when assessing the differences in biomarkers of endothelial dysfunction between patients and control individuals. Of interest, the body mass indices were similar in both groups, but patients had higher waist circumference, blood pressure and triglyceride levels. The latter are features of the metabolic syndrome [5,30]. Although differences in age and exercise habits might have contributed to these findings, features of the metabolic syndrome also cluster in RA because of inflammation-induced insulin resistance [5,30].
In multivariable models, traditional cardiovascular risk factors attenuated the differences in biomarkers of endothelial dysfunction between patients and control individuals to a lesser extent than did nontraditional cardiovascular risk factors. However, the differences in endothelial function between patients and control individuals were no longer significant only after both traditional and nontraditional cardiovascular risk factors had been controlled for. Our results also show the need for comprehensive assessment of cardiovascular risk factors in healthy individuals when comparing their endothelial function with that in RA patients.
Relationship between patient characteristics and biomarkers of endothelial dysfunction in RA patients
We investigated the potential role of an extensive range of patient characteristics in endothelial dysfunction in RA. IL-6, rheumatoid factor titre and low GFR predicted endothelial dysfunction, as assessed using biomarkers.
In multivariable analysis, IL-6 predicted endothelial dysfunction independent of traditional cardiovascular risk factors. Chronic cytokine release from inflamed joints was previously implicated in the increased production of adhesion molecules by endothelial cells in RA [4,31]. Circulating cytokines could impair endothelial function directly [4] or through their effects on insulin sensitivity and on CRP and fibrinogen (a major determinant of erythrocyte sedimentation rate [32]) production by the liver [4]. In the present cohort of unselected RA patients, IL-6 was more strongly associated with endothelial dysfunction than were CRP, erythrocyte sedimentation rate and insulin resistance. In contrast to IL-6, circulating IL-1 and TNF-α concentrations were not associated with endothelial dysfunction. IL-1 and TNF-α are major proinflammatory cytokines in RA joints and stimulate IL-6 production by synovial fibroblasts [33-37], whereas IL-6 is a major circulating cytokine in RA that induces the acute phase response, production of immunoglobulins by B cells and neuroendocrine alterations [33,34,37,38]. IL-6 promotes adhesion molecule expression and stimulates macrophages to secrete monocyte chemotactic protein-1 [39]. Circulating IL-6 concentrations also predict cardiovascular disease in the general population, independent of hs-CRP levels [39].
Apart from IL-6, rheumatoid factor was also predictive of endothelial dysfunction independent of traditional cardiovascular risk factors in the present cohort. The mechanism underlying the strong association between rheumatoid factor and endothelial dysfunction in RA cannot be discerned from our data. However, rheumatoid factor may directly cause endothelial injury [31]. Direct evidence for a role for humoral immunity in atherosclerosis was found by George and coworkers [40]. In their study repeated intraperitoneal administration of IgG from serum of mice immunized with heat shock protein 65 enhanced fatty streak formation in mice in comparison with their control anti-bovine serum albumin injected littermates [40]. Rheumatoid factor is produced by B cells that are highly effective at presenting antigens to T cells [41] and T-cell activation in rheumatoid synovium is B-cell dependent [42]. The recently reported remarkable efficacy of B-cell depletion in rheumatoid factor positive RA with rituximab [43] indicates that B cells are key contributors to the immunopathogenesis of RA. In a recent autopsy report on two RA patients with coronary artery disease the coronary plaques and adventitia contained large numbers of B cells, whereas in coronary artery disease it is typical for lymphocytic infiltrates to consist almost exclusively of T cells [44]. These reports and our findings support a role for humoral mechanisms in RA atherogenesis.
Finally, a low GFR also predicted endothelial dysfunction in our RA patients. Although only four (5%) patients had an elevated serum creatinine (>115 μmol/l), GFR estimation using the Cockcroft–Gault equation [45] revealed that 16 (22%) patients had a GFR under 60 ml/min [26], which is indicative of chronic kidney disease. The high prevalence of cardiovascular disease in individuals with chronic kidney disease has been amply reported [45]. This is attributable to the high prevalence of traditional risk factors such as older age, high total cholesterol, low high-density lipoprotein cholesterol and diabetes, as well as to nontraditional risk factors such as oxidant stress, inflammation and, to a lesser extent, hyperhomocysteinaemia. In the present cohort, varimax rotated factor analysis confirmed strong relationships between low GFR, age and hyperhomocysteinaemia [26]. Also, independent of age as well as other traditional cardiovascular risk factors, a low GFR remained independently predictive of endothelial dysfunction in our patients whereas hs-CRP was not associated with circulating adhesion molecules. In support of an important role of oxidant stress in cardiovascular disease complicating chronic kidney disease, both vitamin E 800 U daily and acetylcysteine 600 mg twice daily were shown to decrease cardiovascular events in randomized controlled trials in haemodialysis patients [45]. Whether such interventions could decrease cardiovascular disease in RA may deserve further study.
Associations between biomarkers of endothelial dysfunction and common carotid artery intima–media thickness and plaque in RA patients
In a previous study conducted by Wallberg-Jonsson and coworkes [23] in 39 RA patients, ICAM-1 and selectin concentrations were found to be related to ultrasonographically detected CCA and femoral artery plaque as well as to haemostatic factors of endothelial origin. We found that VCAM-1 was associated with ultrasonographically determined CCA IMT and plaque. Taken together, these data further support the contention that circulating adhesion molecules are linked to cardiovascular disease in RA.
Study limitations
We assessed cardiovascular risk comprehensively and our results are in keeping with previously reported paradigms of cardiovascular disease in RA [4,5,31]. However, our findings must be reproduced in a longitudinal study and investigations of a larger cohort is likely to reveal more independent associations between biomarkers of endothelial dysfunction and cardiovascular risk factors.
Conclusion
We found that the biomarkers of endothelial dysfunction VCAM-1, ICAM-1 and ELAM-1 were higher in 74 RA patients than in 80 healthy control individuals. In multivariable regression models these differences could be accounted for by nontraditional cardiovascular risk factors (high hs-CRP, IL-1, IL-6 and TNF-α) and unfavourable traditional cardiovascular risk factor profiles in RA patients. IL-6, rheumatoid factor titre and low GFR predicted endothelial dysfunction, as assessed by biomarkers, independent of traditional cardiovascular risk factor in the 74 RA patients. VCAM-1 was associated with CCA atherosclerosis in RA. Those DMARDs that are most effective at suppressing both cytokine and rheumatoid factor production may also be most effective in protecting against cardiovascular disease in RA. In addition, interventions aimed at preserving renal function may need to be considered in cardiovascular disease prevention in RA.
Abbreviations
CCA = common carotid artery; DMARD = disease-modifying antirheumatic drug; ELAM = endothelial leucocyte adhesion molecule; GFR = glomerular filtration rate; hs-CRP = high-sensitivity C-reactive protein; ICAM = intercellular adhesion molecule; IL = interleukin; IMT = intima–media thickness; RA = rheumatoid arthritis; TNF = tumour necrosis factor; VCAM = vascular cell adhesion molecule.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
PD conceived the study, collected the data, performed the statistical analysis and drafted the manuscript. BJ participated in the study design, interpretation of the data and drafting the manuscript. SS conducted the immunoassays.
Acknowledgements
The authors thank Dr Milton Tobias for reading the radiographs, and Ms Belinda Stevens for carrying out the ultrasonographic carotid artery evaluations. The study was supported in part by the South African Circulatory Disorders Research Fund.
Figures and Tables
Table 1 Cardiovascular risk factors in RA patients and control individuals
Risk factors Controls Patients P
Traditional risk factors
Age (years) 44 (20–87) 57 (27–81) <0.0001
Women (n [%]) 64 (80) 64 (86) 0.3
Caucasian:asian (n:n) 87: 6 71: 9 0.3
Smokers (n [%]) 25 (31) 17 (23) 0.3
Smoking (cigarettes/day) 0 (0–50) 0 (0–40) 0.5
Alcohol users (n [%]) 37 (47) 26 (35) 0.2
Alcohol (units/week) 0 (0–40) 0 (0–35) 0.4
Exercisers (n [%]) 43 (54) 19 (26) 0.0004
Exercise (hours/week) 1 (0–20) 0 (0–7) 0.0002
Diabetes (n [%]) 0 (0) 5 (7) 0.02
Body mass index (kg/m2) 24.3 (17.9–33.8) 23.7 (17.9–38.3) 1.0
Waist (cm) 81 (60–109) 85 (66–120) 0.01
SBP (mmHg) 115 (83–150) 123 (98–164) <0.0001
DBP (mmHg) 71 (47–100) 82 (67–109) <0.0001
Total cholesterol (mmol/l) 5.3 (2.1–7.9) 5.1 (3.5–7.5) 0.4
HDL-cholesterol (mmol/l) 1.5 (0.4–2.3) 1.6 (0.8–2.6) 0.4
Triglycerides (mmol/l) 1.0 (0.5–4.7) 1.2 (0.5–2.6) 0.04
Nontraditional risk factors
hs-CRP (mg/l) 1.6 (0.3–8.9) 10.8 (0.3–256) <0.0001
IL-1 (pg/ml) 0.7 (0–42.7) 3.5 (0.1–323) <0.0001
IL-6 (pg/ml) 0.8 (0.1–8.2) 5.4 (0.5–186.3) <0.0001
TNF-α (pg/ml) 0.2 (0.2–32.6) 3.0 (0.3–93.2) <0.0001
Cytokine suppressant therapy
DMARD use (n [%]) 0 (0) 56 (76) <0.0001
Prednisone use (n [%]) 0 (0) 11 (15) 0.0003
Results are expressed as median (range) unless indicated otherwise. Data were analyzed using the Mann–Whitney U-test (continuous variables) or the χ2 test (dichotomous variables). DBP, diastolic blood pressure; DMARD, disease-modifying antirheumatic drug; HDL, high-density lipoprotein; hs-CRP, high-sensitivity C-reactive protein; IL, interleukin; RA, rheumatoid arthritis; SBP, systolic blood pressure; TNF, tumour necrosis factor.
Table 2 Biomarkers of endothelial dysfunction in rheumatoid arthritis patients and control individuals
Marker Controls Patients Unadjusted P (multivariable adjusted)
Model 1a Model 2b Model 3c Model 4d
VCAM-1 (pg/ml) 506 (253–1067) 747 (391–2077) <0.0001 <0.0001 <0.0001 <0.0001 0.08
ICAM-1 (pg/ml) 231 (82–857) 366 (135–993) <0.0001 0.0002 <0.0001 0.0005 0.08
ELAM-1 (pg/ml) 48 (7–178) 58 (12–149) <0.0001 0.02 0.02 0.7 0.08
Results are expressed as median (range). Unadjusted comparisons were done using the Mann–Whitney U-test and adjustments for potentially explanatory variables were made using logistic regression models. aAdjusted for disease-modifying antirheumatic drug and prednisone use. bAdjusted for traditional risk factors (age; sex; race; smoking, alcohol and exercising status; diabetes; waist; systolic blood pressure; total cholesterol; high-density lipoprotein cholesterol; triglycerides). cAdjusted for nontraditional risk factors (high sensitivity C-reactive protein, interleukin IL-1, IL-6 and tumour necrosis factor-α). dAdjusted for traditional and nontraditional risk factors. ELAM, endothelial leukocyte adhesion molecule; ICAM, intercellular adhesion molecule; VCAM, vascular adhesion molecule.
Table 3 Spearman correlations among biomarkers and potential cardiovascular risk factors
Cardiovascular risk factor VCAM-1 (pg/ml) ICAM-1 (pg/ml) ELAM-1 (pg/ml)
Demographic
Age (years) 0.279 0.149 0.148
Lifestyle
Smoking (cigarettes/day) 0.209 0.253 0.062
Alcohol (units/week) 0.209 0.112 -0.072
Exercise (hours/week) 0.034 0.085 -0.055
Systemic inflammation
hs-CRP (mg/l) 0.289 0.189 0.217
ESR (mm/hour) 0.217 0.175 0.195
Disease duration (years) 0.138 0.069 0.079
Disease severity
Radiographic score 0.157 -0.012 0.092
Rheumatoid factor (IU/ml) 0.319** 0.363** 0.290
Cytokines
TNF-α (pg/ml) 0.107 0.204 0.019
IL-1 (pg/ml) -0.123 -0.030 -0.120
IL-6 (pg/ml) 0.248 0.298* 0.359**
Drug therapy
Current prednisone dose (mg/day) 0.156 0.152 0.279
Cumulative prednisone dose (mg) 0.042 0.008 0.071
Cumulative pulsed MP (mg) -0.051 -0.155 -0.096
Metabolic syndrome
Waist circumference (cm) 0.036 0.249 0.127
SBP (mmHg) 0.159 0.011 0.019
DBP (mmHg) -0.042 -0.073 -0.097
HDL-cholesterol (mmol/l) -0.093 -0.103 -0.091
Triglycerides (mmol/l) 0.186 0.181 0.079
QUICKI -0.088 -0.035 -0.019
Uric acid (mmol/l) 0.208 0.271 0.260
Others
Haemoglobin (g/dl) -0.139 -0.130 -0.093
Leucocytes (× 106/l) -0.017 -0.096 0.199
Polymorphonuclear cells (×106/l) 0.041 -0.092 0.212
Platelets (109/l) -0.025 -0.142 0.104
Homocysteine (μmol/l) 0.124 0.236 0.277
Thyrotropin (μIU/ml) -0.132 -0.007 0.020
LDL-cholesterol (mmol/l) -0.139 -0.078 0.175
Apolipoprotein(a) (mg/l) 0.097 0.004 0.063
GFR (ml/min) -0.285* -0.144 -0.121
Urinary albumin/creatinine (mg/mmol) -0.182 -0.163 -0.031
DBP, diastolic blood pressure; ELAM, endothelial leukocyte adhesion molecule; ESR, erythrocyte sedimentation rate; GFR, glomerular filtration rate; HDL, high-density lipoprotein; hs-CRP, high-sensitivity C-reactive protein; ICAM, intercellular adhesion molecule; IL, interleukin; LDL, low-density lipoprotein; MP, methylprednisolone; QUICKI, Quantitative Insulin Sensitivity Check Index; SBP, systolic blood pressure; TNF, tumour necrosis factor; VCAM, vascular adhesion molecule. *P ≤ 0.01; **P < 0.006.
Table 4 Number of patients and biomarkers of endothelial dysfunction by sex, race, diabetes and medications used
Number VCAM-1 (pg/ml) ICAM-1 (pg/ml) ELAM-1 (pg/ml)
Sex
Female 64 750 (391–2073) 356 (135–993) 58 (12–144)
Male 10 698 (463–1814) 480 (235–805) 60 (38–149)
Race
Caucasian 68 750 (391–2073) 366 (135–993) 61 (12–149)
Asian 6 572 (463–971) 352 (235–699) 49 (35–82)
Diabetes
Yes 5 674 (554–1338) 376 (235–455) 55 (52–129)
No 69 749 (391–2073) 365 (135–993) 59 (12–149)
COX-2 inhibitor use
Yes 20 689 (475–2073) 371 (145–749) 79 (25–129)
No 54 750 (391–2001) 363 (135–993) 55 (12–149)
Traditional NSAID use
Yes 19 766 (493–2001) 366 (135–993) 68 (16–149)
No 55 703 (391–2073) 365 (145–761) 55 (12–129)
Aspirin use
Yes 6 719 (463–2073) 346 (219–595) 40 (23–94)
No 68 747 (391–2001) 366 (135–993) 62 (12–149)
Oestrogen use
Yes 24 673 (445–2001) 311 (181–993) 53 (12–83)
No 50 786 (391–2073) 389 (135–805) 64 (20–149)
DMARD use
Yes 56 726 (391–2073) 368 (145–805) 56 (12–149)
No 18 816 (567–2001) 358 (135–993) 64 (34–115)
Prednisone use
Yes 11 930 (391–2073) 455 (181–761) 94 (23–129)
No 63 724 (445–2001) 361 (135–993) 55 (12–149)
Antihypertensive agent use
Yes 23 744 (569–2073) 414 (219–805) 66 (16–149)
No 51 758 (391–2001) 350 (135–993) 54 (12–137)
Results for biomarkers are expressed as median (range). Data were analyzed using the Mann–Whitney U test. No comparisons were significant at P ≤ 0.01. COX-2, cyclooxygenase-2; DMARD, disease-modifying antirheumatic drug; ELAM, endothelial leukocyte adhesion molecule; ICAM, intercellular adhesion molecule; NSAID, nonsteroidal anti-inflammatory drug; VCAM, vascular adhesion molecule.
Table 5 Partial correlation coefficients between IL-6, rheumatoid factor and glomerular filtration rate, and biomarkers of endothelial dysfunction in rheumatoid arthritis patients
Marker IL-6 Rheumatoid factor GFR
Ra P Ra P Ra P
VCAM-1 (pg/ml) 0.262 0.04 0.345 0.006 -0.299 0.02
ICAM-1 (pg/ml) 0.263 0.04 0.373 0.003 -0.351 0.005
ELAM-1 (pg/ml) 0.430 0.0005 0.222 0.09 -0.223 0.07
aThe correlation coefficients shown are controlled for traditional risk factors (age; sex; race; smoking, alcohol and exercising status; diabetes; waist; systolic blood pressure; total cholesterol; high-density lipoprotein cholesterol; triglycerides) in multivariable regression models. ELAM, endothelial leukocyte adhesion molecule; GFR, glomerular filtration rate; ICAM, intercellular adhesion molecule; IL, interleukin; VCAM, vascular adhesion molecule.
==== Refs
Solomon DH Karlson EW Rimm EB Cannuscio CC Mandl LA Manson JE Stampfer MJ Curhan GC Cardiovascular morbidity and mortality in women diagnosed with rheumatoid arthritis Circulation 2003 107 1303 1307 12628952 10.1161/01.CIR.0000054612.26458.B2
Wolfe F Freundlich B Straus WL Increase in cardiovascular and cerebrovascular disease prevalence in rheumatoid arthritis J Rheumatol 2003 30 36 40 12508387
del Rincon ID Williams K Stern MP Freeman GL Escalante A High incidence of cardiovascular events in a rheumatoid arthritis cohort not explained by traditional cardiac risk factors Arthritis Rheum 2001 44 2737 2745 11762933 10.1002/1529-0131(200112)44:12<2737::AID-ART460>3.0.CO;2-#
Sattar N McCarey DW Capell H McInnes IB Explaining how 'high-grade' systemic inflammation accelerates vascular risk in rheumatoid arthritis Circulation 2003 108 2957 2963 14676136 10.1161/01.CIR.0000099844.31524.05
Dessein PH Joffe BI Stanwix AE Inflammation, insulin resistance, and aberrant lipid metabolism as cardiovascular risk factors in rheumatoid arthritis J Rheumatol 2003 30 1403 1405 12858433
Wallberg-Jonsson S Ohman ML Dahlqvist SR Cardiovascular morbidity and mortality in patients with seropositive rheumatoid arthritis in Northern Sweden J Rheumatol 1997 24 445 451 9058647
Kumeda Y Inaba M Goto H Nagata M Henmi Y Furumitsu Y Ishimura E Inui K Yutani Y Miki T Shoji T Nishizawa Y Increased thickness of the arterial intima-media detected by ultrasonography in patients with rheumatoid arthritis Arthritis Rheum 2002 46 1489 1497 12115178 10.1002/art.10269
Nagata-Sakurai M Inaba M Goto H Kumeda y Furumitsu Y Inui K Koyama H Emoto M Ishimura E Shoji T Nishizawa Y Inflammation and bone resorption as independent factors of accelerated arterial wall thickening in patients with rheumatoid arthritis Arthritis Rheum 2003 48 3061 3067 14613267 10.1002/art.11327
Belcaro G Nicolaides AN Ramaswami G Cesarone MR De Sanctis M Incandela L Ferrari P Geroulakos G Barsotti A Griffin M Carotid and femoral ultrasound morphology screening and cardiovascular events in low risk subjects: a 10-year follow-up study (the CAFES-CAVES study (1)) Atherosclerosis 2001 156 379 387 11395035 10.1016/S0021-9150(00)00665-1
Bonetti PO Lerman LO Lerman A Endothelial dysfunction. A marker of atherosclerotic risk Arterioscler Thromb Vasc Biol 2003 23 168 175 12588755 10.1161/01.ATV.0000051384.43104.FC
Gonzalez-Juanatey C Testa A Garcia-Castelo A Garcia-Porrua C Llorca J Vidan J Hajeer AH Ollier WE Mattey DL Gonzalez-Gay MA HLA-DRB1 status affects endothelial function in treated patients with rheumatoid arthritis Am J Med 2003 114 647 652 12798452 10.1016/S0002-9343(03)00133-5
Vaudo G Marchesi S Gerli R Allegrucci R Giordano A Siepi D Pirro M Shoenfeld Y Schillaci G Mannarino E Endothelial dysfunction in young patients with rheumatoid arthritis and low disease activity Ann Rheum Dis 2004 63 31 35 14672888 10.1136/ard.2003.007740
Hurlimann D Forster A Noll G Enseleit F Chenevard R Distler O Bechir M Spieker LE Neidhart M Michel BA Anti-tumor necrosis factor-alpha treatment improves endothelial function in patients with rheumatoid arthritis Circulation 2002 106 2184 2187 12390945 10.1161/01.CIR.0000037521.71373.44
Bergholm R Leirisalo-Repo M Vehkavaara S Makimattila S Taskinen MR Yki-Jarvinen H Impaired responsiveness to NO in newly diagnosed patients with rheumatoid arthritis Arterioscler Thromb Vasc Biol 2002 22 1637 1641 12377742 10.1161/01.ATV.0000033516.73864.4E
Van Doornum S McColl G Jenkins A Screening for atherosclerosis in patients with rheumatoid arthritis Arthritis Rheum 2003 48 72 80 12528106 10.1002/art.10735
Meigs JB Hu FB Rifai N Manson JE Biomarkers of endothelial dysfunction and risk of type 2 diabetes mellitus JAMA 2004 291 1978 1986 15113816 10.1001/jama.291.16.1978
Ponthieux A Herbeth B Droesch S Haddy N Lambert D Visvikis S Biological determinants of serum ICAM-1, E-selectin, P-selectin and L-selectin levels in healthy subjects: the Stanislas study Atherosclerosis 2004 172 299 308 15019540 10.1016/j.atherosclerosis.2003.11.003
Hwang S-J Ballantyne CM Sharrett AR Smith LC Davis CE Gotto AM JrBoerwinkle E Circulating adhesion molecules VCAM-1, ICAM-1, and E-selectin in carotid atherosclerosis and incident coronary heart disease cases Circulation 1997 96 4219 4225 9416885
De Caterina R Basta G Lazzerini G Dell'Omo G Petrucci R Morale M Carmassi F Pedrinelli R Soluble vascular cell adhesion molecule-1 as a biohumoral correlate of atherosclerosis Arterioscler Thromb Vasc Biol 1997 17 2646 2654 9409238
O'Malley T Ludlam CA Riemermsa RA Fox KA Early increase in levels of soluble inter-cellular adhesion molecule-1 (sICAM-1): potential risk factor for the acute coronary syndromes Eur Heart J 2001 22 1226 1234 11440495 10.1053/euhj.2000.2480
Malik I Danesh J Whincup P Bhatia V Papacosta O Walker M Lennon L Thomson A Haskard D Soluble adhesion molecules and prediction of coronary heart disease: a prospective study and meta-analysis Lancet 2001 358 971 976 11583751 10.1016/S0140-6736(01)06104-9
Solomon DH Curhan GC Rimm EB Cannuscio CC Karlson EW Cardiovascular risk factors in women with and without rheumatoid arthritis Arthritis Rheum 2004 50 3444 3449 15529391 10.1002/art.20636
Wallberg-Jonsson S Cvetkovic JT Sundqvist KG Lefvert AK Rantapaa-Dahlqvist S Activation of the immune system and inflammatory activity in relation to markers of atherothrombotic disease and atherosclerosis in rheumatoid arthritis J Rheumatol 2002 29 875 882 12022343
Cush JJ Rothlein R Lindsley HB Mainolfi EA Lipsky PE Increased levels of circulating intercellular adhesion molecule 1 in the sera of patients with rheumatoid arthritis Arthritis Rheum 1993 36 1098 1102 8102057
Arnett FC Edworthy SM Bloch DA McShane DJ Fries JF Cooper NS Healey LA Kaplan SR Liang MH Luthra HS The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis Arthritis Rheum 1988 31 315 324 3358796
Dessein PH Joffe BI Veller MG Stevens BA Tobias M Reddi K Stanwix AE Traditional and non-traditional cardiovascular risk factors are associated with atherosclerosis in rheumatoid arthritis J Rheumatol 2005 32 435 442 15742434
Dessein PH Joffe BI Stanwix AE Subclinical hypothyroidism is associated with insulin resistance in rheumatoid arthritis Thyroid 2004 14 443 446 15242571 10.1089/105072504323150750
Veale DJ Maple C Kirk G McLaren M Belch JJ Soluble cell adhesion molecules-P-selectin and ICAM-1, and disease activity in patients receiving sulphasalazine for active rheumatoid arthritis Scand J Rheumatol 1998 27 296 299 9751471 10.1080/030097498442415
El M Ashour S Moustafa H Ahmed I Altered levels of soluble adhesion molecules in patients with rheumatoid arthritis complicated by peripheral neuropathy J Rheumatol 2002 29 57 61 11824972
Dessein PH Stanwix AE Joffe BI Cardiovascular risk in rheumatoid arthritis versus osteoarthritis: acute phase response related decreased insulin sensitivity and high-density lipoprotein cholesterol as well as clustering of metabolic syndrome features in rheumatoid arthritis Arthritis Res 2002 4 R5 12223108 10.1186/ar428
Van Doornum S McColl G Wicks IP Accelerated atherosclerosis. An extraarticular feature of rheumatoid arthritis? Arthritis Rheum 2002 46 862 783 11953961 10.1002/art.10089
Wolfe F Comparative usefulness of C-reactive protein and erythrocyte sedimentation rate in patients with rheumatoid arthritis J Rheumatol 1997 24 1477 1485 9263138
Feldmann M Brennan FM Maini RN Role of cytokines in rheumatoid arthritis Annu Rev Immunol 1996 14 397 440 8717520 10.1146/annurev.immunol.14.1.397
Papanicolaou DA Wilder RL Manolagas SC Chrousos GP The pathophysiologic roles of interleukin-6 in human disease Ann Intern Med 1998 128 127 137 9441573
Duff GW Cytokines and acute phase proteins in rheumatoid arthritis Scand J Rheumatol Suppl 1994 100 9 19 7532320
Ou LS See LC Wu CJ Kao CC Lin YL Huang JL Association between serum inflammatory cytokines and disease activity in juvenile idiopathic arthritis Clin Rheumatol 2002 21 52 56 11954886 10.1007/s100670200012
Maini RN Feldman M How does infliximab work in rheumatoid arthritis? Arthritis Res 2002 22 28 10.1186/ar549
Barrera P Boerbooms AM Demacker PN van de Putte LB Gallati H van der Meer JW Circulating concentrations and production of cytokines and soluble receptors in rheumatoid arthritis patients: effects of a single dose methotrexate Br J Rheumatol 1994 33 1017 1024 7981988
Rattazzi M Puato M Faggin E Bertipaglia B Zambon A Pauletto P C-reactive protein and interleukin-6 in vascular disease: culprits or passive bystanders? J Hypertens 2003 21 1787 1803 14508181 10.1097/00004872-200310000-00002
George J Afek A Gilburd B Shoenfeld Y Harats D Cellular and humoral immune responses to heat shock protein 65 are both involved in promoting fatty-streak formation in LDL-receptor deficient mice J Am Coll Cardiol 2001 38 900 905 11527651 10.1016/S0735-1097(01)01440-1
Newkirk MM Rheumatoid factors: what do they tell us? J Rheumatol 2002 29 2034 2040 12375308
Takemura S Klimiuk PA Braun A Goronzy JJ Weyand CM T cell activation in rheumatoid synovium is B cell dependent J Immunol 2001 167 4710 4718 11591802
Edwards JC Szczepariski L Szechinski J Filipowicz-Sosnowska A Emery P Close DR Stevens RM Shaw T Efficacy of B-cell-targeted therapy with rituximab in patients with rheumatoid arthritis N Engl J Med 2004 350 2572 2581 15201414 10.1056/NEJMoa032534
Aubry MC Riehle DL Edwards WD Maradit-Kremers H Roger VL Sebo TJ Gabriel SE B-lymphocytes in plaque and adventitia of coronary arteries in two patients with rheumatoid arthritis and coronary atherosclerosis: preliminary observations Cardiovasc Pathol 2004 13 233 236 15210141 10.1016/j.carpath.2004.02.005
Sarnak MJ Levey AS Schoolwerth AC Coresh J Culleton B Hamm LL McCullough PA Kasiske BL Kelepouris E American Heart Association Councils on Kidney in Cardiovascular Disease, High Blood Pressure Research, Clinical Cardiology, and Epidemiology and Prevention: Kidney disease as a risk factor for development of cardiovascular disease Circulation 2003 108 2154 2169 14581387 10.1161/01.CIR.0000095676.90936.80
| 15899050 | PMC1174955 | CC BY | 2021-01-04 16:02:37 | no | Arthritis Res Ther. 2005 Mar 24; 7(3):R634-R643 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1717 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17191589904610.1186/ar1719Research ArticleAnti-Sa antibodies and antibodies against cyclic citrullinated peptide are not equivalent as predictors of severe outcomes in patients with recent-onset polyarthritis Boire Gilles [email protected] Pierre [email protected] Brum-Fernandes Artur J [email protected] Patrick [email protected] Théophile [email protected] Zhijie J [email protected] Nathalie [email protected] Claude [email protected]énard Henri-A [email protected] Department of Medicine, Division of Rheumatology, University of Sherbrooke, Sherbrooke, Quebec, Canada2 Department of Medicine, Division of Internal Medicine, University of Sherbrooke, Sherbrooke, Quebec, Canada3 Centre de recherches cliniques, Centre hospitalier universitaire de Sherbrooke (CHUS), Sherbrooke, Québec, Canada4 Division of Rheumatology, Department of Medicine, McGill University Health Centre, Montréal, Quebec, Canada5 Laboratoire d'histocompatibilité, Université du Québec, INRS-Institut Armand-Frappier, Laval, Québec, Canada2005 17 3 2005 7 3 R592 R603 10 12 2004 6 1 2005 16 2 2005 18 2 2005 Copyright © 2005 Boire et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The prognostic value of two antibodies targeting citrullinated antigens, anti-Sa and anti-cyclic citrullinated peptide (CCP), present at inclusion, was evaluated prospectively in a cohort of 165 consecutive patients with recent-onset or early polyarthritis (EPA) followed for up to 30 months. Patients were treated according to current Good Clinical Practice standards. Predefined outcomes were severe arthritis and persistent arthritis. At inclusion, a median of 3 months after disease onset, 133 (81%) patients fulfilled at least four American College of Rheumatology criteria for rheumatoid arthritis and 30 (18%) had erosive changes on radiographs of hands and feet. Disease-modifying anti-rheumatic drugs were used in close to 80% of the patients at 30 months. Joint damage increased linearly over time, whereas disease activity declined markedly and remained low at each follow-up. Autoantibodies were identified in 76 (46%) patients: rheumatoid factor (RF) in 68 (41%), anti-CCP in 53 (33%), and anti-Sa in 46 (28%). All three antibodies were correlated, but anti-Sa antibodies best predicted severity at 18 and 30 months. RF and anti-CCP performed less well. For both outcomes, anti-Sa alone performed better than any combination of antibodies. The presence of any autoantibody identified about 50 to 60% of the patients with poor outcomes. In multivariate analysis, anti-Sa (odds ratio (OR) 8.83), the presence of erosions at inclusion (OR 3.47) and increasing age (OR 1.06/year) were significantly associated with severity, whereas RF and anti-CCP were not significant predictors. Persistent arthritis was present in up to 84% of patients; autoantibodies were specific but poorly sensitive predictors of this outcome. We conclude that assays for antibodies against citrullinated antigens differ in their ability to predict poorer outcomes in patients with EPA. In our EPA cohort treated in accordance with current standards, detection of anti-Sa but not of RF or anti-CCP antibodies, in combination with clinical and radiological variables present at the first encounter, allowed the identification of a subgroup of EPA patients suffering more rapid and more severe joint damage over 30 months.
==== Body
Introduction
Rheumatoid arthritis (RA) is a chronic inflammatory arthritis that frequently starts at the peak of productive life and is a major cause of invalidity, morbidity, and premature mortality [1]. RA is characterized by an early inflammatory stage that is frequently responsive to disease-modifying anti-rheumatic drugs (DMARDs) [2]. At a temporally undefined later stage, the RA process evolves towards pannus formation responsible for the joint destruction. Once established, pannus may progress on its own, independently of the apparent response to DMARDs. No available treatment can reverse significant joint damage. These observations gave rise to the notion of a therapeutic 'window of opportunity' during which the rheumatoid process would be more likely to be stopped or retarded [2,3]. This notion is supported by the fact that aggressive treatment of early RA decreases both mortality and long-term invalidity and can increase the rate of long-term remission [4-8]. The identification, early into disease, of those patients likely to evolve rapidly to pannus formation and destructive/disabling arthritis would allow the most cost-effective use of expensive treatments and, reciprocally, would minimize unnecessary exposure of spontaneously remitting patients to the risks of aggressive treatments.
Sufficiently specific and sensitive prognostic markers that could be used with confidence in the individual patient with early or recent-onset polyarthritis (EPA) and recent-onset RA are still lacking [9-11]. For example, even within patients fulfilling the 1987 revised classification criteria for RA of the American College of Rheumatology (ACR; formerly the American Rheumatism Association) [12], chronic arthritis presents wide variations in response to treatments, degree of inflammation, and potential for joint destruction and functional impairment [13]. Classification criteria have even more limited value in predicting outcomes of patients with recent-onset polyarthritis [14]. A second challenge is the frequent occurrence of spontaneous remission in early polyarthritis present for up to 3 to 6 months. This relatively benign evolution is well documented, both in population-based studies [15,16] and in cohort studies of patients with polyarthritis of recent onset [17,18]. As a consequence, clinicians frequently adopt a watch-and-see attitude with patients during the first months of disease, delaying treatment with irreversible detrimental consequences [6].
Recently, antibodies targeting determinants resulting from the deimination of peptidylarginine to peptidylcitrulline residues have been described in the serum of patients with RA [19]. These include antibodies targeting cyclic citrullinated peptide (CCP) [20] and antibodies targeting in vivo citrullinated proteins, such as anti-keratin antibodies, antiperinuclear factor, anti-citrullinated (pro)filaggrin and anti-Sa/citrullinated vimentin [21,22]. The specificity of these antibodies for established RA, either rheumatoid factor (RF)-positive or RF-negative, and their presence in patient sera in the preclinical and early clinical phases of disease have been documented [23,24]. The serum of most patients with diseases other than RA, either RF-positive or RF-negative, does not contain antibodies against citrullinated peptides or proteins [20]. Assays to detect these anti-citrullinated peptides/proteins might therefore be useful to predict poor outcomes in patients with early polyarthritis [25-27], although their added value relative to RF was found to be modest or even controversial [28-31].
The purpose of our study was to assess the prognostic role of anti-Sa and of anti-CCP antibodies, relative to RF, in patients with polyarthritis evaluated and treated early after the onset of disease, irrespective of the fulfillment or not of diagnostic criteria for a specific disease entity. We now present an analysis of our cohort at 30 months of follow-up.
Methods
Patients
Consecutive adult patients with synovitis affecting at least three joints for at least 1 month and less than 12 months and evaluated at the Centre hospitalier universitaire de Sherbrooke (CHUS) were asked to participate to the study. The CHUS is the single regional rheumatology referral center for about 400,000 people, and is the site of practice of six of the seven rheumatologists in the area. Patients with bacterial or crystal-induced arthritis, patients with a defined connective tissue disease (looked for both clinically and by autoantibody testing [32]), patients with systemic vasculitis according to ACR criteria [33], and those who declined or were unable to consent were excluded. Failure to fulfill ACR criteria for RA [12], the presence of skin or nail lesions typical of psoriasis or psoriasis-like lesions, inflammatory bowel disease, clinical features suggestive of a spondylarthropathy [34], signs or symptoms suggestive of a connective tissue disease without fulfillment of specific ACR criteria, and the presence of positive HLA-B27 were not reasons for exclusion. Most of the included patients were regularly followed by rheumatologists and treated with the current approach of early and intensive treatment with DMARDs [2,35]. Patients, rheumatologists, coordinating nurses and treating physicians remained blinded to the patients' HLA-DR, anti-CCP and anti-Sa status. Serum and DNA materials were coded to ensure confidentiality and blinding of investigators. All patients gave their informed consent and the ethics review board of the CHUS approved the study.
Disease variables
A rheumatologist completed a structured physical examination, including 68 tender and 66 swollen-joint count assessments and identification of extra-articular manifestations. A trained nurse coordinator performed a structured interview at the inclusion visit and at each of the follow-up visits scheduled at 18, 30, 42 and 60 months after the onset of inflammatory arthritis. The time of onset of arthritis was assumed to be the month during which the patient indicated that joint symptoms/signs had appeared or, in patients with previous musculoskeletal complaints (such as osteoarthritis), at the time that the signs or symptoms of inflammatory arthritis appeared, additive to the previous signs and symptoms. Scheduling the follow-up evaluations relative to the onset of symptoms rather than to the establishment of diagnosis sought to ensure that the cohort was more homogeneous in disease duration on follow-up.
Variables assessed included demographics, tender and swollen joint counts, duration of morning stiffness, use of DMARDs and corticosteroids at and between each visit, modified Health Assessment Questionnaire (M-HAQ) [36], erythrocyte sedimentation rate (according to the Wintrobe method), serum C-reactive protein (CRP; lower limit of detection 3.0 mg/L, upper normal limit 8.0 mg/L), serum RF (latex agglutination, RapiTex RF; Dade Behring Inc, Newark, DE, USA; threshold for positive results set at 40 IU/ml on the basis of clinical experience and confirmed by its optimal prognostic value), genomic typing of the HLA-DR (see below), measure of anti-Sa antibodies (see below) and anti-CCP antibodies (QuantaLite™; Inova Diagnostics, San Diego, CA, USA; in accordance with the manufacturer's recommendations). Standardized radiographs of the hands and feet were obtained at inclusion and at each scheduled assessment. Joint space narrowing and erosions were scored by the Sharp–van der Heijde (SvH) method, with a maximum score of 448 [37]. All radiographs were read in pairs in a known time sequence by a trained reviewer who was blinded to the corresponding clinical information. The smallest detectable difference was shown to be 5 units; this was assumed to be identical to the minimal clinically important difference [38,39]. Functional status was determined at each visit by using validated French-Canadian or English versions of the M-HAQ with a range of 0 to 3 (good functional status to maximal disablement). Disease Activity Score using 28 joints and 3 variables (tender and swollen joint counts and CRP; DAS28-3) was calculated with the appropriate formulas [40]. A DAS28-3 score below 2.6 defined clinical remission [41,42], whereas a score above 3.2 indicated more than mild disease.
Predefined outcomes
Persistent arthritis was defined as the presence of at least one joint with synovitis and/or the current use of DMARDs or at least 10 mg of prednisone equivalent per day [25]. Fulfillment of RA criteria required at least four ACR criteria for RA. Severity required an M-HAQ score of at least 1.0 and/or belonging to the upper third of the SvH radiological score.
Genomic PCR typing at the HLA-DR and HLA-DQ
Genomic DNA was extracted from leukocytes present in 2.5 ml of EDTA-treated whole blood using Wizard DNA extraction kit (Promega Corporation, Madison, WI, USA), then stored at 4°C. Genomic typing was performed by using PCR with sequence-specific primers specific for HLA class II molecules [43,44]. Primer sets for low-resolution typing of HLA-DR and HLA-DQ antigens, and high-resolution typing of DRB1*01, DRB1*04, DRB1*14 and DQB1 loci were from Pel-Freez Clinical Systems (Brown Deer, WI, USA). After amplification, PCR products (10 μl) were resolved in a 2% agarose gel containing ethidium bromide and detected by ultraviolet transillumination. Amplification patterns were interpreted in accordance with the manufacturer's instructions. HLA-DRB1 alleles encoding the amino acid sequences QRRAA (DRB1*0101, DRB1*0102, DRB1*0105, DRB1*0404, DRB1*0405 and DRB1*0408), QKRAA (DRB1*0401 and DRB1*0409) and RRRAA (DRB1*1001) were considered to encode the shared epitope.
Anti-Sa/citrullinated vimentin
The Sa antigen was first detected in human spleen/placenta extracts by western blotting (WB) [21] and subsequently identified as citrullinated vimentin [22]. Most citrullinated arginine-rich proteins are adequate but not absolutely equivalent reagents for the detection of antibodies against citrullinated epitopes, as discussed in [22]. Because access to human tissues and consistent extraction of the Sa antigen from tissues proved problematic, and commercial human vimentin is very expensive, we developed standardized and less expensive assays. We screened several cell lines and identified the ECV 304 endothelial cell line as rich in both vimentin and peptidylarginine deiminases (PADIs). Auto-citrullinated extracts from ECV 304 cells (see below) paralleled human placenta extracts for anti-Sa detection in WB. We then used this ECV 304-based WB assay to monitor the performance of an in-house enzyme-linked immunosorbent assay (ELISA), designed with in vitro citrullinated bovine myelin basic protein (MBP). MBP and vimentin have similarly high proportions of arginine. Most positive results in MBP-ELISA correlated with positive anti-Sa results on the original WB method using human placenta.
Anti-Sa WB assay
ECV 304 cells were cultured at 37°C under 5% CO2 in IMDM (Sigma Chemical Co., St Louis, MO, USA) enriched with 10% fetal bovine serum containing 3.024 g/L sodium bicarbonate (pH 7.2). At confluence, cells were trypsinized, washed, incubated for 5 min on ice in 1 ml of buffer containing 40 mM Tris-HCl pH 7.5, 1 mM EDTA pH 8.0, 0.150 NaCl, and mechanically detached. The resultant cell pellet was submitted to three freeze-thaw cycles, and the supernatant was clarified by centrifugation. The concentration of protein was measured by the Bradford assay (Bio-Rad), and the extracts were stored at -80°C. In preparation for WB, the extracts (2.5 μg per tested serum) were auto-citrullinated for 3 hours at 37°C in 50 mM citrullination buffer (Tris-HCl pH 7.5, 10 mM CaCl2, 10 mM dithiothreitol). The reaction was stopped with EDTA (100 mM final concentration). An equal volume of 2× WB loading buffer was added to the citrullinated extract for long-term storage in aliquots at -20°C. WB was performed as described [21], with ECV 304 native and citrullinated extracts being run in parallel. Sera were tested in duplicate at a dilution of 1:100 and results are expressed as either positive or negative.
Anti-Sa ELISA
Bovine MBP (Sigma) was citrullinated in vitro by incubation for 3 hours at 37°C in citrullination buffer containing 0.2 U rabbit PADI 2 (Sigma) per μg of MBP. Microtiter plates (96 wells; Nunc Maxisorp, WWR International Ltd., Mississauga, ON, Canada) were coated overnight at 4°C with native or citrullinated bovine MBP at 1 μg per well. The plates were blocked for 1 hour at room temperature (20 to 25°C) with PBS containing 1% (w/v) BSA, then washed three times with PBS containing 0.05% Tween 20 (PBST). The plates were incubated for 1 hour at room temperature with 100 μl per well of human serum diluted 1:300 in PBS containing 1% (w/v) BSA (PBS-BSA), and then washed three times with PBST. Bound human IgG was detected with a goat anti-human IgG, alkaline phosphatase-labeled conjugate (Jackson ImmunoResearch Laboratories Inc., West Grove, PA, USA) diluted 1:1000 in PBS-BSA. The reaction was detected with 1 mg/ml p-nitrophenyl phosphate (Sigma) in substrate buffer (10% diethanolamine, 0.5 mM MgCl2, pH 9.8) for 20 min at room temperature. The absorbance, A, was read at 405 nm. Each sample was run in duplicate and in parallel on native MBP (N), on citrullinated MBP (C), and on BSA (B). Results were given by the equation A405 = AC – AN. A positive test was defined as A405 > 0.2. A known positive control serum was used on each plate for reagent quality control. The BSA control was used to detect spurious intrinsic individual serum stickiness for which some RA sera are notorious. When a serum gave high A405 when used on BSA alone, a positive anti-Sa ELISA result had to be confirmed on WB to be reported as positive. All the anti-Sa assays described here were developed and performed on coded sera at the McGill University Autoimmune Research Laboratory.
Statistical analysis
The principal independent variables were anti-Sa (MBP-ELISA), anti-CCP (positive if more than 20 U/ml), and RF (positive if at least 40 IU/ml) antibodies. The degree of association between anti-Sa and anti-CCP assays was measured by using the phi coefficient. Sensitivity, specificity and positive likelihood ratio (LR) were calculated for each of the three outcome measures at 18 and 30 months into disease and for each of the 13 following putative prognostic markers present at the inclusion visit: anti-Sa, anti-CCP, RF, sex, age (in years), duration of disease (in months), duration of morning stiffness (positive when at least 1 hour), symmetry of joint swelling, number of joints with synovitis, M-HAQ score (positive when at least 1.0), RA-associated HLA-DRB1 alleles ('shared epitope'), CRP levels (positive when higher than 8.0 mg/L), and an erosion component of the SvH score of at least 5. The predictive ability of anti-Sa, anti-CCP, and RF in the early identification of patients with each outcome was also studied by using the area under a receiver operating characteristics curve. Confidence intervals were calculated for these estimates. After these estimations, a logistic regression approach was used to estimate the independent contribution of RF, anti-Sa, and anti-CCP present at inclusion to predict the development of outcomes at 30 months. Each autoantibody was evaluated separately and in combination, after inclusion in the multivariate statistical model of prognostic factors listed above. A final statistical model was constructed with the stepwise logistic regression approach. By using the variables included in the final model, with and without the inclusion of anti-Sa and anti-CCP antibodies, the odds ratio (OR) was calculated to estimate the contribution of each marker adjusted for the other markers in the prediction of the severity outcome.
Results
Patient characteristics at study entry
Up to May 2004, 260 patients agreed to participate and were included. Eleven additional patients declined to participate or had very incomplete inclusion data and were excluded. At this time, 165 patients had reached the 30-month evaluation mark. Of these patients,5 had died and 11 failed to come to their scheduled appointment. We therefore report on the 149 patients (retention rate 90.3%) with a complete 30-month assessment.
The baseline characteristics at entry of the 165 patients are summarized in Table 1. The 16 patients who died or who were lost to follow-up did not differ significantly on any of the inclusion variables from the 149 who were followed up (data not shown). Men (42.4% of our patients) were slightly older than women (61.7 versus 58.0 years). These patients had a disease of short duration (mean 4.4 months, median 3 months). Nevertheless, at inclusion, 80.6% of the patients already fulfilled at least four ACR criteria for RA, 47% were functionally disabled (M-HAQ at least 1.0), and 18% had significant erosive changes (SvH erosion score of 5 or more).
The prevalence of autoantibodies was low. RF was found in 68 (41.2%) patients, anti-CCP in 53 (33.1%), and anti-Sa in 46 (28.0%). As reported previously [20,29], the presence of anti-CCP, RF, and anti-Sa was moderately to highly correlated (Kendall's tau-b > 0.6 for all comparisons). A total of 76 sera contained either one of RF, anti-CCP, or anti-Sa antibodies, and 35 sera contained all three. RF and anti-CCP co-existed in 46 patients, RF and anti-Sa in 41, and anti-Sa and anti-CCP in 39. This means that, despite a good correlation between assays, 13 (19.1%) of the 68 sera with RF did not contain antibodies against citrullinated antigens, CCP or Sa. Interestingly, the degree of association between anti-Sa and anti-CCP (titer higher than 21 U/ml; phi coefficient 0.6881) was very similar to the degree of association between anti-Sa and high titer anti-CCP only (titer higher than 100 U/ml; phi coefficient 0.6339). This absence of variation of the degree of association with increasing titers suggests that anti-Sa and anti-CCP assays differ qualitatively. The shared epitope at the HLA DRB1 locus was present in 77 (47%) of all patients, and in 62 (47%) of those fulfilling criteria for RA at inclusion. Twenty-two patients (13.4%) carried two shared epitope alleles.
Evolution of the clinical status from inclusion to the 18-month and 30-month evaluations
Upon follow-up, the clinical status of the patients improved markedly (Table 2). The DAS28-3, the M-HAQ, and the swollen joint counts all significantly decreased. Despite this apparent control of clinical disease activity relative to inclusion, the radiological SvH score increased steadily at each follow-up, both in the total score and in its erosion component (Fig. 1). A high proportion (77.9%) of patients was still being actively treated with DMARDs at 30 months. Interestingly, despite blinding of their treating physicians, the patients with anti-Sa and anti-CCP were treated as intensively as RF-positive patients. Most treated patients were on 15 to 25 mg/week methotrexate, with about two-thirds (77 of 116) on combination therapies. Despite this liberal use of DMARDs, nine patients with at least one swollen joint were not being treated with DMARDs at 30 months. Those untreated patients represented 13% of patients with synovitis at this follow-up. According to our definition (see above and similar to that in [25]), persistent arthritis was said to be present when swelling was detectable in at least one joint and/or when at least one DMARD or at least a moderate dose of corticosteroids was used at the time of follow-up. Thus, 88% and 84% of the patients were considered as having persistent arthritis at 18 and 30 months, indicating that long-term spontaneous remission was unusual in our EPA cohort, as reported previously in [45]. Close to half (namely 53 of 137 and 58 of 125, at 18 and 30 months, respectively) of the patients labeled 'persistent arthritis' had no swollen joints at each follow-up. The good clinical control of disease in our patients was also evident when using the DAS28-3 score: 76 (48.7%) had a score of less than 2.6 ('remission' [41,42]) at 18 months, and 82 (55.0%) at 30 months. Thus, more than half of our initial cohort was in remission as defined by DAS28-3 at 30 months, arguably the highest proportion of remission in a cohort of patients with EPA [6].
Severe arthritis was present in 56 (38%) patients at 30 months. Our inclusive definition of severity (upper third of SvH score and/or M-HAQ ≥ 1.0) did not make any a priori assumption about the rate of damage progression in our patients, as this should be highly dependent on the characteristics of the patients included and (possibly) the treatments used. The threshold for the upper third of SvH score was calculated as a total score of more than 10 at 18 months and more than 15 at 30 months, that is, about 2% and 3.5%, respectively, of the maximal SvH score. Not surprisingly, functional impairment was the criterion most determinant for selection in the group with severe disease at inclusion, whereas radiological damage became the major reason for selection during follow-up.
Only 28 (18.79%) patients still fulfilled at least four ACR criteria for RA at 30 months. At inclusion or on follow-up, 20 patients fulfilled criteria for inflammatory rheumatic diseases other than RA: 13 had arthritis associated with skin psoriasis (present at inclusion in 12), 3 had benign sarcoid arthritis, 3 had spondylarthropathy, and 1 had scleroderma. Autoantibodies were present at low frequency in these patients: RF in four, anti-Sa in two, and anti-CCP in one. At each of the follow-up visits, more than two-thirds of the patients (namely 101 of 156 and 101 of 149 at 18 and 30 months, respectively) did not fulfill criteria for a specific diagnosis and might thus be classified as undifferentiated arthritis.
Autoantibodies as predictors of poor outcomes
We first determined the sensitivity, specificity and positive LR of the presence of autoantibodies at onset as prognostic markers for predefined outcomes (Table 3). Anti-Sa stood out as the single moderately good marker (positive LR 2.16) to predict the development of severe arthritis. In this regard, anti-CCP (positive LR 1.38) and RF (positive LR 1.50) gave similarly poor results. It is remarkable that, despite the lower prevalence of anti-Sa in our cohort, its sensitivity for the severity outcome was very similar to that of RF and anti-CCP. Any combination of RF and/or anti-CCP with anti-Sa was no better than anti-Sa alone in predicting severe outcomes. In contrast, antibodies present at inclusion were specific but poorly sensitive predictors of persistent arthritis at 30 months. Owing to our inclusive a priori definition (see above and similar to that in [25]), persistent arthritis was observed in 84% of the patients at 30 months. Thus, despite their strong positive LR, antibodies were not clinically helpful in predicting persistence in our cohort at the 30-month evaluation mark. We therefore looked at alternative definitions of persistent arthritis, such as non-remission defined by DAS28 (namely presenting a DAS28-3 score of at least 2.6). Non-remission defined by DAS28-3 was present in 45% of our aggressively DMARD-treated patients at the 30-month evaluation. None of the three antibodies was correlated with non-remission defined by DAS-28 (data not shown), suggesting that the presence of antibodies does not correlate with a poorer response to (conventional) DMARD treatments. Persistence defined as non-remission defined by DAS28-3 was therefore not used further. Finally, no autoantibody correlated with the fulfillment of RA criteria, mostly because of poor specificity. The best serological markers for the fulfillment of criteria for RA at 30 months were RF and anti-Sa, which gave low but very similar positive LR values (1.53 and 1.54, respectively). This result probably reflects the dampening impact of treatment on the clinical activity of arthritis.
In all three evaluated outcomes, RF had a higher sensitivity than anti-Sa, whereas anti-Sa had a higher specificity than RF (except for the fulfillment of RA criteria). Anti-CCP also had a slightly better specificity than RF for predefined outcomes. However, despite their high specificity for established RA, anti-CCP antibodies performed poorly. The same differences between anti-CCP and anti-Sa antibodies, although present, were less marked at the 18-month evaluation (data not shown). As noted above, anti-Sa alone performed better than any combination of antibodies in predicting severity. The conclusions were not significantly different when the analysis was restricted to patients fulfilling RA criteria at inclusion (data not shown). As noted in Table 2, specific avoidance of DMARDs in anti-Sa patients could be discarded as a possible explanation for their more rapid radiographic deterioration, because these patients were as intensively treated as patients with RF.
Multivariate prognostic models of severity
As illustrated in the multiple regression model for severity (Table 4, model A), anti-Sa (OR 8.832) clearly outperformed RF and anti-CCP. Indeed, the presence of anti-Sa conferred an even higher OR than erosion SvH abnormalities (OR 3.472), and was the most powerful independent predictive variable for the severity outcome. The OR of anti-Sa increased after adjustment for all other included variables. Although anti-Sa was the best individual marker for severity, a combination of anti-Sa, an SvH erosion score of at least 5 and increasing age best explained severe disease development. Worth noting is the absence of a HLA-DR 'shared-epitope', disability defined by M-HAQ, and the fulfillment of ACR criteria for RA from the short list of variables present at inclusion that were predictive of the predefined severe outcome.
However, anti-Sa antibodies are not widely available yet, and their presence in the model might have wiped out significant associations of RF and/or anti-CCP with severe arthritis. After the removal of anti-Sa from the analysis, neither RF nor anti-CCP was significantly predictive of severe arthritis (Table 4, model B). At that time, we suspected some heterogeneity among patients with anti-CCP. Receiver operating characteristics curves indicated that a cut-off of 61 U/ml of anti-CCP, rather than 21 U/ml, would yield an improved positive LR of 2.01 for the severity outcome (data not shown). Similarly, after deletion of anti-Sa results and restriction to titers of anti-CCP higher than 60 U/ml, the predictive value of anti-CCP improved but remained below that of RF; neither reached a statistically significant level (Table 4, model C).
Finally, we tested the interaction of erosion SvH score and anti-Sa, the two major initial predictors of severe arthritis on the development of joint damage (Fig. 1). Each of the two predictors influenced the 30-month SvH score, with significant erosive damage at inclusion being the most determinant for the rate of progression of radiological damage. In patients in whom erosion SvH scores were abnormal at inclusion, the rate of yearly increase of the total SvH score was about 10 units. This rate was about 4 units/year in those whose initial erosion SvH scores were less than 5 units. These data support the importance of early damage as evidence for aggressive disease.
In the multiple regression model for persistence at 30 months (data not shown), disease duration for at least 4 months at inclusion was the single significant prognostic marker. No antibody was statistically associated with the outcome of persistence at that time.
Discussion
The main conclusion from our EPA study is that assays targeting different citrullinated antigens have distinct prognostic values for poor outcomes, in addition to their differences in sensitivity and specificity [20]. When present, anti-Sa antibodies are useful markers of poor prognosis in EPA patients, even when rapidly treated intensively with conventional DMARDs. In our EPA population, anti-Sa outperformed anti-CCP antibodies for the prediction of each of the predetermined clinical outcomes, at least up to 30 months into disease. Despite their more specific association with RA [20,46], anti-CCP antibodies did not carry any significant advantage over RF. It might thus be premature to consider replacing RF by anti-CCP antibodies in the clinical evaluation of EPA patients [47]. Differences between RF and anti-Sa antibodies were also significant. As a rule, RF proved to be the most sensitive assay for all outcomes, whereas anti-Sa was the most specific.
At first sight, the lower prevalence of anti-Sa (28% versus 41% for RF) seemed a significant disadvantage. However, the sensitivity of anti-Sa for severe arthritis increased over time in our cohort, as 29%, 39%, and 45% of all severe patients had anti-Sa at inclusion, 18 months, and 30 months, respectively. This observation held true even though anti-Sa positive patients received as aggressive DMARD treatments as the other patients with persistent disease. The lack of statistically significant association of antibodies with persistence was probably due in part to the dual definition of persistence that we used, namely the presence of synovitis and/or being treated. In the absence of a more specific diagnostic test for persistence, we consider this definition to be the best currently available in treated recent-onset arthritis. The major limitation with our definition of persistence was the unexpectedly high (80%) frequency of ongoing use of DMARDs at 30 months. We expect that most patients in protracted remission will slowly taper their DMARDs over the planned follow-up of up to 5 years. The confounding influence of prolonged treatment on our definition of persistence should thus progressively attenuate over these additional 30 months, and the prognostic value of antibodies for persistence might then become assessable.
The second important conclusion is that autoantibodies were present early in only 50 to 60% of the patients who developed any of the predefined poor outcomes at 30 months. Autoantibodies present early into disease therefore characterize a large, but limited, subset of EPA patients with poor outcomes. Detection of autoantibodies in EPA patients must thus be used in combination with additional variables present at inclusion (for example elevated erosion SvH scores and increasing age) to best predict, and possibly prevent, the development of severe disease at 30 months (and beyond).
The third conclusion is that our results should be generally applicable to EPA patients evaluated by rheumatologists. Several reasons support that opinion. First, selection biases were minimized by the strict exclusion of patients with diseases known to have a good prognosis (such as crystal-induced arthritis and monoarthritis) and by a retention rate of 90% at 30 months, without individual missing data. Second, the patients were thoroughly evaluated with widely available tools (except for anti-Sa), early in their disease (median 3 months), and at consistent intervals relative to disease onset. Third, the predefined severity outcome included both functional disability and radiographic damage. Functional impairment and invalidity are the best indicators of severity, but they occur too late into disease to be useful prognostic indicators. Radiographic changes, such as those quantified with the SvH score, are therefore used as surrogates of severity in early disease. Because disability and radiographic damage are poorly correlated in early arthritis [45,48], both contribute in relatively independent ways to the full picture of severity during the first years of chronic arthritis.
Our observations regarding the prognostic usefulness of anti-CCP in EPA patients are somewhat at variance with other reports. The reasons for differences in prognostic value between anti-CCP and anti-Sa assays therefore merit further discussion. A first possible explanation relates to deficiencies of the specific commercial anti-CCP assay used in this paper. That explanation is unlikely. We observed no significant differences in sensitivities or specificities between commercial anti-CCP assays (G. Boire, unpublished observations). Such differences would be unlikely, because all commercially available anti-CCP assays use the same antigen strips [20] and differ only in technical aspects of the ELISA itself. Thus, the prognostic performance of the particular anti-CCP assay we used is likely to be generally applicable to commercial anti-CCP assays.
A second and more likely possibility for the poor performance of anti-CCP might reside in the design of the assay. The objectives of current anti-CCP assays are to attain the maximal sensitivity for patients with established RA or RA-like disease, while maintaining a reasonable specificity. This approach favors a low threshold to report positive results. Because EPA and RA are very heterogeneous diseases, a more appropriate design would be to identify subsets of patients with worse outcomes. In our cohort, most anti-CCP positive sera had high titers (mean 135.6 U/ml; median 116 U/ml) but, significantly, 11 patients had low or moderate titers (60 U/ml and below). The presence of low or moderate titers of anti-CCP at inclusion did not seem to be associated with severe arthritis at 30 months. Indeed, using a cut-off of 61 U/ml afforded the best OR estimate for the development of severe arthritis, although the predictive value of this level of anti-CCP was still not statistically significant (Table 4, model C). It was previously suggested that many non-RA sera with low anti-CCP titers bind similarly to citrullinated and non-modified peptides [49]. These sera would present as false positive anti-CCP results. As a consequence, we suggest that the threshold for positive results of commercial anti-CCP assays should be adjusted upward to increase their prognostic value in EPA patients.
A third explanation for the prognostic discrepancy between anti-CCP and anti-Sa implies hypothetical qualitative differences between the anti-Sa and the anti-CCP assays, as suggested by the moderate degree of association between anti-Sa and anti-CCP results remaining constant across different cut-offs for positive anti-CCP (phi coefficient 0.6881 for titers higher than 21 U/ml, and 0.6339 for titers higher than 100 U/ml). These qualitative differences between assays would translate into genuine consequent associations with outcomes. Antibodies present in sera from RA patients recognize citrullinated peptidic residues in the context of adjoining amino acid residues and of peptide conformation [19,20]. The short synthetic peptides used in anti-CCP ELISA were circularized and covalently bound to plastic strips to circumvent some of the problems inherent to the use of peptides in solid-phase assays [46]. This design might not be appropriate in very early polyarthritis, at a time when the immune response to citrullinated antigens is still being matured in vivo by modified arginine-rich proteins. At present it remains to be confirmed whether citrullinated proteins in general might perform better than CCP in ELISA at identifying severe subsets of EPA patients.
A fourth contributing factor to explain the poor performance of anti-CCP in our cohort is the influence of aggressive treatment on EPA prognosis. An earlier time of introduction, a higher intensity of use, and a more prolonged maintenance of DMARDs, with the objective of controlling disease, distinguish our current cohort from previous EPA cohorts established during the early 1990s. This more aggressive approach is illustrated by the high prevalence of DMARDs still used at 30 months, as well as by the marked decrease in disease activity over time, as assessed by the DAS28-3, the swollen joint counts, and the M-HAQ. Possibly because of this relatively intensive use of DMARDs, more than half of the patients were in clinical remission at 30 months. According to this hypothesis, the presence of anti-CCP antibodies would indeed identify patients with a poor natural course, as suggested in previous EPA or early RA cohorts [25-27,50]. In contrast, anti-CCP would fail to identify patients who do poorly when exposed relatively early to intensive treatment with current DMARDs. Such an effect of aggressive treatment was previously reported to explain the loss of HLA-DR association with severity in aggressively DMARD-treated patients [51]. If this is true for anti-CCP, however, patients with anti-Sa would not exhibit the same generally good response to DMARDs or, alternatively, the association of anti-Sa with severe outcomes would be so large that the use of DMARDs would not completely erase it. Because intensive and effective DMARD treatments capable of slowing the progression of radiological and functional deterioration are increasingly being used in clinical practice, progressively longer periods of observation in larger samples of patients should therefore be required to delineate the true predictive value of potential prognostic markers.
Finally, our data stress the importance of intrinsic aggressiveness of arthritis in causing progressive joint damage. At inclusion, 18% of the patients already demonstrated clinically significant erosive changes. That represents a very early progression to the pannus phase, and an already missed opportunity to intervene during the optimal 'therapeutic window' in these patients. Severe outcomes at 30 months happened mostly in those patients who already had significant erosive joint damage at first evaluation. Early erosive changes are therefore an excellent surrogate marker for an aggressive arthritis. Whether the use of more sensitive imaging techniques such as ultrasound or magnetic resonance imaging [52] would increase the sensitivity of detection without affecting its specificity for poor outcomes remains to be defined [52-54]. In contrast, a serological marker such as anti-Sa antibodies, more highly associated with severe outcomes than RF or anti-CCP, and present at inclusion in a patient without detectable erosions, would probably represent a pre-pannus surrogate marker for aggressive arthritis. Such a marker would be extremely useful in clinical trials and in real-world practice as well, to select patients for appropriate treatment. In that situation, additional independent markers (for example inflammatory markers, possibly genetic markers) will still be needed if we are to attain a more complete prediction of the outcomes in individual EPA patients.
Conclusion
This study reports the first direct comparison of RF, anti-CCP and anti-Sa antibodies as prognostic markers in a cohort of patients with recent-onset inflammatory polyarthritis treated according to current standards. In this cohort, anti-Sa antibodies were present at inclusion in 45% of the patients who subsequently developed severe outcomes at 30 months. Anti-Sa antibodies were the single serological marker independently predictive of poor outcomes and, together with the presence of joint erosions at inclusion and with increasing age, characterized a subgroup of patients with more rapid and more severe joint damage. In contrast, anti-CCP antibodies were not independently associated with severe outcomes in our patients. Using higher thresholds for positive results slightly improved the performance of the anti-CCP assay, but the predictive value of anti-CCP antibodies remained inferior to that of RF. Although anti-Sa and anti-CCP assays both use citrullinated antigens, our data suggest that the two assays differ in their ability to predict poor outcomes in patients treated aggressively early after the onset of inflammatory polyarthritis.
Abbreviations
ACR = American College of Rheumatology; BSA = bovine serum albumin; CCP = cyclic citrullinated peptide; CHUS = Centre hospitalier universitaire de Sherbrooke; CRP = C-reactive protein; DAS28-3 = Disease Activity Score using 28 joints and 3 variables (tender and swollen joint counts and CRP); DMARD = disease-modifying anti-rheumatic drug; ELISA = enzyme-linked immunosorbent assay; EPA = early (or recent-onset) polyarthritis; LR = likelihood ratio; MBP = myelin basic protein; M-HAQ = modified Health Assessment Questionnaire; OR = odds ratio; PADI = peptidylarginine deiminase; PBS = phosphate-buffered saline; RA = rheumatoid arthritis; RF = rheumatoid factor; SvH = Sharp–van der Heijde; WB = western blotting.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
GB designed the research protocol, played a leading role in the design and coordination of the study, and drafted the manuscript. NC performed the statistical analysis and helped to draft the manuscript. PC helped in the design of the research protocol, longitudinally contributed to the coordination of the study, and helped in drafting the manuscript. AJF participated in the performance and the coordination of the study, and helped to draft the manuscript. PL participated in the performance and the coordination of the study, and helped to draft the manuscript. TN participated in the design of the protocol, contributed to the coordination of the study, supervised the statistical analysis, and reviewed the draft manuscript. ZJZ developed the anti-Sa assays and performed a large portion of the assays for the study. CD participated in the design of the protocol and performed the immunogenetic tests. HAM designed the research protocol, contributed to the performance of the study, supervised the development and performance of anti-Sa assays, and helped to draft the manuscript. All authors read and approved the final manuscript.
Acknowledgements
Supported by The Arthritis Society (Grant No. 00/201) and by the Fonds pour la recherche en santé du Québec (FRSQ). We are indebted to our study nurse coordinators, Céline Boulet BSc and Julie Robindaine BSc, and to Marthe Filion for their continuing involvement and dedication. We thank Dr David Hercelin MD DEA, Dr Daniel Myhal MD, and Dr Sophie Roux MD PhD for their contribution to patient recruitment and follow-up. We thank Dr Maximilien Lora PhD for performance of some of the anti-Sa assays. We thank Dr Michel Gingras MD for his contribution to scoring radiographs in accordance with the Sharp–van der Heijde method.
Figures and Tables
Figure 1 Total Sharp–van der Heijde (SvH) score at inclusion and at 18 and 30 months. Mean values and 95% confidence limits are illustrated. The four curves represent groups of patients at inclusion who had an erosion SvH score of at least 5 and positive anti-Sa (15, 14, and 14 patients at each visit) (squares), an erosion SvH score of at least 5 and negative anti-Sa (15, 13, and 13 patients at each visit) (circles), an erosion SvH score of 4 or less and positive anti-Sa (32, 30, and 30 patients at each visit) (triangles), and an erosion SvH score of 4 or less and negative anti-Sa (103, 97, and 92 patients at each visit) (diamonds).
Table 1 Baseline characteristics of the study population (n = 165)
Characteristic Value
No. female (%) 95 (57.6)
Median age, years (range) 58.8 (19–85)
Median duration of symptoms, months (range) 3 (1–12)
No. with disease duration of 3 months or less (%) 87 (52.7)
Median number of joints with synovitis (range) 10 (3–58)
No. fulfilling at least four ACR criteria for RA (%) 133 (80.6)
Mean M-HAQ score (range) 0.93 (0–2.75)
No. with M-HAQ score ≥ 1.0 (%) 77 (47.0)
Median DAS28-3 score (range) 5.01 (2.09–7.85)
No. with RF ≥ 40 IU/mL (%) 68 (41.2)
No. with anti-CCP ≥ 20 U/mL; positive (%) 53 (33.1)
No. with positive anti-Sa antibodies (%) 46 (28.0)
No. with CRP ≥ 8.0 mg/L (%) 119 (72.6)
No. with at least one HLA-DR 'shared epitope' (%) 77 (47.0)
No. with two HLA-DR 'shared epitopes' (%) 22 (13.4)
Median SvH score (range) 4 (0–54)
No. with total SvH score ≥ 10 (%) 31 (18.8)
No. with erosion SvH score ≥ 5 (%) 30 (18.2)
ACR, American College of Rheumatology; CCP, cyclic citrullinated peptide; CRP, C-reactive protein; DAS28-3, Disease Activity Score using 28 joints and 3 variables (tender and swollen joint counts and C-reactive protein); M-HAQ, modified Health Assessment Questionnaire; RA, rheumatoid arthritis; RF, rheumatoid factor; SvH, Sharp–van der Heijde.
Table 2 Clinical assessments and outcomes at inclusion and during follow-up
Measure Inclusion (n = 165) 18 months (n = 156) 30 months (n = 149)
Mean M-HAQ score (median) 0.93 ± 0.64 (0.875) 0.42 ± 0.50 (0.25) 0.35 ± 0.41 (0.25)
No. with M-HAQ ≥ 1.0 (%) 76 (46.1) 18 (11.6%) 17 (11.5%)
Mean DAS28-3 score (median) 5.01 ± 1.27 (4.94) 2.94 ± 1.26 (2.66) 2.81 ± 1.06 (2.46)
No. with DAS28-3 < 2.6 (%) 5 (3.0) 76 (48.7) 82 (55.0)
Mean swollen joint count (median) 12.55 ± 9.57 (10) 2.83 ± 5.51 (0) 2.04 ± 3.64 (0)
Number without synovitis (%) n.a. 72 (46.2) 82 (55.0)
Number with persistent arthritis (%) n.a. 137 (87.8) 125 (83.9)
No. fulfilling criteria for RA (%) 133 (80.6) 36 (23.1) 28 (18.8)
No. with severe arthritis (%) 106 (64.2) 65 (41.7) 56 (37.8)
Mean total SvH score (median) 6.05 ± 8.38 (4) 11.34 ± 13.82 (7) 15.76 ± 18.38 (10)
Mean erosion SvH score (median) 3.02 ± 4.99 (1) 6.94 ± 9.37 (4) 10.31 ± 13.14 (7)
Upper third of the total SvH score >5 >10 >15
No. on DMARDs (%)
Total 31 (18.8) 120 (76.9) 116 (77.9)
RF-positive 13/68 (19.1) 60/63 (95.2) 60/61 (98.4)
Anti-CCP-positive 10/53 (18.9) 49/52 (94.2) 50/51 (98.0)
Anti-Sa-positive 6/46 (13.0) 43/45 (95.6) 44/44 (100.0)
CCP, cyclic citrullinated peptide; DAS28-3, Disease Activity Score using 28 joints and 3 variables (tender and swollen joint counts and C-reactive protein); M-HAQ, modified Health Assessment Questionnaire; n.a., not applicable; RF, rheumatoid factor; SvH, Sharp–van der Heijde. Where errors are shown, results are means ± SD.
Table 3 Autoantibodies as prognostic markers of poor outcomes
Autoantibody Measure Outcomes at 30 months
Persistence RA criteria Severity
RF Sensitivity 47.2 57.1 55.4
Specificity 91.7 62.8 63.0
Positive LR 5.69 1.53 1.50
Anti-Sa Sensitivity 34.4 39.3 44.6
Specificity 95.8 72.7 79.3
Positive LR 8.26 1.44 2.16
Anti-CCP Sensitivity 39.2 39.3 42.9
Specificity 91.7 66.9 70.7
Positive LR 4.70 1.19 1.46
Combinations of antibodies (rheumatoid factor (RF) and/or anti-Sa and/or anti-cyclic citrullinated peptide (anti-CCP)) are not shown, because none brought any improvement over the detection of anti-Sa alone. LR, likelihood ratio; RA, rheumatoid arthritis.
Table 4 Odds ratio estimates of severe arthritis at 30 months from independent variables present at inclusion
Variable at presentation Odds ratio estimate (95% confidence limits)
Model A Model B Model C
RF ≥ 40 IU/ml 1.788 (0.549–5.825) 2.946 (0.957–9.073) 2.110 (0.715–6.231)
Anti-CCP ≥ 20 U/ml 0.320 (0.075–1.357) 0.917 (0.285–2.953) -
Anti-Sa-positive 8.832 (2.141–36.436)*** - -
Anti-CCP ≥ 60 U/ml - - 1.693 (0.551–5.201)
Age (per year) 1.063 (1.025–1.102)*** 1.049 (1.015–1.084)*** 1.048 (1.014–1.084)***
Male sex 1.741 (0.718–4.223) 1.627 (0.703–3.767) 1.700 (0.733–3.945)
Disease duration 1.686 (0.701–4.058) 1.647 (0.715–3.793) 1.603 (0.700–3.673)
Morning stiffness 0.415 (0.154–1.118) 0.554 (0.218–1.406) 0.502 (0.196–1.284)
Symmetry of arthritis 0.798 (0.174–3.657) 1.085 (0.241–4.880) 1.315 (0.282–6.124)
Swollen joint count 0.985 (0.940–1.033) 0.989 (0.946–1.033) 0.988 (0.945–1.033)
M-HAQ ≥ 1.0 0.666 (0.266–1.666) 0.750 (0.315–1.789) 0.791 (0.330–1.897)
CRP ≥ 8 mg/L 2.594 (0.920–7.312) 2.107 (0.795–5.581) 2.139 (0.807–5.667)
HLA shared epitope 1.096 (0.444–2.704) 0.982 (0.417–2.313) 0.937 (0.401–2.186)
Erosion SvH score 3.472 (1.236–9.575)* 3.751 (1.388–10.135)* 3.611 (1.340–9.729)**
Multivariate regression analysis was performed using all three autoantibodies (model A), without anti-Sa (model B), and without anti-Sa but with a higher threshold of 61 U/ml for anti-cyclic citrullinated peptide (anti-CCP) (model C). Confidence limits are 95% Wald confidence limits. Disease duration was set as positive when present for at least 4 months. Morning stiffness was set as positive when it lasted for at least 60 min. The swollen joint count represents the contribution of each additional joint above the inclusion requirement of three. HLA shared epitope indicates the presence of at least one of the HLA DR alleles associated with RA, as listed in the Methods section. Erosion Sharp–van der Heijde (SvH) score was set as positive when the value was 5 or more. CRP, C-reactive protein; M-HAQ, modified Health Assessment Questionnaire; RF, rheumatoid factor. *P < 0.02; **P < 0.01; ***P < 0.005.
==== Refs
Callahan LF Yelin EH Klippel JD The social and economic consequences of rheumatic diseases Primer on the Rheumatic Diseases 2001 12 Atlanta, GA: Arthritis Foundation 1 4
American College of Rheumatology Subcommittee on Rheumatoid Arthritis Guidelines Guidelines for management of rheumatoid arthritis, 2002 update Arthritis Rheum 2002 46 328 346 11840435 10.1002/art.10148
Harris ED Rheumatoid arthritis. Pathophysiology and implications for therapy N Engl J Med 1990 322 1277 1289 2271017
Landewe RB Boers M Verhoeven AC Westhovens R van de Laar MA Markusse HM van Denderen JC Westedt ML Peeters AJ Dijkmans BA COBRA combination therapy in patients with early rheumatoid arthritis: long-term structural benefits of a brief intervention Arthritis Rheum 2002 46 347 356 11840436 10.1002/art.10083
Mottonen T Hannonen P Korpela M Nissila M Kautiainen H Ilonen J Laasonen L Kaipiainen-Seppanen O Franzen P Helve T Delay to institution of therapy and induction of remission using single-drug or combination-disease-modifying antirheumatic drug therapy in early rheumatoid arthritis Arthritis Rheum 2002 46 894 898 11953964 10.1002/art.10135
Korpela M Laasonen L Hannonen P Kautiainen H Leirisalo-Repo M Hakala M Paimela L Blafield H Puolakka K Mottonen T Retardation of joint damage in patients with early rheumatoid arthritis by initial aggressive treatment with disease-modifying antirheumatic drugs: five-year experience from the FIN-RACo study Arthritis Rheum 2004 50 2072 2081 15248204 10.1002/art.20351
Puolakka K Kautiainen H Mottonen T Hannonen P Korpela M Julkunen H Luukkainen R Vuori K Paimela L Blafield H Impact of initial aggressive drug treatment with a combination of disease-modifying antirheumatic drugs on the development of work disability in early rheumatoid arthritis: a five-year randomized followup trial Arthritis Rheum 2004 50 55 62 14730599 10.1002/art.11436
van Aken J Lard LR le Cessie S Hazes JM Breedveld FC Huizinga TW Radiological outcome after four years of early versus delayed treatment strategy in patients with recent onset rheumatoid arthritis Ann Rheum Dis 2004 63 274 279 14962962 10.1136/ard.2003.010298
Wolfe F Hawley DJ The longterm outcomes of rheumatoid arthritis: Work disability: a prospective 18 year study of 823 patients J Rheumatol 1998 25 2108 2117 9818651
Thomson W Harrison B Ollier B Wiles N Payton T Barrett J Symmons D Silman A Quantifying the exact role of HLA-DRB1 alleles in susceptibility to inflammatory polyarthritis: results from a large, population-based study Arthritis Rheum 1999 42 757 762 10211891 10.1002/1529-0131(199904)42:4<757::AID-ANR20>3.0.CO;2-X
Drossaers-Bakker KW Zwinderman AH Vlieland TP Van Zeben D Vos K Breedveld FC Hazes JM Long-term outcome in rheumatoid arthritis: a simple algorithm of baseline parameters can predict radiographic damage, disability, and disease course at 12-year followup Arthritis Rheum 2002 47 383 390 12209484 10.1002/art.10513
Arnett FC Edworthy SM Bloch DA McShane DJ Fries JF Cooper NS others The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis Arthritis Rheum 1988 31 315 324 3358796
Pincus T Callahan LF How many types of patients meet classification criteria for rheumatoid arthritis? J Rheumatol 1994 21 1385 1389 7983634
Wiles N Symmons DP Harrison B Barrett E Barrett JH Scott DG Silman AJ Estimating the incidence of rheumatoid arthritis: trying to hit a moving target? Arthritis Rheum 1999 42 1339 1346 10403260 10.1002/1529-0131(199907)42:7<1339::AID-ANR6>3.0.CO;2-Y
Mikkelsen WM Dodge HJ A four year follow up of suspected rheumatoid arthritis: the Tecumseh, Michigan, Community Health Study Arthritis Rheum 1969 12 87 91 5777773
O'Sullivan JB Cathcart ES The prevalence of rheumatoid arthritis: follow up evaluation of the effect of criteria on rates in Sudbury, Massachusetts Ann Intern Med 1972 76 573 577 4640316
Wolfe F Ross K Hawley DJ Roberts FK Cathey MA The prognosis of rheumatoid arthritis and undifferentiated polyarthritis syndrome in the clinic: a study of 1141 patients J Rheumatol 1993 20 2005 2009 8014925
Alarcon GS Willkens RF Ward JR Clegg DO Morgan JG Ma KN Singer JZ Steen VD Paulus HE Luggen ME Early undifferentiated connective tissue disease. IV. Musculoskeletal manifestations in a large cohort of patients with undifferentiated connective tissue diseases compared with cohorts of patients with well-established connective tissue diseases: follow up analyses in patients with unexplained polyarthritis and patients with rheumatoid arthritis at baseline Arthritis Rheum 1996 39 403 414 8607889
Schellekens GA de Jong BA van den Hoogen FH van de Putte LB van Venrooij WJ Citrulline is an essential constituent of antigenic determinants recognized by rheumatoid arthritis-specific autoantibodies J Clin Invest 1998 101 273 281 9421490
van Boekel MAM Vossenaar ER van den Hoogen FHJ van Venrooij WJ Autoantibody systems in rheumatoid arthritis: specificity, sensitivity and diagnostic value Arthritis Res 2002 4 87 93 11879544 10.1186/ar395
Despres N Boire G Lopez-Longo FJ Menard HA The Sa system: a novel antigen-antibody system specific for rheumatoid arthritis J Rheumatol 1994 21 1027 1033 7932409
Vossenaar ER Despres N Lapointe E Van Der Heijden A Lora M Senshu T Van Venrooij WJ Menard HA Rheumatoid arthritis specific anti-Sa antibodies target citrullinated vimentin Arthritis Res Ther 2004 6 R142 R150 15059278 10.1186/ar1149
Rantapaa-Dahlqvist S de Jong BA Berglin E Hallmans G Wadell G Stenlund H Sundin U van Venrooij WJ Antibodies against cyclic citrullinated peptide and IgA rheumatoid factor predict the development of rheumatoid arthritis Arthritis Rheum 2003 48 2741 2749 14558078 10.1002/art.11223
Berglin E Padyukov L Sundin U Hallmans G Stenlund H Van Venrooij WJ Klareskog L Dahlqvist SR A combination of autoantibodies to cyclic citrullinated peptide (CCP) and HLA-DRB1 locus antigens is strongly associated with future onset of rheumatoid arthritis Arthritis Res Ther 2004 6 R303 R308 15225365 10.1186/ar1187
Visser H le Cessie S Vos K Breedveld FC Hazes JM How to diagnose rheumatoid arthritis early: a prediction model for persistent (erosive) arthritis Arthritis Rheum 2002 46 357 365 11840437 10.1002/art.10117
Meyer O Labarre C Dougados M Goupille P Cantagrel A Dubois A others Anti-citrullinated protein/peptide antibody assays in early rheumatoid arthritis for predicting five year radiographic damage Ann Rheum Dis 2003 62 120 126 12525380 10.1136/ard.62.2.120
van Gaalen FA Linn-Rasker SP van Venrooij WJ de Jong BA Breedveld FC Verweij CL Toes RE Huizinga TW Autoantibodies to cyclic citrullinated peptides predict progression to rheumatoid arthritis in patients with undifferentiated arthritis: a prospective cohort study Arthritis Rheum 2004 50 709 715 15022309 10.1002/art.20044
van Jaarsveld CH ter Borg EJ Jacobs JW Schellekens GA Gmelig-Meyling FH van Booma-Frankfort C de Jong BA van Venrooij WJ Bijlsma JW The prognostic value of the antiperinuclear factor, anti-citrullinated peptide antibodies and rheumatoid factor in early rheumatoid arthritis Clin Exp Rheumatol 1999 17 689 697 10609067
Goldbach-Mansky R Lee J McCoy A Hoxworth J Yarboro C Smolen JS Steiner G Rosen A Zhang C Menard HA Rheumatoid arthritis associated autoantibodies in patients with synovitis of recent onset Arthritis Res 2000 2 236 243 11056669 10.1186/ar93
Kroot EJ de Jong BA van Leeuwen MA Swinkels H van den Hoogen FH van 't Hof M van de Putte LB van Rijswijk MH van Venrooij WJ van Riel PL The prognostic value of anti-cyclic citrullinated peptide antibody in patients with recent-onset rheumatoid arthritis Arthritis Rheum 2000 43 1831 1835 10943873 10.1002/1529-0131(200008)43:8<1831::AID-ANR19>3.0.CO;2-6
Gossec L Dougados M Goupille P Cantagrel A Sibilia J Meyer O Sany J Daures JP Combe B Prognostic factors for remission in early rheumatoid arthritis: a multiparameter prospective study Ann Rheum Dis 2004 63 675 680 15140774 10.1136/ard.2003.010611
Boire G Menard HA Gendron M Lussier A Myhal D Rheumatoid arthritis: anti-Ro antibodies define a non-HLA-DR4 associated clinicoserological cluster J Rheumatol 1993 20 1654 1660 8295174
The American College of Rheumatology Klippel JD The American College of Rheumatology 1990 criteria for the classification of vasculitis (summary): Appendix I. Criteria for the classification and diagnosis of the rheumatic diseases Primer on the Rheumatic Diseases 1997 11 Atlanta, GA: Arthritis Foundation 459 461
Dougados M van der Linden S Juhlin R Huitfeld B Amor B Calin A the European Spondylarthropathy Study Group The European Spondylarthropathy study Group preliminary criteria for classification of spondylarthropathy Arthritis Rheum 1991 34 1218 1227 1930310
O'Dell JR Therapeutic strategies for rheumatoid arthritis N Engl J Med 2004 350 2591 2602 15201416 10.1056/NEJMra040226
Pincus T Summey JA Soraci SA JrWallston KA Hummon NP Assessment of patient satisfaction in activities of daily living using a modified Stanford Health Assessment Questionnaire Arthritis Rheum 1983 26 1346 1353 6639693
van der Heijde D How to read radiographs according to the Sharp/van der Heijde method J Rheumatol 1999 26 743 745 10090194
Bruynesteyn K Van Der Heijde D Boers M Saudan A Peloso P Paulus H Houben H Griffiths B Edmonds J Bresnihan B Detecting radiological changes in rheumatoid arthritis that are considered important by clinical experts: influence of reading with or without known sequence J Rheumatol 2002 29 2306 2312 12415585
Bruynesteyn K Van Der Linden S Landewe R Gubler F Weijers R Van Der Heijde D Progression of rheumatoid arthritis on plain radiographs judged differently by expert radiologists and rheumatologists J Rheumatol 2004 31 1088 1094 15170919
Prevoo ML van 't Hof MA Kuper HH van Leeuwen MA van de Putte LB van Riel PL Modified disease activity scores that include twenty-eight-joint counts. Development and validation in a prospective longitudinal study of patients with rheumatoid arthritis Arthritis Rheum 1995 38 44 48 7818570
Balsa A Carmona L Gonzalez-Alvaro I Belmonte MA Tena X Sanmarti R Value of Disease Activity Score 28 (DAS28) and DAS28-3 compared to American College of Rheumatology-defined remission in rheumatoid arthritis J Rheumatol 2004 31 40 46 14705217
Fransen J Creemers MC Van Riel PL Remission in rheumatoid arthritis: agreement of the disease activity score (DAS28) with the ARA preliminary remission criteria Rheumatology (Oxford) 2004 43 1252 1255 15238643
Zetterquist H Olerup O Identification of the HLA-DRB1*04, -DRB1*07, and -DRB1*09 alleles by PCR amplification with sequence-specific primers (PCR-SSP) in 2 hours Hum Immunol 1992 34 64 74 1356957 10.1016/0198-8859(92)90086-3
Olerup O Zetterquist H HLA-DR typing by PCR amplification with sequence-specific primers (PCR-SSP) in 2 hours: an alternative to serological DR typing in clinical practice including donor-recipient matching in cadaveric transplantation Tissue Antigens 1992 39 225 235 1357775
Harrison BJ Symmons DPM Brennan P Bankhead CR Barrett EM Scott DGI Silman AJ Inflammatory polyarthritis in the community is not a benign disease: predicting functional disability one year after presentation J Rheumatol 1996 23 1326 1331 8856609
Schellekens GA Visser H de Jong BA van den Hoogen FH Hazes JM Breedveld FC van Venrooij WJ The diagnostic properties of rheumatoid arthritis antibodies recognizing a cyclic citrullinated peptide Arthritis Rheum 2000 43 155 163 10643712 10.1002/1529-0131(200001)43:1<155::AID-ANR20>3.0.CO;2-3
Backhaus M Burmester GR Sandrock D Loreck D Hess D Scholz A Blind S Hamm B Bollow M Prospective two year follow up study comparing novel and conventional imaging procedures in patients with arthritic finger joints Ann Rheum Dis 2002 61 895 904 12228160 10.1136/ard.61.10.895
Wiles N Barrett J Barrett E Silman A Symmons D Disability in patients with early inflammatory polyarthritis cannot be 'tracked' from year to year: an examination of the hypothesis underlying percentile reference charts J Rheumatol 1999 26 800 804 10229399
Yamasaki Y Akaogi J Narain S Richard H Reeves W Satoh M Reactivity of antibodies to cyclic citrullinated peptide (CCP) and unmodified peptide in systemic rheumatic diseases [abstract] Arthritis Rheum 2004 50 Suppl 1 S523
van Gaalen FA van Aken J Huizinga TW Schreuder GM Breedveld FC Zanelli E van Venrooij WJ Verweij CL Toes RE de Vries RR Association between HLA class II genes and autoantibodies to cyclic citrullinated peptides (CCPs) influences the severity of rheumatoid arthritis Arthritis Rheum 2004 50 2113 2121 15248208 10.1002/art.20316
Lard LR Boers M Verhoeven A Vos K Visser H Hazes JM Zwinderman AH Schreuder GM Breedveld FC De Vries RR Early and aggressive treatment of rheumatoid arthritis patients affects the association of HLA class II antigens with progression of joint damage Arthritis Rheum 2002 46 899 905 11953965 10.1002/art.10151
Ostergaard M Wiell C Ultrasonography in rheumatoid arthritis: a very promising method still needing more validation Curr Opin Rheumatol 2004 16 223 230 15103249 10.1097/00002281-200405000-00010
Goldbach-Mansky R Woodburn J Yao L Lipsky PE Magnetic resonance imaging in the evaluation of bone damage in rheumatoid arthritis: a more precise image or just a more expensive one? Arthritis Rheum 2003 48 585 589 12632407 10.1002/art.10819
Szkudlarek M Narvestad E Klarlund M Court-Payen M Thomsen HS Ostergaard M Ultrasonography of the metatarsophalangeal joints in rheumatoid arthritis: comparison with magnetic resonance imaging, conventional radiography, and clinical examination Arthritis Rheum 2004 50 2103 2112 15248207 10.1002/art.20333
| 15899046 | PMC1174957 | CC BY | 2021-01-04 16:02:37 | no | Arthritis Res Ther. 2005 Mar 17; 7(3):R592-R603 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1719 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17211589904810.1186/ar1721Research ArticleIncreased serum HO-1 in hemophagocytic syndrome and adult-onset Still's disease: use in the differential diagnosis of hyperferritinemia Kirino Yohei [email protected] Mitsuhiro [email protected] Mika 1Ueda Atsuhisa 1Ohno Shigeru 1Shirai Akira 1Kanamori Heiwa 1Tanaka Katsuaki 2Ishigatsubo Yoshiaki [email protected] Department of Internal Medicine and Clinical Immunology, Yokohama City University Graduate School of Medicine, Yokohama, Japan2 Yokohama City University Medical Center, Department of Gastroenterological Center, Yokohama, Japan2005 21 3 2005 7 3 R616 R624 8 12 2004 26 1 2005 17 2 2005 21 2 2005 Copyright © 2005 Kirino et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Heme oxygenase-1 (HO-1), an inducible heme-degrading enzyme, is expressed by macrophages and endothelial cells in response to various stresses. Because ferritin synthesis is stimulated by Fe2+, which is a product of heme degradation, we examined the relation between HO-1 and ferritin levels in the serum of patients with hemophagocytic syndrome (HPS), adult-onset Still's disease (ASD), and other diseases that may cause hyperferritinemia. Seven patients with HPS, 10 with ASD, 73 with other rheumatic diseases, 20 with liver diseases, 10 recipients of repeated blood transfusion because of hematological disorders, and 22 healthy volunteers were enrolled. Serum HO-1 and ferritin levels were determined by ELISA. Expression of HO-1 mRNA and protein by peripheral blood mononuclear cells (PBMCs) was determined by real-time PCR and immunocytochemical techniques, respectively. Serum levels of HO-1 were significantly higher in patients with active HPS and ASD than in the other groups (P < 0.01). HO-1 levels were not elevated in patients with other causes of hyperferritinemia but were moderately elevated in patients with dermatomyositis/polymyositis. Among patients with HPS and ASD, serum HO-1 levels correlated closely with serum ferritin levels, and the levels of both returned to normal after therapy had induced remission. Increased expression of HO-1 mRNA was confirmed in PBMCs from some patients with HPS and ASD. Hyperferritinemia correlated closely with increased serum HO-1 in patients with HPS and ASD but not other conditions, indicating that measurement of serum HO-1 and ferritin levels would be useful in the differential diagnosis of hyperferritinemia and perhaps also in monitoring disease activity in HPS and ASD.
==== Body
Introduction
Heme oxygenase (HO) is an enzyme that catalyzes the conversion of heme into CO, Fe2+, and biliverdin [1,2]. HO-1, an inducible form of HO, is a 32-kD heat shock protein expressed in response to various noxious stimuli including heavy metals, hyperoxia, hypoxia, endotoxin, hydrogen peroxide, and inflammatory cytokines [1,2]. Evidence suggests that increased expression of HO-1 can benefit the host in a variety of pathological conditions [1-5]. In this context, our research team has found that HO-1 gene therapy is useful for lipopolysaccharide-induced lung injury [6], influenza viral pneumonia [7], bleomycin-induced pulmonary fibrosis [8], and chronic respiratory infection with Pseudomonas aeruginosa in mice [9]. We also found that chemically induced HO-1 was of benefit in lupus nephritis [10]. On the other hand, a deficiency in HO-1 expression is associated with severe chronic inflammation, as shown in studies of HO-1 knockout mice (mice in which the gene for HO-1 had been inactivated) and a patient with HO-1 deficiency [11-13]. This observation is consistent with HO-1 having a physiological effect in protecting against inflammation.
Products of heme degradation mediate the protective effects of HO-1. CO suppresses apoptosis, macrophage activation, and the synthesis of proinflammatory cytokines, nitrite oxide, and prostaglandins [1,2,14]. Biliverdin is converted into bilirubin, an antioxidant [1,2,15-18]. Fe2+, which itself has toxic effects by inducing the formation of free radicals, stimulates the production of ferritin [19]. Ferritin acts as an antioxidant and detoxifies Fe2+ [19]. Thus, the heme degradation products and the metabolic derivatives generated by HO-1 suppress toxic events in cells.
Regulation of HO-1 is of particular interest in the inflammation associated with hyperferritinemia, as is the case in hemophagocytic syndrome (HPS) and adult-onset Still's disease (ASD), because HO-1 can be involved in increased ferritin in these conditions [1,2]. HPS is a serious, life-threatening condition, which is characterized by cytopenia due to hemophagocytosis [20-22]. The disease is subdivided into two categories, familial lymphohistiocytosis and secondary HPS, the latter of which is associated with rheumatic diseases such as systemic-onset juvenile idiopathic arthritis, viral infection, and certain malignancies [20].
Like children with Still's disease, patients with ASD present with high fever, arthralgia, typical skin rash, hepatosplenomegaly, and leukocytosis [20,21]. HPS and ASD share several clinical features, including high fever, hepatosplenomegaly, lymphadenopathy, liver injury, and coagulopathy [20,21]. The observation that severe ASD is sometimes complicated by HPS is consistent with the suggestion that a common pathophysiology may link these two diseases [20,21,23].
Recent studies have shown that dysfunction of natural killer (NK) cells due to mutations of the genes for perforin and Munc 13-4 leads to familial lymphohistiocytosis, whereas it has been suggested that decreased NK cell activity and abnormal levels of perforin are involved in the macrophage activation syndrome of systemic-onset juvenile rheumatoid arthritis [20]. Dysfunction of NK and cytotoxic cells may lead to inadequate control of cellular immune responses, resulting in systemic macrophage activation, which is implicated in the development of both diseases of HPS and ASD. Subsequently, excessive production of proinflammatory cytokines and active infiltration of macrophages into vital organs have been observed [20,21]. Increased serum ferritin is characteristic of, but not specific for, both diseases, because it is also elevated in various other conditions [23,24]. For example, patients with hyperferritinemia who have rheumatic or liver disease or who receive frequent transfusions because of hematological diseases often develop cytopenia and high fever resembling these signs in HPS.
Lack of specific disease markers often delays diagnosis of HPS and ASD, with potentially lethal consequences [21]. The present study shows that serum HO-1 levels are significantly increased in patients with active HPS and ASD but not in patients with hyperferritinemia due to other causes. Moreover, there is a close correlation between serum HO-1 levels and the disease activity in HPS and ASD.
Materials and methods
Patients
All patients enrolled in this study were being treated at the Yokohama City University Hospital, the Yokohama City University Medical Center Hospital, or the National Hospital Organization Yokohama Medical Center (Table 1). Seven patients with secondary HPS met the diagnostic guideline for hemophagocytic lymphohistiocytosis [22,25], except as regards hypertriglyceridemia and hypofibrinogenemia, neither of which is generally applicable to secondary HPS in adults. In these seven patients, the underlying diseases were systemic lupus erythematosus (SLE) in two; hematological malignancy, including non-Hodgkin's lymphoma, multiple myeloma, and acute myeloid leukemia, in three; and ASD and viral infection in the others. The patients having more than two lineages of cytopenia, liver dysfunction, fever above 39°C, and hyperferritinemia were categorized as having active disease. Remission of the diseases was defined as disappearance of these findings after therapy. Ten patients with ASD met the criteria of Cush [26] and Yamaguchi [27] and their colleagues. An ASD patient who also met the diagnostic guidelines for hemophagocytic lymphohistiocytosis was classified in the HPS group in this study. Patients with active ASD were those presenting with polyarthritis, typical skin rashes, and fever above 39°C, in addition to hyperferritinemia. When the symptoms and signs had subsided, the patients were considered to be in remission.
We also studied 73 patients with other rheumatic diseases, including 30 with rheumatoid arthritis (RA), 18 with SLE, 9 with dermatomyositis/polymyositis (DM/PM), and 16 with Behçet's disease (BD). The diagnosis of individual diseases was based on the following criteria: for RA, the 1987 American College of Rheumatology (formerly, the American Rheumatism Association) criteria [28]; for SLE, the 1997 updating of the American College of Rheumatology revised criteria for the classification of systemic lupus erythematosus [29]; for polymyositis and dermatomyositis, the diagnostic criteria described by Bohan and Peter [30,31]; and for Behçet's disease, the International Study Group criteria for diagnosis of Behçet's disease [32]. The disease activity was evaluated at the time of blood sampling. All of the RA patients were considered to have active disease, because their disease activity scores (DAS) based on 28 joints and C-reactive protein (CRP) (DAS28-CRP) were more than 3.2 [33]. The mean CRP level at the time of blood sampling was 2.4 ± 2.7 mg/dl. On the basis of the Systemic Lupus Disease Activity Index [34], 12 of the 18 SLE patients had a score above 9 and were regarded as having active disease, while the other 6 were in remission. Two other SLE patients who met the diagnostic guidelines for hemophagocytic lymphohistiocytosis were included in the HPS group [22,25]. All of the DM/PM patients had active diseases, inasmuch as their creatine kinase concentrations were more than twice the normal upper limit and they had muscle weakness and/or active interstitial pneumonia. Six of the 16 BD patients presented active symptoms of uveitis, erythema nodosum, genital ulcers, deep vein thrombus, central nervous system involvement, arterial occlusion, or gastrointestinal lesions in addition to positive CRP, indicating active disease; the other 10 were regarded as having inactive disease.
Twenty patients with liver diseases were enrolled in this study (Table 1). Of the five with acute hepatitis, three had hepatitis B, one had drug-induced hepatitis, and one had Epstein–Barr viral hepatitis. The seven patients with chronic hepatitis included one with hepatitis B and six with hepatitis C. Serum alanine aminotransferase levels were measured as an indicator of liver injury. The means ± standard deviations (IU/l) found for these 20 patients were as follows: acute hepatitis, 770.0 ± 568.6; chronic hepatitis, 61.9 ± 28.9; liver cirrhosis, 59.0 ± 24.0; hepatocellular carcinoma, 27.5 ± 14.8; primary biliary cirrhosis, 98.5 ± 14.1; autoimmune hepatitis, 259; and alcoholic hepatitis, 56.
Ten patients who had received frequent blood transfusions were also included. The underlying hematological diseases were myelodysplastic syndrome in six patients and aplastic anemia in four. Healthy volunteers served as normal controls. All the studies were performed after obtaining written informed consent, which was approved by the local Institutional Review Board.
ELISA
Serum ferritin and HO-1 levels were measured by an EIA detection system (Tosoh, Tokyo, Japan), and a human HO-1 ELISA kit (Stressgen, Victoria, Canada), respectively. Concentrations of serum tumor necrosis factor (TNF)-α were determined by specific ELISA systems using pairs of capture and biotin-conjugated detecting antibodies, which were purchased from R&D (Minneapolis, MN, USA). Serum IL-18 level was determined using a human IL-18 ELISA kit in accordance with the manufacturer's protocol (MBL, Nagoya, Japan).
Cell preparation and culture
Peripheral blood mononuclear cells (PBMCs) were isolated by centrifugation over Ficoll-Hypaque (ICN, Aurora, OH, USA). 106cells/ml were cultured with 100 μM hemin (Sigma-Aldrich, Saint Louis, MO, USA) in Hepes modified RPMI 1640 (Sigma-Aldrich) containing 10% fetal calf serum (Equitech-Bio, Kerrville, TX, USA), 2 mM L-glutamine (Sigma-Aldrich), 100 U/ml penicillin, plus 100 μg/ml streptomycin (Sigma-Aldrich) in a 5% CO2 in an air incubator at 37°C for 24 hours.
Real-time PCR
Total RNA was isolated from cells by using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). Reverse transcription was performed using a SuperScript™ reverse transcriptase (Invitrogen). Panels of primers of human HO-1 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA were purchased from PE Applied Biosystems (Foster City, CA, USA). Real-time PCR was performed using a TaqMan Universal Master Mix (PE Applied Biosystems), and the data were analyzed by the ABI prism 7700 sequence detection system (PE Applied Biosystems). Briefly, 1/50 amounts of of cDNA derived from 1 μg of total RNA, 200 nmol/l of probe, and 800 nmol/l of primers were incubated in 25 μl at 50°C for 2min and 95°C for 10min, followed by 40 cycles of 95°C for 15s and 60°C for 1 min. The amounts of cDNA obtained from transcriptions of mRNA were semiquantified in comparison with those of serially diluted standard cDNA, which was prepared using a conventional PCR technique. The expression level of HO-1 mRNA in a sample was expressed as arbitrary units, which were determined by the formula 1AU = (HO-1 mRNA/GAPDH mRNA) × 100.
Immunocytochemistry
Cells expressing HO-1 were determined with anti-HO-1 monoclonal antibody (Stressgen) using a Dako LSAB2 kit (Dako, Glostrup, Denmark).
Statistical analysis
The Mann–Whitney U test, the Wilcoxon signed rank test, and multiple regression analyses were used to test for differences. P values less than 0.05 were considered significant. Values are reported as means ± standard deviations.
Results
Increased serum HO-1 levels in patients with HPS and ASD
Serum HO-1 levels in patients with inflammatory rheumatic diseases were monitored by ELISA. In the healthy controls, only very low levels of serum HO-1 were detectable (2.6 ± 1.3 ng/ml) (Fig. 1). Age and sex did not influence HO-1 levels. In contrast, HO-1 levels were significantly elevated in patients with active ASD and HPS (Table 1; Fig. 1). HO-1 protein levels exceeded 10 ng/ml in all but one patient with ASD, who was classified as having active disease in the study because of high fever with elevated levels of CRP and ferritin during maintenance therapy with a low dose of prednisolone (PSL). However, the clinical manifestations were less serious and serum ferritin was lower (1201 ng/ml) than in any other patient with active ASD in this study. Although subjects with active DM/PM also had significantly increased serum HO-1 levels (Table 1; P = 0.001), these were still significantly lower than in subjects with active HPS or ADS (P = 0.0007 and P = 0.003, respectively). Serum HO-1 levels were not increased in other rheumatic diseases including RA, SLE, and BD, regardless of disease activity (except for two patients with SLE complicated by HPS) (Table 1). These findings suggest that increased serum HO-1 levels are characteristic of active ASD and HPS.
Serum HO-1 is a marker of disease severity in HPS and ASD
Serum HO-1 levels were monitored before and after remission-inducing therapy that included corticosteroids with or without cyclosporin A in three patients with HPS and five with ASD. Serum HO-1 levels were significantly reduced after successful therapy (Fig. 2a) (P = 0.0078).
Serum HO-1 and ferritin were serially monitored in one patient with ASD and one with HPS during the course of disease (Fig. 2b,c). A 34-year-old man admitted with fever, polyarthralgia, sore throat, and salmon-pink rashes was diagnosed with ASD (Fig. 2b). When this patient was admitted, his serum concentrations of both HO-1 and ferritin were extremely elevated (182 ng/ml and 6,855 ng/ml, respectively). Treatment with methylprednisolone (mPSL) pulse therapy (1,000 mg/day for 3 days) followed by oral PSL (60 mg/day) and cyclosporin A (200 mg/day) led to clinical remission. Associated with this response to therapy, serum HO-1 levels gradually decreased to the normal range over 2 months, as did levels of ferritin and CRP. PSL was tapered to 30 mg/day without relapse.
In a 45-year-old woman with SLE admitted with high fever and cytopenia (Fig. 2c), bone marrow aspiration revealed hemophagocytosis, and her serum ferritin level was 4,588 ng/ml, resulting in a diagnosis of HPS complicated with SLE (the Systemic Lupus Disease Activity Index score was 9). On admission, increased serum HO-1 (74.8 ng/ml) was noted. mPSL pulse therapy (1,000 mg/day for 3 days) followed by oral PSL (60 mg/day) and intravenous gamma globulin (17.5 g/day for 5 days) temporarily reduced her fever and CRP levels. Despite these treatments, serum ferritin and HO-1 peaked at 25,070 ng/ml and 214 ng/ml, respectively. A second course of mPSL pulse therapy also failed, but the patient's condition gradually improved after initiation of cyclosporin A (200 mg/day). Serum levels of CRP, ferritin, and HO-1 reached normal levels by two months after admission. PSL was tapered to 30 mg/day without exacerbation. These findings suggest that the serum HO-1 level is closely correlated with disease activity during the clinical course in patients with HPS and ASD.
We next examined the relation between the serum HO-1 level and other laboratory parameters in the patients with HPS and ASD. Because serum ferritin was widely accepted as a monitoring marker for the diseases, the data included in the analysis were those found when the ferritin level was highest in individual patients during the whole study. The results indicate that serum HO-1 correlates closely with serum ferritin (P = 0.0048, Fig. 3a) but not CRP or lactate dehydrogenase (LDH) levels (Fig. 3b,c), a finding consistent with an association between HO-1 and hyperferritinemia in patients with HPS and ASD. We also measured serum levels of IL-18 and TNF-α, both of which have been shown to be elevated in patients with HPS and ASD [35,36]. However, we did not find any correlation between serum level of HO-1 and those of cytokines (data not shown).
Increased serum HO-1 level is not always associated with hyperferritinemia
Besides being found in patients with HPS and ASD, hyperferritinemia is also found in patients with liver diseases and in recipients of frequent blood transfusions. Because ferritin synthesis is stimulated by Fe2+, which is generated by HO-1-mediated heme degradation, hyperferritinemia might be caused by high HO-1 activity, irrespective of the underlying diseases. To examine this possibility, the relation between serum HO-1 and ferritin was evaluated in all patient groups. A total of 37 patients had serum ferritin levels >500 ng/ml, which is the cutoff level in the revised diagnostic criteria for HLH [22,25]. Serum HO-1 levels exceeded 10 ng/ml in 7 of 7 HPS patients and in 9 of 10 ASD patients but in only 2 of 20 patients with other diseases (one with dermatomyositis and the other with Epstein–Barr hepatitis) (Fig. 4). Of all the subjects studied, only one person, with dermatomyositis, had serum HO-1 >10 ng/ml but serum ferritin <500 ng/ml. Thus, simultaneous elevation of serum ferritin and HO-1 was much more common in patients with ASD and HPS than any other disease studied.
HO-1 is up-regulated in PBMCs from some, but not all, patients with active HPS and ASD
Yachie and colleagues reported that PBMCs from children with acute inflammatory illness express elevated HO-1 mRNA levels [37]. In our study, HO-1 mRNA expression in PBMCs was semiquantified using real-time PCR. We found that PBMCs from 3 of 5 patients with active HPS and 3 of 10 with active ASD had HO-1 mRNA expression exceeding the mean + 2 standard deviations of healthy controls, whereas no such elevations were found in PBMCs from patients with other rheumatic diseases, irrespective of disease activity (Fig. 5a). The six patients with increased mRNA expression universally manifested elevated serum HO-1 protein levels. Moreover, HO-1 mRNA expression fell when remission was induced in two patients with HPS and one with ASD (Fig. 5b). Changes in HO-1 mRNA expression mirrored changes in serum HO-1 protein levels in a 45-year-old woman (Fig. 5c). This patient, who had had ASD for 4 years and maintained remission with PSL (20 mg/day), was admitted to our hospital because of high fever and cytopenia. Bone marrow aspiration revealed hemophagocytosis, indicating that the patient's ASD was complicated with HPS. Besides increased serum ferritin (8,690 ng/ml) and HO-1 (40.4 ng/ml), HO-1 mRNA expression in PBMCs was much higher than that of healthy controls. Immunocytochemistry showed that HO-1-expressing cells were found in hemin-treated, but not untreated, PBMCs from normal donors (Fig. 6a,b), whereas HO-1 proteins were stained in freshly isolated PBMCs, mainly monocytes, from the patient (Fig. 6c). After clinical remission was achieved by mPSL pulse therapy and subsequent oral PSL, HO-1 mRNA in PBMCs was reduced in parallel with serum HO-1 and ferritin levels (Fig. 5c). These data indicate that circulating PBMCs may contribute to increased serum HO-1 protein levels in some subjects. However, since HO-1 mRNA expression was normal in PBMCs from 9 of 15 patients with active HPS and ASD, despite elevated serum HO-1, it is clear that PBMCs are not a critical source of circulating HO-1.
Discussion
This study demonstrates that serum HO-1 levels are elevated in patients with active HPS and ASD, and that these levels correlate closely with disease activity, irrespective of underlying conditions and clinical phenotypes. Serum HO-1 levels were also slightly elevated in some patients with DM/PM, but not to the degree of patients with HPS or ASD.
Yachie and colleagues reported that HO-1 mRNA levels were elevated in PBMCs from children with acute inflammatory illness and suggested that HO-1 is up-regulated when cells are stressed [37]. It has been shown that HO-1 is cytoprotective in a number of pathological conditions [1,2], although an excess of HO-1 can also injure cells [38-40]. In the current study, increased serum HO-1 was present only in patients with active disease, although it is unclear whether HO-1 was playing a protective or harmful role in these subjects.
Very high levels of serum ferritin are widely used as a marker for HPS and ASD [20,21,23], although the mechanism underlying this increase in ferritin is unknown. The current work documents a significant correlation between serum HO-1 and ferritin levels in HPS and ASD patients. Increased HO-1 activity generates Fe2+, a heme catabolyzed product of HO-1, which acts as a potent stimulator of ferritin synthesis [19]. Indeed, it has been shown that more Fe2+ is sequestered by ferritin in ASD patients than in healthy controls, whereas the iron saturation of individual ferritin molecules was decreased [41]. These findings are compatible with the hypothesis that increased HO-1 contributes to hyperferritinemia in ASD and HPS. Alternatively, because Nrf2 (nuclear factor, erythroid derived 2, like 2) regulates transcription of HO-1 and ferritin genes, activation of the transcription factor may be involved in simultaneous overproduction of both molecules [42,43]. On the other hand, it is plausible that an HO-1-independent or an Nrf2-independent mechanism or both are responsible for the elevation in serum ferritin level in subjects with liver disease and frequent transfusions.
Sources of circulating HO-1 in patients with HPS and ASD remain undetermined. These diseases are recognized as macrophage-activation diseases, because increased proinflammatory cytokines such as IL-6, TNF-α, and IL-18 are dominantly produced by macrophages [20,21,25,35,36]. Moreover, HPS and severe ASD are characterized by the proliferation of macrophages that phagocytose hematopoietic cells in the bone marrow and their subsequent infiltration into other organs, accounting in part for the systemic clinical symptoms of these diseases [20]. In response to various stresses, HO-1 is strongly expressed in cells of the macrophage lineage, including circulating monocytes [37]. We found that PBMCs from some, but not all, HPS and ASD patients with elevated serum HO-1 levels overexpressed HO-1 mRNA. It therefore seems that serum HO-1 proteins may be partly derived from circulating monocytes in ASD and HPS patients, although other sources of HO-1 must also be involved.
Useful diagnostic criteria for familial hemophagocytic lymphohistiocytosis [22,25] are well established, whereas it is sometimes hard to diagnose secondary HPS, especially in adults. Although the diagnosis requires the histological identification of hemophagocytosis in organs, the findings are often difficult to prove even by biopsies of the bone marrow, lymph nodes, and liver [21,44]. Depressed NK cell activity and increased soluble IL-2 receptor levels are helpful but are not specific for the disease. In the early stage of ASD, the diagnostic criteria [26,27] are not satisfied in some patients.
Hyperferritinemia is found not only in HPS and ASD, but also in other rheumatic diseases, liver diseases, and hematological disorders with frequent transfusions. All of these diseases can be accompanied by cytopenia and/or high fever, leading to difficulty of differential diagnosis. Since no disease-specific findings have been established, it is important to exclude other diseases. The delay associated with examinations may delay the initiation of critically needed therapies. On the other hand, it is prompt, simple, noninvasive, and informative to measure serum HO-1 levels by ELISA in such situations.
In contrast to the case with HPS and ASD, hyperferritinemia is not associated with elevated serum HO-1 levels in patients with liver disease or hematological diseases requiring frequent transfusions. This clear distinction suggests that the combination of increased serum HO-1 plus ferritin provides greater specificity in the diagnosis of HPS and ASD.
Conclusion
The present study shows that serum HO-1 is a novel marker for the diagnosis of HPS and ASD and for monitoring disease activity. Further studies are required to determine the mechanism and sources of increased serum HO-1 in these diseases. Clarification of the relation between HO-1 and ferritin metabolism will shed further light on the pathogenesis of HPS and ASD.
Abbreviations
ASD = adult-onset Still's disease; BD = Behçet's disease; CO = carbon monoxide; CRP = C-reactive protein; DM/PM = dermatomyositis/polymyositis; ELISA = enzyme-linked immunosorbent assay; GAPDH = glyceraldehyde-3-phosphate dehydrogenase; HO = heme oxygenase; HPS = hemophagocytic syndrome; IL = interleukin; mPSL = methylprednisolone; NK = natural killer; PBMC = peripheral blood mononuclear cell; PCR = polymerase chain reaction; PSL = prednisolone; RA = rheumatoid arthritis; SLE = systemic lupus erythematosus; TNF = tumor necrosis factor.
Competing interests
The authors have received no financial support or other benefits from commercial sources for the work reported in the manuscript, and no other financial interests that any of the authors may have could create a potential conflict of interest or the appearance of a conflict of interest with regard to the work.
Authors' contributions
YI designed and organized the study. YK, MT, and MI, conducted the laboratory work. YK, MT, AU, SO, AS, HK, KT, and YI were involved in the analysis and interpretation of data. YK, MT, and YI were involved in writing the report. All authors read and approved the final manuscript.
Acknowledgements
This work was supported in part by grants from The Yokohama City University Center of Excellence Program of the Ministry of Education, Culture, Sports, Science and Technology of Japan (to Y Ishigatsubo), Research on Specific Disease of the Health Science Research Grants of the Ministry of Health, Labour, and Welfare (to Y Ishigatsubo), and 2004 grant in aid for scientific research project No. 16590991 from the Ministry of Education, Culture, Sports, and Technology of Japan (to M Takeno). The sources of funding had no role in the writing of the report and did not participate in the decision to publish the results. The authors would like to thank Hideo Kobayashi, Yukiko Taked, Ryusuke Yoshimi, Hiroshi Kobayashi, and Kyosuke Motoji, who were involved in collecting blood samples from the patients. The authors are greatly indebted to Dr Dennis M Klinman, Center for Biologics Evaluation and Research, Food and Drug Administration, Bethesda, MD, USA, for his review and invaluable suggestions in preparing the manuscript. We also thank Dr Kenji Ohshige, Yokohama City University School of Medicine, Department of Public Health, Yokohama, Japan, for statistical advice.
Figures and Tables
Figure 1 Serum heme oxygenase-1 in patients with hemophagocytic syndrome or adult-onset Still's disease. Also studied were normal controls (NC) and people with other rheumatic diseases including rheumatoid arthritis (RA) (n = 30), systemic lupus erythematosus (SLE) (n = 18), dermatomyositis/polymyositis (DM/PM) (n = 9), and Behçet's disease (BD) (n = 16). Filled circles and open circles represent patients with active and inactive disease, respectively. *P < 0.0001, **P = 0.0001, §P = 0.001, †P = 0.0007, ‡P = 0.003, as determined by the nonpaired Mann–Whitney U test. ASD, adult-onset Still's disease; HO-1, heme oxygenase 1; HPS, hemophagocytic syndrome.
Figure 2 Serum heme oxygenase (HO)-1 and ferritin in hemophagocytic syndrome (HPS) and adult-onset Still's disease (ASD). (a) Serum HO-1 levels of HPS patients (open circles) and ASD patients (filled circles) before and after remission. *P = 0.0078, as determined by the Wilcoxon signed-rank test. (b) Clinical course of ASD in a 34-year-old man. (c) Clinical course in a 45-year-old woman with HPS and systemic lupus erythematosus. 'Pulse' represents intravenous infusion of methylprednisolone at 1,000 mg/day for 3 days. ALT, alanine aminotransferase; CRP, C-reactive protein.
Figure 3 Correlation between serum heme oxygenase (HO)-1 and other serum constituents. (a) Correlation between serum HO-1 and ferritin in the patients with hemophagocytic syndrome (HPS) and adult-onset Still's disease (ASD) at the time when the serum ferritin was highest during the study. P = 0.0048, as determined by multiple regression analyses. (b,c) Correlations between HO-1 and (b) lactate dehydrogenase (LDH), and (c) C-reactive protein (CRP) in the same patients at the same point in the study.
Figure 4 Serum heme oxygenase (HO)-1 and ferritin levels in all the patients studied. Filled triangles stand for patients with active hemophagocytic syndrome (HPS). Open circles stand for those with active adult-onset Still's disease (ASD). The horizontal dotted line indicates 500 ng/ml of ferritin, which was determined on the basis of revised Diagnostic Guidelines for hemophagocytic lymphohistiocytosis [22,25], and the vertical dotted line indicates the arbitrary cutoff value 10 ng/ml of HO-1.
Figure 5 Expression of HO-1 mRNA in PBMCs semiquantified by real-time PCR. The data are expressed as arbitrary units (AU). (a) Heme oxygenase (HO)-1 mRNA levels in patients with hemophagocytic syndrome (HPS) (n = 5), adult-onset Still's disease (ASD) (n = 10), rheumatoid arthritis (RA) (n = 15), systemic lupus erythematosus (SLE) (n = 6), or Behçet's disease (BD) (n = 13), and in normal controls (NC) (n = 20). Filled circles and open circles represent patients with active and inactive disease, respectively. *P < 0.05 as determined by the Mann–Whitney U test. The horizontal dotted line represents the mean + 2 standard deviations of the mRNA level in healthy controls. (b) HO-1 mRNA levels in peripheral blood mononuclear cells (PBMCs) from HPS and ASD patients (filled and open circles, respectively) before and after remission. (c) Clinical course of ASD and HPS in one patient. The numbers on the vertical axis representng/ml (serum ferritin and HO-1 concentrations) and AU (HO-1 mRNA). 'Pulse' represents intravenous infusion of methylprednisolone at 1,000 mg/day for 3 days.
Figure 6 Expression of heme oxygenase (HO)-1 protein in PBMCs, determined using anti-HO-1 monoclonal antibody. (a) Untreated peripheral blood mononuclear cells (PBMCs) from a healthy control. (b) 100 μM hemin-treated PBMCs from a healthy control. (c) PBMCs from a patient with active adult-onset Still's disease (ASD) complicated by hemophagocytic syndrome (HPS). HO-1-expressing monocytes (stained red) were found in (b) and (c). Original magnification × 400.
Table 1 Characteristics of the patients enrolled in the study
Diagnosis No. Age Sex (M/F) Serum HO-1 (ng/ml) Serum ferritin (ng/ml)
Hemophagocytic syndrome
7 42.7 (15.5) 1/6 71.2 (72.7) 8485.3 (8388.0)
Adult-onset Still's disease
10 41.0 (11.9) 5/5 102.8 (102.6) 9658.5 (17042.1)
Rheumatic diseases
73 48.2 (15.9) 22/51 3.4 (2.7) 225.5 (709.9)
Rheumatoid arthritis 30 53.1 (13.2) 8/22 2.8 (1.4) 83.0 (95.3)
Systemic lupus erythematosus 18 38.0 (16.8) 1/17 3.2 (1.9) 188.0 (286.3)
Active 12 42.5 (16.4) 1/11 3.6 (2.1) 220.4 (329.8)
Inactive 6 29.0 (15.1) 0/6 2.2 (1.0) 100.6 (122.1)
Behçet's disease 16 47.3 (14.4) 10/6 2.4 (0.8) 48.3 (43.0)
Active 6 46.5 (14.3) 5/1 2.5 (0.5) 61.4 (45.8)
Inactive 10 47.7 (15.3) 5/5 2.3 (1.0) 40.4 (41.7)
Dermatomyositis/polymyositis 9 54.4 (16.6) 3/6 7.4 (5.2) 1097.2 (1827.9)
Liver diseases
20 47.8 (18.1) 16/4 3.7 (2.4) 1032.2 (2496.9)
Acute hepatitis 5 27.8 (5.6) 4/1 4.8 (3.9) 1347.0 (861.0)
Chronic hepatitis 7 51.9 (12.5) 6/1 3.8 (1.6) 159.9 (128.0)
Liver cirrhosis 2 69.0 (7.1) 2/0 5.8 (1.4) 102.0 (28.3)
Hepatocellular carcinoma 2 59.5 (10.6) 2/0 1.9 (0.7) 260.0 (134.4)
Primary biliary cirrhosis 2 46.0 (21.2) 1/1 2.8 (3.0) 373.5 (892.4)
Autoimmune hepatitis 1 34 0/1 2.9 11262
Alcoholic hepatitis 1 71 1/0 2.8 56
Hematological diseases
10 62.3 (15.3) 6/4 4.3 (2.4) 2822.6 (2817.3)
Myelodysplastic syndrome 6 59.3 (22.5) 4/2 3.7 (2.6) 1883.8 (1885.9)
Aplastic anemia 4 54.0 (18.8) 2/2 5.1 (2.1) 4230.8 (3671.1)
Healthy controls
22 30.8 (7.6) 16/6 2.6 (1.3) 93.0 (56.9)
Data are shown as means (standard deviations). F, female; HO, heme oxygenase; M, male.
==== Refs
Morse D Choi AM Heme oxygenase-1: the "emerging molecule" has arrived Am J Respir Cell Mol Biol 2002 27 8 16 12091240
Otterbein LE Soares MP Yamashita K Bach FH Heme oxygenase-1: unleashing the protective properties of heme Trends Immunol 2003 24 449 455 12909459 10.1016/S1471-4906(03)00181-9
Maines MD Heme oxygenase: function, multiplicity, regulatory mechanisms, and clinical applications FASEB J 1988 2 2557 2568 3290025
Willis D Moore AR Frederick R Willoughby DA Heme oxygenase: a novel target for the modulation of the inflammatory response Nat Med 1996 2 87 90 8564848 10.1038/nm0196-87
Nath KA Balla G Vercellotti GM Balla J Jacob HS Levitt MD Rosenberg ME Induction of heme oxygenase is a rapid, protective response in rhabdomyolysis in the rat J Clin Invest 1992 90 267 270 1634613
Inoue S Suzuki M Nagashima Y Suzuki S Hashiba T Tsuburai T Ikehara K Matsuse T Ishigatsubo Y Transfer of heme oxygenase 1 cDNA by a replication-deficient adenovirus enhances interleukin 10 production from alveolar macrophages that attenuates lipopolysaccharide-induced acute lung injury in mice Hum Gene Ther 2001 12 967 979 11387061 10.1089/104303401750195926
Hashiba T Suzuki M Nagashima Y Suzuki S Inoue S Tsuburai T Matsuse T Ishigatsubo Y Adenovirus-mediated transfer of heme oxygenase-1 cDNA attenuates severe lung injury induced by the influenza virus in mice Gene Ther 2001 8 1499 1507 11593363 10.1038/sj.gt.3301540
Tsuburai T Suzuki M Nagashima Y Suzuki S Inoue S Hashiba T Ueda A Ikehara K Matsuse T Ishigatsubo Y Adenovirus-mediated transfer and overexpressioin of heme oxygenase 1 cDNA in lung prevents bleomycin-induced pulmonary fibrosis via a Fas-Fas ligand-independent pathway Hum Gene Ther 2002 13 1945 1960 12449100 10.1089/10430340260355356
Tsuburai T Kanako T Nagashima Y Ueda A Tagawa A Shinohara T Ishigatsubo Y Pseudomonas aeruginosa-induced neutrophilic lung inflammation is attenuated by adenovirus-mediated transfer of the heme oxygenase 1 cDNA in mice Hum Gene Ther 2004 15 273 285 15018736 10.1089/104303404322886129
Takeda Y Takeno M Iwasaki M Kobayashi H Kirino Y Ueda A Nagahama K Aoki I Ishigatsubo Y Chemical induction of HO-1 suppresses lupus nephritis by reducing local iNOS expression and synthesis of anti-dsDNA antibody Clin Exp Immunol 2004 138 237 244 15498032 10.1111/j.1365-2249.2004.02594.x
Yachie A Niida Y Wada T Igarashi N Kaneda H Toma T Ohta K Kasahara Y Koizumi S Oxidative stress causes enhanced endothelial cell injury in human heme oxygenase-1 deficiency J Clin Invest 1999 103 129 135 9884342
Kawashima A Oda Y Yachie A Koizumi S Nakanishi I Heme oxygenase-1 deficiency: the first autopsy case Hum Pathol 2002 33 125 130 11823983 10.1053/hupa.2002.30217
Poss KD Tonegawa S Reduced stress defense in heme oxygenase 1-deficient cells Proc Natl Acad Sci USA 1997 94 10925 10930 9380736 10.1073/pnas.94.20.10925
Otterbein LE Bach FH Alam J Soares M Tao LH Wysk M Davis RJ Flavell RA Choi AM Carbon monoxide has anti-inflammatory effects involving the mitogen-activated protein kinase pathway Nat Med 2000 6 422 428 10742149 10.1038/74680
Tenhunen R Marver HS Schmid R The enzymatic conversion of heme to bilirubin by microsomal heme oxygenase Proc Natl Acad Sci USA 1968 61 748 755 4386763
Stocker R Yamamoto Y McDonagh AF Glazer AN Ames BN Bilirubin is an antioxidant of possible physiological importance Science 1987 235 1043 1046 3029864
Foresti R Goatly H Green CJ Motterlin R Role of heme oxygenase-1 in hypoxia-reoxygenation: requirement of substrate heme to promote cardioprotection Am J Physiol Heart Circ Physiol 2001 281 H1976 1984 11668058
Clark JE Foresti R Sarathchandra P Kaur H Green CJ Motterlini R Heme oxygenase-1-derived bilirubin ameliorates postischemic myocardial dysfunction Am J Physiol Heart Circ Physiol 2000 278 H643 651 10666097
Harrison PM Arosio P The ferritins: molecular properties, iron storage function and cellular regulation Biochim Biophys Acta 1996 1275 161 203 8695634
Grom AA Natural killer cell dysfunction: A common pathway in systemic-onset juvenile rheumatoid arthritis, macrophage activation syndrome, and hemophagocytic lymphohistiocytosis? Arthritis Rheum 2004 50 689 698 15022306 10.1002/art.20198
Ramanan AV Schneider R Macrophage activation syndrome – what's in a name! J Rheumatol 2003 30 2513 2516 14719185
Henter JI Elinder G Ost A Diagnostic guidelines for hemophagocytic lymphohistiocytosis. The FHL Study Group of the Histiocyte Society Semin Oncol 1991 18 29 33 1992521
Coffernils M Soupart A Pradier O Feremans W Neve P Decaux G Hyperferritinemia in adult onset Still's disease and the hemophagocytic syndrome J Rheumatol 1992 19 1425 1427 1433011
Lee MH Means RT Jr Extremely elevated serum ferritin levels in a university hospital: associated diseases and clinical significance Am J Med 1995 98 566 571 7778572 10.1016/S0002-9343(99)80015-1
Billiau AD Roskams T Van Damme-Lombaerts R Matthys P Wouters C Macrophage activation syndrome: characteristic findings on liver biopsy illustrating the key role of activated, IFN-γ-producing lymphocytes and IL-6- and TNF-α-producing macrophages Blood 2005 105 1648 1651 15466922 10.1182/blood-2004-08-2997
Cush JJ Medsger TA JrChristy WC Herbert DC Cooperstein LA Adult-onset Still's disease. Clinical course and outcome Arthritis Rheum 1987 30 186 194 3827959
Yamaguchi M Ohta A Tsunematsu T Kasukawa R Mizushima Y Kashiwagi H Kashiwazaki S Tanimoto K Matsumoto Y Ota T Preliminary criteria for classification of adult Still's disease J Rheumatol 1992 19 424 430 1578458
Arnett FC Edworthy SM Bloch DA McShane DJ Fries JF Cooper NS Healey LA Kaplan SR Liang MH Luthra HS The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis Arthritis Rheum 1988 31 315 324 3358796
Hochberg MC Updating the American College of Rheumatology revised criteria for the classification of systemic lupus erythmatosus Arthritis Rheum 1997 40 1725 9324032
Bohan A Peter JB Polymyositis and dermatomyositis N Engl J Med 1975 292 344 347 1090839
Bohan A Peter JB Polymyositis and dermatomyositis N Engl J Med 1975 292 403 407 1089199
International Study Group for Behçet's Disease Criteria for diagnosis of Behçet's disease Lancet 1990 335 1078 1080 1970380
van der Heijde DM van't Hof MA van Riel PL Theunisse LA Lubberts EW van Leeuwen MA van Rijswijk MH van de Putte LB Judging disease activity in clinical practice in rheumatoid arthritis: first step in the development of a disease activity score Ann Rheum Dis 1990 49 916 920 2256738
Bombardier C Gladman DD Urowitz MB Caron D Chang CH Derivation of the SLEDAI. A disease activity index for lupus patients. The Committee on Prognosis Studies in SLE Arthritis Rheum 1992 35 630 640 1599520
Kawashima M Yamamura M Taniai M Yamauchi H Tanimoto T Kurimoto M Miyawaki S Amano T Takeuchi T Makino H Levels of interleukin-18 and its binding inhibitors in the blood circulation of patients with adult-onset Still's disease Arthritis Rheum 2001 44 550 560 11263769 10.1002/1529-0131(200103)44:3<550::AID-ANR103>3.0.CO;2-5
Larroche C Mouthon L Pathogenesis of hemophagocytic syndrome (HPS) Autoimmun Rev 2004 3 69 75 15003190 10.1016/S1568-9972(03)00091-0
Yachie A Toma T Mizuno K Okamoto H Shimura S Ohta K Kasahara Y Koizumi S Heme oxygenase-1 production by peripheral blood monocytes during acute inflammatory illnesses of children Exp Biol Med (Maywood) 2003 228 550 556 12709585
Suttner DM Dennery PA Reversal of HO-1 related cytoprotection with increased expression is due to reactive iron FASEB J 1999 13 1800 1809 10506583
Dong Z Lavrovsky Y Venkatachalam MA Roy AK Heme oxygenase-1 in tissue pathology: the Yin and Yang Am J Pathol 2000 156 1485 1488 10793059
Maruhashi K Kasahara Y Ohta K Wada T Ohta K Nakamura N Toma T Koizumi S Yachie A Paradoxical enhancement of oxidative cell injury by overexpression of heme oxygenase-1 in an anchorage-dependent cell ECV304 J Cell Biochem 2004 93 552 562 15378604 10.1002/jcb.20170
ten Kate J Drenth JP Kahn MF van Deursen C Iron saturation of serum ferritin in patients with adult onset Still's disease J Rheumatol 2001 28 2213 2215 11669158
Alam J Camhi S Choi AM Identification of a second region upstream of the mouse heme oxygenase-1 gene that functions as a basal level and inducer-dependent transcription enhancer J Biol Chem 1995 270 11977 11984 7538129 10.1074/jbc.270.20.11977
Wasserman WW Fahl WE Functional antioxidant responsive elements Proc Natl Acad Sci USA 1997 94 5361 5366 9144242 10.1073/pnas.94.10.5361
Janka GE Familial hemophagocytic lymphohistiocytosis Eur J Pediatr 1983 140 221 230 6354720 10.1007/BF00443367
| 15899048 | PMC1174958 | CC BY | 2021-01-04 16:02:36 | no | Arthritis Res Ther. 2005 Mar 21; 7(3):R616-R624 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1721 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17221589904710.1186/ar1722Research ArticleCD134 as target for specific drug delivery to auto-aggressive CD4+ T cells in adjuvant arthritis Boot Elmieke PJ [email protected] Gerben A [email protected] Gert [email protected] Josée PA [email protected] Eden Willem [email protected] Linda A [email protected] Marca HM [email protected] Department of Pharmaceutics, Utrecht Institute for Pharmaceutical Sciences, Utrecht University, Utrecht, The Netherlands2 Division of Immunology, Department of Infectious Diseases and Immunology, Faculty of Veterinary Medicine, Utrecht University, Utrecht, The Netherlands2005 21 3 2005 7 3 R604 R615 7 12 2004 18 1 2005 3 2 2005 24 2 2005 Copyright © 2005 Boot et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
T cells have an important role during the development of autoimmune diseases. In adjuvant arthritis, a model for rheumatoid arthritis, we found that the percentage of CD4+ T cells expressing the activation marker CD134 (OX40 antigen) was elevated before disease onset. Moreover, these CD134+ T cells showed a specific proliferative response to the disease-associated epitope of mycobacterial heat shock protein 60, indicating that this subset contains auto-aggressive T cells. We studied the usefulness of CD134 as a molecular target for immune intervention in arthritis by using liposomes coated with a CD134-directed monoclonal antibody as a drug targeting system. Injection of anti-CD134 liposomes subcutaneously in the hind paws of pre-arthritic rats resulted in targeting of the majority of CD4+CD134+ T cells in the popliteal lymph nodes. Furthermore, we showed that anti-CD134 liposomes bound to activated T cells were not internalized. However, drug delivery by these liposomes could be established by loading anti-CD134 liposomes with the dipalmitate-derivatized cytostatic agent 5'-fluorodeoxyuridine. These liposomes specifically inhibited the proliferation of activated CD134+ T cells in vitro, and treatment with anti-CD134 liposomes containing 5'-fluorodeoxyuridine resulted in the amelioration of adjuvant arthritis. Thus, CD134 can be used as a marker for auto-aggressive CD4+ T cells early in arthritis, and specific liposomal targeting of drugs to these cells via CD134 can be employed to downregulate disease development.
==== Body
Introduction
In several autoimmune diseases, for example rheumatoid arthritis, the involvement of CD4+ T cells in disease induction has been suggested [1]. As a treatment strategy, the manipulation of CD4+ T cells by CD4-directed antibodies has therefore been studied extensively [2]. However, because anti-CD4 therapy targets the whole CD4+ population, CD4+ T cells not related to the disease or involved in disease regulation will also be affected. Ideally, only the auto-aggressive CD4+ T cells that are involved in the disease process should be targeted. Because for many human autoimmune diseases the exact antigens recognized by these cells are not known, a therapy would be favorable that specifically targets the auto-aggressive CD4+ T cells and does not depend on the definition of the crucial auto-antigen.
Because auto-reactive CD4+ T cells become activated upon recognition of their cognate antigen in the periphery, they will be transiently marked by the expression of T cell activation markers. In this respect, CD134 (OX40 antigen) is an interesting candidate target molecule, because CD134 is expressed in vivo exclusively on activated CD4+ T cells (reviewed in [3]). In experimental autoimmune encephalomyelitis, a disease model for multiple sclerosis, it has been shown that CD134 is preferentially expressed on pathogenic CD4+ T cells that home to the target organ (namely the central nervous system) [4], and transiently marks the auto-aggressive T cells specific for myelin basic protein [5]. Moreover, in this T cell transfer model, depletion of CD134+ T cells with an anti-CD134 immunotoxin results in the amelioration of paralytic symptoms [6]. Interestingly, in patients with rheumatoid arthritis a high percentage of CD4+ T cells in synovial fluid express CD134 in comparison with peripheral blood T cells [6,7], suggesting that auto-aggressive CD4+ T cells may be transiently marked by surface expression of CD134 in arthritis too.
Here, we investigated whether CD134 can be used as a target for specific drug delivery to activated auto-aggressive CD4+ T cells in arthritis. For this purpose, the rat adjuvant arthritis (AA) model was studied. In this model, a syndrome resembling rheumatoid arthritis is actively induced in Lewis rats after immunization with Mycobacterium tuberculosis (Mt) in adjuvant [8]. We first analyzed the CD134 expression on CD4+ T cells during AA, and investigated the presence of auto-aggressive T cells within the CD134+CD4+ T cell subset. We also studied drug delivery to CD134+ T cells both in vitro and in vivo using liposomes coated with a CD134-directed monoclonal antibody (mAb) as a drug targeting system. To investigate the possibility for therapeutic intervention in arthritis, anti-CD134 liposomes were loaded with a cytostatic drug and administered early in actively induced arthritis. We show that CD134 can be used as a marker for activated auto-aggressive T cells early in AA, that targeting of these cells in vivo can be achieved with anti-CD134 liposomes, and that the course of AA could be affected with drug-containing anti-CD134 liposomes.
Materials and methods
Animals
Male inbred Lewis rats were obtained from the University of Limburg (Maastricht, The Netherlands) and were used between 7 and 10 weeks of age. The animals were kept under conventional conditions and had access to standard pelleted rat chow and acidified water ad libitum. The Utrecht University Animal Ethics Committee approved all animal experiments.
Antigens
Heat-killed Mt, strain H37RA, was obtained from Difco Laboratories (Detroit, Michigan, USA). For immunization, Mt was suspended in incomplete Freund's adjuvant (Difco Laboratories). Peptides Mt HSP60176–190 (EESNTFGLQLELTEG; one-letter amino acid codes) (HSP60 stands for heat shock protein 60), Mt HSP60211–225 (AVLEDPYILLVSSKV) and OVA323–339 (ISQAVHAAHAEINEAGR) (OVA stands for Ovalbumin) were obtained from Isogen Bioscience (Maarssen, The Netherlands).
mAbs and second-step reagents
The anti-CD134 (OX40) and anti-CD25 (OX39) hybridomas were obtained from the ECACC (Salisbury, UK) [9]. The 12CA5 hybridoma producing IgG2b isotype control mAb was kindly provided by Dr GJ Strous (Department of Cell Biology and Institute of Biomembranes, University Medical Center, Utrecht, The Netherlands). mAbs were isolated from hybridoma supernatant by affinity chromatography with GammaBind Plus Sepharose (Roche Pharmacia, Uppsala, Sweden). For ease of flow cytometric detection, some purified mAbs were biotinylated with D-biotinoyl-ε-aminohexanoic acid-N-hydroxy-succinimide ester (Roche Molecular Biochemicals, Basel, Switzerland). Fluorescein isothiocyanate (FITC)-conjugated anti-CD4 (OX35) and anti-CD45RA (OX33), phycoerythrin (PE)-conjugated goat-anti-mouse immunoglobulin, PE-conjugated streptavidin, peridinin chlorophyll protein (PerCP)-conjugated anti-T-cell antigen receptor (anti-TCR)-αβ (R73) and IgG1 isotype control (A112), and allophycocyanin-conjugated streptavidin were purchased from BD Pharmingen (San Diego, California, USA).
Culture of rat CD4+ T cell clone A2b
The isolation, maintenance, and properties of rat CD4+ T cell clone A2b have been described previously [10]. The arthritogenic T cell clone A2b recognizes the 180 to 188 epitope of Mycobacterium tuberculosis HSP60 [11]. Cells were cultured in medium (Iscove's modified Dulbecco's medium (Invitrogen, Merelbeke, Belgium), supplemented with L-glutamine (2 mM), 2-mercaptoethanol (50 μM), penicillin (50 U/ml) and streptomycin (50 μg/ml)) with 2% heat-inactivated normal rat serum.
Induction of AA
Rats were injected intradermally with 100 μl of Mt in incomplete Freund's adjuvant at the base of the tail. For studying cell-surface marker expression, CD4+ subset specificity during AA and liposome binding in vivo, 10 mg/ml Mt was used. For AA treatment studies, rats were immunized with 5 mg/ml Mt (yielding 100% disease incidence, but lower maximum disease scores in comparison with 10 mg/ml Mt). Rats were weighed and examined for clinical signs of arthritis in a semi-blinded set-up. Severity of arthritis was scored by grading each paw from 0 to 4 based on erythema, swelling and immobility of the joints, resulting in a maximum score of 16 per animal [12].
Ex vivo analysis of cell-surface marker expression
Before Mt immunization or 7, 10, 14, 21 or 35 days afterwards, rats were killed and popliteal lymph nodes (PLN), inguinal lymph nodes (ILN), spleen, and peripheral blood were isolated. Single-cell suspensions were prepared by mechanically forcing the organs through a 70 μm mesh; erythrocytes were removed from the splenocyte and blood suspensions by Ficoll-Isopaque gradient centrifugation. Cells (2 × 105 per sample) were labeled with anti-CD134 for 30 min on ice, followed by incubation with PE-conjugated goat anti-mouse immunoglobulin and subsequently with anti-CD4-FITC. The cells were incubated and washed (between each labeling step) in blocking buffer (PBS (Cambrex Bio Science, Verviers, Belgium) containing 4% heat-inactivated rat serum, 1% fraction V BSA (Sigma-Aldrich Chemie, Zwijndrecht, The Netherlands) and 0.1% NaN3). Finally, cells were washed in PBS, fixed in 2% paraformaldehyde and stored at 4°C in the dark. Cell-associated fluorescence was analyzed within 10 days on a FACSCalibur using Cell Quest software (Becton Dickinson, Brussels, Belgium).
Cell sorting and ex vivo proliferation of CD4+ T cell subsets
At 7 and 10 days after Mt immunization, PLN, ILN, and spleens of 10 to 15 rats were isolated and each organ type was pooled. Single-cell suspensions were prepared as described above. Cells were washed and incubated in PBS containing 4% heat-inactivated rat serum, and stained with anti-CD134-biotin/streptavidin-allophycocyanin and anti-CD4-FITC. CD4+, CD4+CD134- and CD4+CD134+ cells were sorted with a FACS Vantage and Cell Quest software (Becton Dickinson), resulting in fractions that were 87 to 97% pure.
Sorted cells were washed and incubated for 72 hours in medium with 2% heat-inactivated normal rat serum in flat-bottomed 96-well plates (Corning-Costar, Schiphol, The Netherlands) at 5 × 104 cells per well in the presence of 30 Gy-irradiated thymocytes as antigen-presenting cells (APC) (106 cells per well) and concanavalin A (Con A; 2.5 μg/ml) or antigen (20 μg/ml Mt HSP60176–190, 20 μg/ml Mt HSP60211–225). Finally, cells were pulsed for 18 to 20 hours with [3H]thymidine, 0.4 μCi per well (specific radioactivity 1 Ci/mmol; Amersham Biosciences, Roosendaal, The Netherlands), after which [3H]thymidine incorporation was measured. Results are presented as the mean stimulation index (SI, defined as [3H]thymidine incorporation in the presence of antigen or Con A divided by [3H]thymidine incorporation in the absence of antigen or Con A) of triplicate wells. For logistic reasons, ILN single-cell suspensions were kept overnight on ice and were washed, stained and sorted on the following day.
Preparation of (mAb-coupled) PEG-liposomes
Liposomes were composed of egg phosphatidylcholine, cholesterol, poly(ethyleneglycol)2000-distearoylphosphatidylethanolamine (PEG2000-DSPE) and maleimide-PEG2000-DSPE in a molar ratio of 2:1:0.075:0.075. Egg phosphatidylcholine was kindly provided by Lipoid (Ludwigshafen, Germany), PEG2000-DSPE was purchased from Avanti Polar Lipids (Birmingham, Alabama, USA), cholesterol from Sigma-Aldrich Chemie, and Maleimide-PEG2000-DSPE from Shearwater Polymers (Huntsville, Alabama, USA). Liposomes used for investigating liposome binding in vivo contained 0.1 mol% 1,1'-dioctadecyl-3,3,3',3'-tetramethylindodicarbocyanine, 4-chlorobenzenesulfonate salt (DiD; Molecular Probes Europe, Leiden, The Netherlands). Liposomes used for investigating in vitro binding and internalization contained 0.1 mol% Texas red-phosphatidylethanolamine (Molecular Probes Europe). Liposomes used for studying drug delivery in vitro and for the treatment of AA contained 2 mol% 5'-fluoro-2'-deoxyuridine dipalmitate (FUdR-dP; that is, 0.06 mol FudR-dP per 3 mol main lipid constituents) (synthesized as described previously [13]).
Lipids (and FUdR-dP or DiD) were dissolved in chloroform/methanol (9:1) and mixed. A lipid film was prepared through rotary evaporation under vacuum and dried under nitrogen. The lipids were hydrated with HN buffer (4-(2-hydroxyethyl)-1-piperazine ethanesulphonic acid (HEPES) and 135 mM NaCl) at pH 6.7. The resulting vesicles were sized by repeated extrusion through 100 nm polycarbonate filters. Particle size and size distribution were determined by dynamic laser light scattering with an Autosizer 4700 Spectrometer (Malvern Instruments, Malvern, Worcestershire, UK). Liposome preparations had a mean particle diameter ranging from 100 to 200 nm (polydispersity between 0.1 and 0.2). Typically, the mean liposomal diameter varied by less than 20% within any given experiment. The anti-CD134 or IgG2b isotype control mAbs were coupled to liposomes by a thiol-maleimide method described previously [13]. In brief, free thiol groups were introduced in the mAbs using the heterobifunctional reagent N-succinimidyl-S-acetylthioacetate (SATA; Sigma-Aldrich Chemie). Free SATA was separated from the derivatized mAbs by gel permeation chromatography, resulting in ATA-derivatized mAbs dissolved in HN buffer at pH 7.4. mAbs with reactive thiol groups, induced by deacetylating the ATA-protein, were incubated with liposomes at 4°C overnight at a ratio of 0.05 to 0.1 mg of mAbs per μmol lipid. N-ethylmaleimide (8 mM in HN buffer, pH 7.4) was added to cap unreacted thiol groups. Unconjugated mAbs were removed by gel-permeation chromatography or by centrifugation at 100,000 g. The liposomal protein content was determined as described previously [14]. Liposomes contained 25 to 125 μg of mAbs per μmol of lipid. Typically, the mAb content of the different liposome preparations within any given experiment varied by less than 20%.
Liposome binding to CD4+ T cells in vivo
On day 7 after Mt immunization, rats received saline or 5 μmol (lipid) DiD-labeled liposomes subcutaneously (s.c.) in each hind paw. After 30 min the rats were killed, and the PLN, ILN, and spleens were isolated. Single-cell suspensions were prepared as described above. Subsequently, cells were stained with anti-CD4-FITC, anti-CD134-biotin/streptavidin-PE and anti-TCR-αβ-PerCP, or with anti-CD45RA-FITC, anti-CD134-biotin/streptavidin-PE and anti-TCR-αβ-PerCP. Cell-associated fluorescence was measured by flow cytometry.
Liposomal drug delivery to T cells in vitro
A2b T cells were activated in vitro to induce CD134 expression by stimulation overnight with 2.5 μg/ml Con A (Sigma-Aldrich Chemie) in the presence of 30 Gy-irradiated Lewis thymocytes as APC (ratio of T cells to APC = 1:25). Alternatively, a spleen cell suspension (at 2 × 105 cells/ml) was stimulated for 3 days with 2.5 μg/ml Con A to induce CD134 expression on splenic T cells. Next, viable cells were collected from the culture by Ficoll-Isopaque gradient centrifugation and transferred to round-bottomed 96-well plates at 105 cells per sample.
For studying anti-CD134 liposome binding to T cells in vitro by flow cytometry, A2b cells were incubated with 5 nmol (lipid) of anti-CD134 liposomes or IgG2b isotype control liposomes, or anti-CD134 mAb or IgG2b isotype control mAb for 30 min on ice. Cells were then washed and incubated on ice with FITC-conjugated goat anti-mouse immunoglobulin to label the cell-bound mAbs or liposomes. Finally, cell-associated fluorescence was measured. For analysis of the interaction of CD4+ T cells and liposomes by confocal microscopy, activated A2b cells were incubated with 50 nmol of the different liposomal formulations in medium for 30 min on ice, washed and cultured in medium with 2% heat-inactivated normal rat serum at 37°C in 5% CO2. Activated spleen cells were incubated with 100 nmol of liposomes. At the indicated time points, cell-associated fluorescence was assessed.
For assessment of in vitro drug delivery by anti-CD134 liposomes, activated A2b cells were incubated at 37°C in 5% CO2 without or with 1 nmol (lipid) of the different liposomal formulations per well or with an equal concentration (100 nM) of free FUdR (Sigma-Aldrich Chemie) in 200 μl of medium without serum. After 30 min, cells were washed three times in medium and cultured for 48 hours in 200 μl of conditioned medium (medium supplemented with 10% heat-inactivated fetal calf serum (Bodinco, Alkmaar, The Netherlands), 10% culture supernatant of the EL-4 lymphoma (containing murine IL-2) and 1% non-essential amino acids (Invitrogen)). Finally, cells were pulsed for 18 to 20 hours with [3H]thymidine as described above, after which [3H]thymidine incorporation was measured. Results are expressed as the mean percentage of inhibition of proliferation of duplicate cultures relative to the incubation without liposomes (defined as 0%).
Treatment of AA with liposomes and ex vivo proliferation assay of LN cells after liposomal treatment
AA was induced in Lewis rats as described above. Rats received 5 μmol of the different liposome formulations s.c. in each hind paw or HN buffer (see below) as a control on days 3 and 7 or on days 3, 7, and 10 after Mt immunization. Rats were followed for arthritis development as described above.
Proliferation of lymphocytes from liposome-treated animals was measured in quadruple cultures of 2 × 105 cells per well without additional APC. Cells were cultured in 96-well flat-bottomed plates in 200 μl of medium containing 2% heat-inactivated normal rat serum in the absence or presence of antigen (20 μg/ml Mt HSP60176–190 or 20 μg/ml OVA232–339) or Con A (2.5 μg/ml). After 72 hours, cells were pulsed for 18 to 20 hours with [3H]thymidine as described above, after which [3H]thymidine incorporation was measured.
Statistical evaluation
The statistical significance of differences was evaluated with GraphPad Prism 3.02 (GraphPad Software, San Diego, California, USA). For statistical analysis of CD134 expression in vivo and liposomal drug delivery in vitro, a one-way analysis of variance with Dunnett's post-hoc test was used. For analysis of differences in the development of AA, a Mann-Whitney test was used for arthritis scores and an unpaired Student's t-test for body weight.
Results
Expression of CD134 on CD4+ T cells during AA
To study the expression of CD134 and CD4 during AA, Lewis rats were immunized with Mt in adjuvant. The first signs of inflammation of the paw joints were observed between days 10 and 14, and the disease reached maximum severity at days 20 to 22. After this, inflammation of paw joints gradually decreased and resolved macroscopically at days 35 to 40. At several time points during AA development, the PLN (which drain the foot and ankle joints), the ILN (which drain the Mt immunization site), the spleen, and blood were isolated and examined by flow cytometry.
Seven days after Mt immunization, before the clinical onset of AA, the percentage of CD134+ T cells was increased both in the PLN and ILN in comparison with naive animals (day 0; Fig. 1a,b). In the ILN this percentage remained elevated throughout the active disease phase between days 10 and 30 (Fig. 1b). In the PLN a decrease in the percentage of CD134+ T cells on days 10 and 14 was observed. On day 21 the percentage of CD134+CD4+ cells was found to increase again (Fig. 1a). The total cell number in the PLN at day 7 (8.5 × 106 ± 1.6 × 106, mean ± SEM) and day 10 (8.3 × 106 ± 2.7 × 106) was comparable, as well as the percentage of CD4+ cells (percentage of live lymphocytes; Fig. 1d). This indicated that the absolute number of CD134+CD4+ T cells decreased during the interval from day 7 to day 10. In the spleen, the main increase in the percentage of CD134-expressing T cells was observed at about day 14 (Fig. 1c). In peripheral blood, no changes were detected in the percentage of CD134+ cells during the onset of AA (data not shown). We did not observe a significant increase in the percentage of CD4+ cells during AA in any of the organs tested (Fig. 1d–f). The data shown in Fig. 1 represent a compilation of four separate experiments, in which all rats were immunized on one day and flow cytometric analysis was performed on separate days. Another experiment in which rats were immunized on separate days (n = 4 rats per time point), and flow cytometric analysis was performed on one day, yielded similar results (data not shown).
Specific responsiveness of CD134+ T cells to the disease-associated epitope of Mt HSP60
Previously, it has been shown that a CD4+ T cell clone (clone A2b [15]), derived from a Lewis rat after Mt immunization and capable of transferring arthritis to naive rats, recognized a T cell epitope present in the 176–190 region of Mt HSP60 (Mt HSP60176–190) [11]. To investigate whether CD134+ T cells early in AA were potentially arthritogenic, we tested CD134+CD4+ T cells isolated at days 7 and 10 after Mt immunization – that is, just before the onset of clinical disease – for their proliferative response to peptide Mt HSP60176–190. The results for day 10 are presented in Fig. 2. PLN-derived CD4+ cells showed a low proliferative response to the disease-associated peptide (SI ≈ 3; Fig. 2). When the CD4+ population was divided into CD134+ and CD134- fractions, the Mt HSP60176–190 response of the CD4+ cells was entirely attributable to the CD134+ cells, as these cells showed a high proliferative activity to Mt HSP60176–190, whereas this response was absent in the CD134- population (SI < 2). Similar results were found for the CD4+ subsets isolated from ILN and spleen (Fig. 2). The isolated CD134+CD4+ cells also showed a response to another mycobacterial HSP60 epitope, peptide 211–225, which has been reported not to be related to AA [16]. However, this response was much lower than the Mt HSP60176–190 response (Fig. 2). Data obtained at day 7 (data not shown) and day 10 were similar. Thus, the CD134+ T cell population found early in AA was enriched for activated auto-aggressive CD4+ T cells, as shown by the specific response to the disease-associated epitope Mt HSP60176–190.
Specific targeting to CD134+ T cells in draining lymph nodes with anti-CD134 liposomes
For delivery of modulating compounds to the potentially arthritogenic CD134+ T cells, we selected a mAb-targeted liposomal system. To investigate whether the CD134+ CD4+ T cells in the draining LN could be targeted in vivo, fluorescent anti-CD134 liposomes were injected s.c. in the hind paws of rats on day 7 after Mt immunization. After 30 min, rats were killed, and the T cells in the joint-draining PLN, the immunization site-draining ILN and spleen were studied for CD134 expression and for the presence of cell-bound liposomes by flow cytometry. In the PLN, 10.7% of the cells were found to be both CD4+ and CD134+, whereas 7.5% of the cells had bound anti-CD134 liposomes and were CD4+ (Fig. 3a). Competitive counterstaining of liposome+CD4+ cells with anti-CD134 mAbs showed that virtually all these cells were indeed CD134+ (Fig. 3b). This implies that the vast majority of CD134+ T cells in the PLN was targeted. In addition, also CD4- cells were targeted in the PLN (Fig. 3a). In this case, however, binding of both anti-CD134- and isotype control liposomes was comparable and these cells were determined to be CD45RA+ B cells (Fig. 3c). Because B cells do not express CD134, the anti-CD134 binding could not be due to CD134 binding. The similar staining pattern of isotype control liposomes and anti-CD134 liposomes on B cells suggested that this binding was due to Fc-mediated binding (because whole mAb was used to coat liposomes). In the ILN or the spleen virtually no CD134+ CD4+ T cells were targeted by anti-CD134 liposomes administered s.c. in the paw (Fig. 3a).
Drug delivery to CD134+ T cells in vitro using anti-CD134 liposomes containing the dipalmitate-anchored cytostatic agent FUdR
The fate of anti-CD134 liposomes after binding to activated CD4+ T cells was studied in vitro by incubation of activated T cells of clone A2b with anti-CD134 liposomes. Activated CD134+ A2b T cells were shown to specifically bind anti-CD134 liposomes (Fig. 4a). Interestingly, although resting A2b cells seemed CD134- after conventional mAb staining, anti-CD134 liposomes did bind to these cells to a small extent.
Using confocal microscopy, anti-CD134 liposomes were shown to bind specifically to activated CD4+ T cells in a diffuse pattern; that is, spread out over the plasma membrane (Fig. 4b). When cells that had bound liposomes were incubated at 37°C, the staining pattern of anti-CD134 liposomes changed from diffuse to a more focal pattern after 2 hours of culture at 37°C (Fig. 4b, 2 hours). However, no internalization of anti-CD134 liposomes was observed at any of the time points evaluated. This was also observed with Con A-activated splenic T cells (data not shown). As a positive control for liposome internalization, liposomes targeting CD25, the α-subunit of the IL-2 receptor, which is also expressed on activated CD4+ T cells, were used (Fig. 4b, 4 hours). Furthermore, activated CD4+ T cells were able to internalize anti-CD134 mAbs (and anti-CD25 mAbs) within 2 hours of binding (data not shown). This indicated that although the CD134 receptor itself was internalized, cell-bound anti-CD134 liposomes were not internalized by the targeted T cells.
Our finding that anti-CD134 liposomes were not internalized by the target T cells had major implications for the strategy of drug delivery. We decided to use the mechanism of lipid-coupled drug transfer between membranes to achieve intracellular drug delivery. The lipid-derivatized cytostatic agent FUdR-dP was used as a model drug [13]. Activated, CD134+ rat T cells of clone A2b were found to be very sensitive to free FUdR (90% growth inhibition with 100 nM FUdR) when FUdR was continuously present during culture for 48 hours. However, when the cells were incubated for only 30 min with free FUdR and then washed to remove extracellular FUdR, no significant growth inhibition was detected during subsequent culture (Table 1). When the equivalent amount of FUdR was present in 1 nmol anti-CD134 liposomes (anti-CD134-FUdR-dP liposomes), which were incubated for 30 min with the cells, the proliferation of activated A2b cells was inhibited by more than 30% (Table 1). Incubation of activated T cells with anti-CD134 liposomes without FUdR-dP or with non-targeted FUdR-dP liposomes did not significantly affect the proliferation of the cells (P > 0.05; Table 1).
Modulation of AA by treatment with drug-containing anti-CD134 liposomes
Next, we investigated whether local targeting to CD134+ T cells in the joint-draining PLN would affect the course of actively induced arthritis in the AA model. Rats were injected with the different liposomal formulations on days 3 and 7 after Mt immunization and were followed for arthritis development. Injection of anti-CD134-FUdR-dP liposomes resulted in less severe disease development than in rats injected with anti-CD134 liposomes without FUdR-dP (Fig. 5a). This effect was increased by the administration of anti-CD134-FudR-dP liposomes on three occasions (Fig. 5b). The improved well-being of the anti-CD134-FUdR-dP liposome-treated rats was also reflected in a faster recovery of weight (Fig. 5a). The isotype control FUdR-dP liposomes also seemed to affect the progression of AA, although anti-CD134-FUdR-dP liposomes were more effective. No difference in AA scores was found between rats treated with empty anti-CD134 liposomes and rats treated with empty, bare liposomes (data not shown).
The modulation of the course of AA after treatment with anti-CD134-FUdR-dP liposomes was correlated with a decreased proliferative response to Mt HSP60176–190 of joint-draining PLN cells isolated at day 42 (Fig. 5c). This is indicative of successful targeting and deletion of Mt HSP60176–190-reactive, CD134+ T cells in vivo.
Discussion
In the present study we investigated whether CD134 can be used as a (transient) marker for targeting auto-aggressive CD4+ T cells in actively induced experimental arthritis. Before the onset of clinical arthritis, an elevated percentage of CD134+ CD4+ T cells was found in the PLN, which drain the foot and ankle joints, and in the ILN, which drain the Mt immunization site. In the ILN, this percentage remained elevated throughout the active disease phase, indicating a continuous activation of T cells, probably because of the presence of an Mt depot at the base of the tail. However, in the PLN the percentage and absolute number of CD134+ T cells were decreased at days 10 and 14 after immunization in comparison with the initial elevation on day 7. In this arthritis model, the first signs of clinical disease become manifest between days 10 and 14, and at about this time T cells start to infiltrate the joints [17]. The present data therefore suggest that early in AA, CD134+ T cells, after activation in the PLN, can migrate to the joints, where they subsequently become involved in joint inflammation. The second increase in the percentage of CD134-expressing T cells in the PLN on days 21 and 35 could reflect the recirculation or generation of activated auto-aggressive T cells, or the emergence of an activated regulatory population.
The presence of arthritogenic cells in the CD134+ subset of CD4+ T cells isolated from pre-arthritic rats was deduced from the high proliferative response to the Mt HSP60176–190 peptide, which was previously linked to the induction of AA [11]. However, the CD134+ T cell subset also includes activated (auto-aggressive) T cells with a different specificity. The low but evident response of isolated CD4+CD134+ cells to Mt HSP60211–225, which was reported not to be related to AA [16], indeed indicated that a part of the CD134+ T cells was not activated in relation to clinical disease but responded to other epitopes present in the immunization mix. This underlines the fact that, although CD134 may be used to select for activated pathogenic CD4+ T cells in an autoimmune setting, this molecule is primarily a marker for activated CD4+ T cells in general [18-20]. Nevertheless, by using CD134 as marker for targeting, we expected to affect all recently activated auto-aggressive CD4+ T cells present at the time of targeting, including arthritogenic T cells with a different specificity from that of Mt HSP60176–190. The preferential expression of CD134 on synovial fluid CD4+ T cells from patients with rheumatoid arthritis, as has been demonstrated by others [6,7,21], indicates that also in humans auto-reactive T cells might be (transiently) marked by CD134. Because CD134 ligand expression has been demonstrated both on vascular endothelial cells [22,23] and in synovial tissue of rheumatoid arthritis patients [21], the recruitment and in situ restimulation of activated T cells through CD134 possibly contributes to the inflammatory process in arthritis. Indeed, in a mouse collagen-induced arthritis model, treatment with a mAb blocking anti-CD134 ligand did inhibit disease development [21].
To explore the possibility for modulating auto-aggressive T cells in arthritis, we examined the potential of drug targeting directly to CD134+ T cells in AA by using liposomes as drug carriers. To study the ability of anti-CD134 liposomes to reach the potentially auto-reactive CD134+ T cells in vivo, the active disease model was employed, because this would allow targeting of the target T cells during priming in situ; that is, in the secondary lymphoid organs. When anti-CD134 liposomes were injected s.c. in the hind paws, the majority of the CD4+CD134+ T cells in the joint-draining PLN could indeed be targeted. The non-T cells in the PLN that were found to bind both anti-CD134 liposomes and isotype control liposomes were determined as being B cells that most probably bound the liposomes in a Fc-mediated fashion.
Activated CD134+ T cells targeted by anti-CD134 liposomes in vitro did not internalize the cell-bound liposomes. This lack of internalization determined the strategy of drug delivery. We here employed the mechanism of lipid-coupled drug transfer between membranes to achieve intracellular drug delivery. When anti-CD134 liposomes carried the lipid-coupled cytostatic agent FUdR as a model drug, a 30% inhibition of proliferation of activated CD134+ T cells was observed in vitro. This inhibitory effect on the proliferation of CD134+ T cells in vitro was correlated with a moderate suppression of AA in rats treated with anti-CD134-FUdR-dP liposomes. The effect of these liposomes on AA development was supported by a downregulation of the disease-associated Mt HSP60176–190 response in the PLN of anti-CD134-FUdR-dP liposome-treated animals.
The effect of isotype control-FUdR-dP on clinical disease might be due to their association with B cells in vivo, probably through binding to Fc receptors [24]. Although B cells have not been described as having a crucial role in the development of AA [25], contrary to collagen-induced arthritis in mice, for example [26], these cells can function as APC and as such may affect the response of auto-aggressive T cells in vivo. Optimizing the therapeutic entity, for example by coupling anti-CD134 Fab fragments to the liposomal carrier instead of the entire anti-CD134 antibody, would circumvent B cell targeting.
It has recently been shown that CD4+CD25+ regulatory T cells involved in the control of autoimmunity [27] can express CD134 [28-30]. Although CD134+ regulatory T cells may go through a proliferative phase in vivo [31,32], in general these cells display a hypoproliferative phenotype in vitro as well as in vivo [33]. Because cytostatic agents act primarily on proliferating cells, it is possible that by employing FUdR-dP-containing anti-CD134 liposomes we largely preserved this regulatory T cell subset.
Conclusion
We show here that CD134 can be used as a marker for recently activated CD4+ T cells with auto-aggressive potential in arthritis, and that anti-CD134 liposomes can be used to target drugs directly to these T cells. Thus, anti-CD134 liposomes represent an attractive method for the development of therapies aiming at the modulation of auto-aggressive T cells for intervention in autoimmune diseases.
Abbreviations
AA = adjuvant arthritis; APC = antigen-presenting cells; Con A = concanavalin A; DiD = 1,1'-dioctadecyl-3,3,3',3'-tetramethylindodicarbocyanine, 4-chlorobenzenesulfonate salt; FITC = fluorescein isothiocyanate; FUdR(-dP) = 5'-fluoro-2'-deoxyuridine (dipalmitate); HSP = heat shock protein; ILN = inguinal lymph nodes; mAb = monoclonal antibody; Mt = Mycobacterium tuberculosis; PBS = phosphate-buffered saline; PE = phycoerythrin; PerCP = peridinin chlorophyll protein; PLN = popliteal lymph nodes; s.c. = subcutaneously; SI = stimulation index; TCR = T-cell antigen receptor.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
EB participated in designing performing the experiments and prepared the manuscript. GK participated in, and supervised, the liposome preparations. GS participated in the design and coordination of the study, and in the interpretion of the results. JWH carried out immunological experiments and participated in the interpretation of the results. WVE participated in the design of the study and in its coordination. LE participated in the experiments, statistical analysis and interpretation of the results. MW supervised the design and execution of the study, and drafted the manuscript. All authors read and approved the final manuscript.
Acknowledgements
We thank Dr Ger Arkesteijn for technical assistance with cell sorting, Ing. Louis van Bloois for technical assistance with preparing liposomes, Ing. Peter J van Kooten and Ing. Mayken CJT Grosfeld-Stulemeijer for technical assistance in production and purification of mAbs, and Dr Martijn A Nolte for valuable advice. The research described in this study is part of the UNYPHAR project, a research network between Yamanouchi Europe BV and the Universities of Groningen, Leiden, and Utrecht.
Figures and Tables
Figure 1 CD134 is differentially expressed on CD4+ cells in secondary lymphoid organs during adjuvant arthritis. Popliteal lymph nodes (a,d), inguinal lymph nodes (b,e), and spleens (c,f) were isolated from Lewis rats before or during adjuvant arthritis (AA) development. Cell suspensions were stained for CD4 and CD134, and cell-associated fluorescence was analyzed by flow cytometry. Results for CD134 (black bars) are depicted as percentages of CD134+ cells of the CD4+ cell population and are expressed as means ± SEM (corrected for isotype control fluorescence). Results for CD4 (white bars) are shown as percentages of CD4+ cells of the live lymphocytes population and are expressed as means ± SEM (live lymphocytes gated based on forward scatter (FSC) and side scatter (SSC) profiles). The data shown are derived from four independent experiments and represent n = 5 to 7 rats per group for t = 0, n = 2 rats per group for t = 7, n = 3 to 7 rats per group for t = 10, n = 5 to 8 rats per group for t = 14, n = 5 to 9 rats per group for t = 21, and n = 4 to 6 rats per group for t = 35. *P < 0.05 compared with t = 0, **P < 0.01 compared with t = 0.
Figure 2 CD134+ T cells recognize the disease-associated mycobacterial epitope early in adjuvant arthritis. Popliteal lymph nodes, inguinal lymph nodes, and spleens were isolated from n = 13 rats at day 10 after immunization with Mycobacterium tuberculosis (Mt). The organs were pooled by organ type, and single-cell suspensions were stained for CD4 and CD134. The cells were sorted into CD4+ (white bars), CD4+CD134- (hatched bars), and CD4+CD134+ (black bars) fractions. Proliferative responses to 20 μg/ml Mt HSP60176–190 (in which HSP60 stands for heat shock protein 60) were tested in a [3H]thymidine incorporation assay. As a control, the proliferation in response to 20 μg/ml Mt HSP60211–225 (not related to disease) was tested. Results are expressed as the mean SI of triplicate wells. The cut-off value for proliferation was set at SI = 2 (indicated by the horizontal line). Shown is one representative experiment of three.
Figure 3 CD134+ T cells in joint-draining lymph nodes are targeted by subcutaneous injection of anti-CD134 liposomes. On day 7 after immunization with Mycobacterium tuberculosis (Mt), rats were injected subcutaneously with fluorescent isotype control liposomes or anti-CD134 liposomes. Rats were killed 30 min later, and popliteal lymph nodes (PLN), inguinal lymph nodes, and spleens were isolated. (a) Cells were stained for CD4 and T-cell antigen receptor (TCR)-αβ and cell-associated fluorescence was analyzed by flow cytometry. Dot plots show cell-associated fluorescence due to in vitro monoclonal antibody (mAb) staining (left panels) or in vivo liposome binding (right panels). Cells were gated for live TCR-αβ+ CD4+ cells. The numbers in the dot plots indicate the percentage of cells above the cut-off line, which was set by using non-stained cells from sham-injected animals. Three rats were analyzed per group; representative stainings of one rat per group were selected and are shown here. (b) PLN cells of anti-CD134 liposome-injected rats were stained with anti-CD4 and anti-CD134 or its isotype control. Cells were gated for live CD4+liposome+ cells. Histograms show cell-associated fluorescence due to the binding of anti-CD134 (filled) or isotype control mAb (open). Representative stainings of one rat of three are shown. (c) PLN cells were stained with anti-TCR-αβ and anti-CD45RA (rat B cells). Cells were gated for live, TCR-αβ-, and liposome+ cells. Histograms show cell-associated fluorescence due to ex vivo CD45RA (filled histogram) or isotype control mAb staining (thin line) on anti-CD134 liposome+ cells, or CD45RA (thick line) mAb staining on isotype control liposome+ cells. Representative stainings of one rat of three are shown.
Figure 4 Anti-CD134-mediated targeting does not lead to liposome internalization by activated CD4+ T cells in vitro. A2b T cells were cultured with antigen-presenting cells and Con A to induce CD134 expression; CD25 is expressed constitutively on these cells. (a) Viable T cells were incubated with isotype control liposomes (filled histogram) or with anti-CD134 liposomes (black line). As a control, the binding of anti-CD134 liposomes or anti-CD134 monoclonal antibodies to resting T cells was assessed (gray lines). Cell-associated fluorescence was analyzed by flow cytometry, with live cells gated on the basis of forward scatter (FSC) and side scatter (SSC) profiles. One representative experiment of three is shown. (b) Viable T cells were incubated for 30 min with anti-CD134 liposomes on ice. After the removal of non-bound liposomes by washing, cells were cultured subsequently at 37°C. Samples were taken at the indicated time points and analyzed for the cellular localization of the liposomal fluorescence with the use of confocal microscopy. A representative cell from each time point is shown. As a positive control for cellular internalization of liposomes, cells incubated with anti-CD25 liposomes are shown. One of two experiments, yielding similar results, is shown.
Figure 5 Adjuvant arthritis is modulated by treatment with anti-CD134 5'-fluoro-2'-deoxyuridine dipalmitate (FUdR-dP) liposomes. (a) Rats were immunized with Mycobacterium tuberculosis (Mt) to induce arthritis. On days 3 and 7, rats received isotype control FUdR-dP liposomes (filled triangles), anti-CD134-FUdR-dP liposomes (filled circles), or empty anti-CD134 liposomes (open circles) subcutaneously (s.c.) in both hind paws. (b) Alternatively, after immunization with Mt, on days 3 and 7 rats received anti-CD134-FUdR-dP liposomes (filled circles) or empty bare liposomes (open diamonds), or on days 3, 7, and 10 anti-CD134-FUdR-dP liposomes (filled squares), s.c. in both hind paws. Rats were followed for the development of clinical disease and body weight until the disease resolved spontaneously (day 37 to 42). Results are expressed as the arthritis score and the mean body weight (percentage of day 0) per group of n = 5 rats and are presented as means ± SEM. Statistical differences are indicated in the plots. (c) On day 42, rats shown in (a) were killed; popliteal lymph node cells were isolated and pooled from each treatment group. Cells from isotype control FUdR-dP liposome-treated rats (white bars), anti-CD134-FUdR-dP liposome-treated rats (black bars), and empty anti-CD134 liposome-treated rats (hatched bars) were tested for their proliferative response to 20 μg/ml Mt HSP60176–190 peptide (in which HSP60 stands for heat shock protein 60) in a [3H]thymidine incorporation assay. The proliferation to 20 μg/ml peptide OVA323–339 is shown as a negative control. Results are expressed as the mean SI for quadruple wells. The cut-off value for proliferation was set at SI 2 (indicated by line).
Table 1 Anti-CD134 5'-fluoro-2'-deoxyuridine dipalmitate liposomes inhibit the proliferation of CD134+ T cells in vitro
Drug Drug carrier Targeting moiety Inhibition of proliferation (%)
None None None 0.0 ± 8.2
FUdR None Anti-CD134 8.9 ± 8.7
FUdR-dP Liposomes None 6.3 ± 0.7
None Liposomes Anti-CD134 0.6 ± 1.1
FUdR-dP Liposomes Anti-CD134 31.0 ± 2.3*
Con A-activated CD134+ T cells of clone A2b were incubated for 30 min without ('none' under the heading 'drug') or with either free 5'-fluoro-2'-deoxyuridine (FUdR), 5'-fluoro-2'-deoxyuridine dipalmitate (FUdR-dP) liposomes, anti-CD134 liposomes, or anti-CD134-FUdR-dP liposomes. To each well was added 100 nM FUdR or 1 nmol of liposomal lipid (for FUdR-dP liposomes this equals 100 nM FUdR). Subsequently, cells were washed and cultured for 48 hours, followed by [3H]thymidine incorporation as a measure of proliferation. Results are expressed as the mean percentage of inhibition of proliferation relative to the incubation without liposomes (defined as 0%; mean [3H]thymidine incorporation of 53,330 c.p.m.). Results in the last column are means ± SEM. One representative experiment of three is shown. *P < 0.001 compared with incubation without liposomes.
==== Refs
Panayi GS Corrigall VM Pitzalis C Pathogenesis of rheumatoid arthritis. The role of T cells and other beasts Rheum Dis Clin North Am 2001 27 317 334 11396095
Schulze-Koops H Lipsky PE Anti-CD4 monoclonal antibody therapy in human autoimmune diseases Curr Dir Autoimmun 2000 2 24 49 11791458
Sugamura K Ishii N Weinberg AD Therapeutic targeting of the effector T-cell co-stimulatory molecule OX40 Nat Rev Immunol 2004 4 420 431 15173831 10.1038/nri1371
Weinberg AD Wallin JJ Jones RE Sullivan TJ Bourdette DN Vandenbark AA Offner H Target organ-specific up-regulation of the MRC OX-40 marker and selective production of Th1 lymphokine mRNA by encephalitogenic T helper cells isolated from the spinal cord of rats with experimental autoimmune encephalomyelitis J Immunol 1994 152 4712 4721 7512604
Weinberg AD Lemon M Jones AJ Vainiene M Celnik B Buenafe AC Culbertson N Bakke A Vandenbark AA Offner H OX-40 antibody enhances for autoantigen specific V beta 8.2+ T cells within the spinal cord of Lewis rats with autoimmune encephalomyelitis J Neurosci Res 1996 43 42 49 8838572 10.1002/jnr.490430105
Weinberg AD Bourdette DN Sullivan TJ Lemon M Wallin JJ Maziarz R Davey M Palida F Godfrey W Engleman E Selective depletion of myelin-reactive T cells with the anti-OX-40 antibody ameliorates autoimmune encephalomyelitis Nat Med 1996 2 183 189 8574963 10.1038/nm0296-183
Brugnoni D Bettinardi A Malacarne F Airo P Cattaneo R CD134/OX40 expression by synovial fluid CD4+ T lymphocytes in chronic synovitis Br J Rheumatol 1998 37 584 585 9651095 10.1093/rheumatology/37.5.584
Pearson CM Development of arthritis, periarthritis and periostitis in rats given adjuvant Proc Soc Exp Biol Med 1956 91 95 101 13297719
Paterson DJ Jefferies WA Green JR Brandon MR Corthesy P Puklavec M Williams AF Antigens of activated rat T lymphocytes including a molecule of 50,000 Mr detected only on CD4 positive T blasts Mol Immunol 1987 24 1281 1290 2828930 10.1016/0161-5890(87)90122-2
Ben-Nun A Wekerle H Cohen IR The rapid isolation of clonable antigen-specific T lymphocyte lines capable of mediating autoimmune encephalomyelitis Eur J Immunol 1981 11 195 199 6165588
van Eden W Thole JE van der Zee R Noordzij A van Embden JD Hensen EJ Cohen IR Cloning of the mycobacterial epitope recognized by T lymphocytes in adjuvant arthritis Nature 1988 331 171 173 2448638 10.1038/331171a0
Wood FD Pearson CM Tanaka A Capacity of mycobacterial wax D and its subfractions to induce adjuvant arthritis in rats Int Arch Allergy Appl Immunol 1969 35 456 467 4976512
Koning GA Morselt HW Velinova MJ Donga J Gorter A Allen TM Zalipsky S Kamps JA Scherphof GL Selective transfer of a lipophilic prodrug of 5-fluorodeoxyuridine from immunoliposomes to colon cancer cells Biochim Biophys Acta 1999 1420 153 167 10446299
Peterson GL A simplification of the protein assay method of Lowry et al. which is more generally applicable Anal Biochem 1977 83 346 356 603028 10.1016/0003-2697(77)90043-4
Holoshitz J Matitiau A Cohen IR Arthritis induced in rats by cloned T lymphocytes responsive to mycobacteria but not to collagen type II J Clin Invest 1984 73 211 215 6228565
Anderton SM van der Zee R Prakken B Noordzij A van Eden W Activation of T cells recognizing self 60-kD heat shock protein can protect against experimental arthritis J Exp Med 1995 181 943 952 7869052 10.1084/jem.181.3.943
Issekutz TB Issekutz AC T lymphocyte migration to arthritic joints and dermal inflammation in the rat: differing migration patterns and the involvement of VLA-4 Clin Immunol Immunopathol 1991 61 436 447 1934631
Gramaglia I Weinberg AD Lemon M Croft M Ox-40 ligand: a potent costimulatory molecule for sustaining primary CD4 T cell responses J Immunol 1998 161 6510 6517 9862675
Maxwell JR Weinberg A Prell RA Vella AT Danger and OX40 receptor signaling synergize to enhance memory T cell survival by inhibiting peripheral deletion J Immunol 2000 164 107 112 10605000
Rogers PR Song J Gramaglia I Killeen N Croft M Ox40 promotes Bcl-xl and Bcl-2 expression and is essential for long-term survival of CD4 T cells Immunity 2001 15 445 455 11567634 10.1016/S1074-7613(01)00191-1
Yoshioka T Nakajima A Akiba H Ishiwata T Asano G Yoshino S Yagita H Okumura K Contribution of OX40/OX40 ligand interaction to the pathogenesis of rheumatoid arthritis Eur J Immunol 2000 30 2815 2823 11069062 10.1002/1521-4141(200010)30:10<2815::AID-IMMU2815>3.0.CO;2-#
Imura A Hori T Imada K Ishikawa T Tanaka Y Maeda M Imamura S Uchiyama T The human OX40/gp34 system directly mediates adhesion of activated T cells to vascular endothelial cells J Exp Med 1996 183 2185 2195 8642328 10.1084/jem.183.5.2185
Souza HS Elia CC Spencer J MacDonald TT Expression of lymphocyte-endothelial receptor-ligand pairs, alpha4beta7/MAdCAM-1 and OX40/OX40 ligand in the colon and jejunum of patients with inflammatory bowel disease Gut 1999 45 856 863 10562584
Phillips NE Parker DC Cross-linking of B lymphocyte Fc gamma receptors and membrane immunoglobulin inhibits anti-immunoglobulin-induced blastogenesis J Immunol 1984 132 627 632 6228594
Taurog JD Sandberg GP Mahowald ML The cellular basis of adjuvant arthritis. II. Characterization of the cells mediating passive transfer Cell Immunol 1983 80 198 204 6603276 10.1016/0008-8749(83)90106-5
Durie FH Fava RA Noelle RJ Collagen-induced arthritis as a model of rheumatoid arthritis Clin Immunol Immunopathol 1994 73 11 18 7923907 10.1006/clin.1994.1164
Sakaguchi S Sakaguchi N Asano M Itoh M Toda M Immunologic self-tolerance maintained by activated T cells expressing IL-2 receptor alpha-chains (CD25). Breakdown of a single mechanism of self-tolerance causes various autoimmune diseases J Immunol 1995 155 1151 1164 7636184
McHugh RS Whitters MJ Piccirillo CA Young DA Shevach EM Collins M Byrne MC CD4+CD25+ immunoregulatory T cells: gene expression analysis reveals a functional role for the glucocorticoid-induced TNF receptor Immunity 2002 16 311 323 11869690 10.1016/S1074-7613(02)00280-7
Stephens LA Barclay AN Mason D Phenotypic characterization of regulatory CD4+CD25+ T cells in rats Int Immunol 2004 16 365 375 14734622 10.1093/intimm/dxh033
Nolte-'t Hoen EN Wagenaar-Hilbers JP Boot EP Lin CH Arkesteijn GJ van Eden W Taams LS Wauben MH Identification of a CD4+CD25+ T cell subset committed in vivo to suppress antigen-specific T cell responses without additional stimulation Eur J Immunol 2004 34 3016 3027 15376196 10.1002/eji.200324725
Fisson S Darrasse-Jeze G Litvinova E Septier F Klatzmann D Liblau R Salomon BL Continuous activation of autoreactive CD4+ CD25+ regulatory T cells in the steady state J Exp Med 2003 198 737 746 12939344 10.1084/jem.20030686
Klein L Khazaie K von Boehmer H In vivo dynamics of antigen-specific regulatory T cells not predicted from behavior in vitro Proc Natl Acad Sci U S A 2003 100 8886 8891 12857946 10.1073/pnas.1533365100
Gavin MA Clarke SR Negrou E Gallegos A Rudensky A Homeostasis and anergy of CD4+CD25+ suppressor T cells in vivo Nat Immunol 2002 3 33 41 11740498 10.1038/ni743
| 15899047 | PMC1174959 | CC BY | 2021-01-04 16:02:36 | no | Arthritis Res Ther. 2005 Mar 21; 7(3):R604-R615 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1722 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17241589905210.1186/ar1724Research ArticleDermatological conditions during TNF-α-blocking therapy in patients with rheumatoid arthritis: a prospective study Flendrie Marcel [email protected] Wynand HPM [email protected] Marjonne CW [email protected] Jong Elke MGJ [email protected] de Kerkhof Peter CM 2van Riel Piet LCM 11 Department of Rheumatology, Radboud University Nijmegen Medical Centre, Nijmegen, the Netherlands2 Department of Dermatology, Radboud University Nijmegen Medical Centre, Nijmegen, the Netherlands2005 4 4 2005 7 3 R666 R676 3 1 2005 20 1 2005 25 2 2005 1 3 2005 Copyright © 2005 Flendrie et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Various dermatological conditions have been reported during tumor necrosis factor (TNF)-α-blocking therapy, but until now no prospective studies have been focused on this aspect. The present study was set up to investigate the number and nature of clinically important dermatological conditions during TNF-α-blocking therapy in patients with rheumatoid arthritis (RA). RA patients starting on TNF-α-blocking therapy were prospectively followed up. The numbers and natures of dermatological events giving rise to a dermatological consultation were recorded. The patients with a dermatological event were compared with a group of prospectively followed up RA control patients, naive to TNF-α-blocking therapy and matched for follow-up period. 289 RA patients started TNF-α-blocking therapy. 128 dermatological events were recorded in 72 patients (25%) during 911 patient-years of follow-up. TNF-α-blocking therapy was stopped in 19 (26%) of these 72 patients because of the dermatological event. More of the RA patients given TNF-α-blocking therapy (25%) than of the anti-TNF-α-naive patients (13%) visited a dermatologist during follow-up (P < 0.0005). Events were recorded more often during active treatment (0.16 events per patient-year) than during the period of withdrawal of TNF-α-blocking therapy (0.09 events per patient-year, P < 0.0005). The events recorded most frequently were skin infections (n = 33), eczema (n = 20), and drug-related eruptions (n = 15). Other events with a possible relation to TNF-α-blocking therapy included vasculitis, psoriasis, drug-induced systemic lupus erythematosus, dermatomyositis, and a lymphomatoid-papulosis-like eruption. This study is the first large prospective study focusing on dermatological conditions during TNF-α-blocking therapy. It shows that dermatological conditions are a significant and clinically important problem in RA patients receiving TNF-α-blocking therapy.
==== Body
Introduction
The introduction of biological agents such as TNF-α-blocking agents has dramatically changed the therapeutic approach to rheumatic diseases in recent years. TNF-α-blocking therapy has had a remarkable effect on disease activity in an increasing number of rheumatic diseases, including rheumatoid arthritis (RA) [1-3], juvenile idiopathic arthritis [4], ankylosing spondylitis [5,6], and psoriatic arthritis [7]. At present, two monoclonal anti-TNF-α antibodies (infliximab and adalimumab) and one soluble p75 TNF-α receptor (etanercept) are being used in rheumatological practice.
Various skin conditions have been reported in clinical trials, including urticaria, rash, and stomatitis (during infliximab therapy) [8]; rash and injection-site reactions (during adalimumab therapy) [3,9]; and injection-site reactions (during etanercept therapy) [2].
However, clinical trials are not designed to provide information about the occurrence of rare adverse events associated with TNF-α-blocking therapy. More severe cutaneous reactions, such as erythema multiforme, discoid and subacute cutaneous lupus erythematosus, atopic dermatitis, necrotizing vasculitis, and bullous skin lesions, have been reported, mostly as single-case observations [10-15]. Larger observational studies such as biological registries are needed to provide information on the nature and number of such dermatological adverse events during TNF-α-blocking therapy.
The aim of this study was to investigate whether dermatological conditions after TNF-α-blocking therapy are a significant and clinically important problem in RA patients receiving TNF-α-blocking therapy.
Materials and methods
Study design
In a prospective cohort study, all consecutive patients with a diagnosis of RA according to the criteria of the American Rheumatism Association [16] who were starting on TNF-α-blocking therapy at the Department Of Rheumatology of the Radboud University Nijmegen Medical Centre were followed as part of a Biological Registry [17]. Approval was obtained by the hospital's ethics committee.
Patients were required to meet the criteria set out in the Dutch guidelines for biological therapies: a moderate to high disease activity score (DAS) based on 28 joints (DAS28 ≥ 3.2), and failure or intolerability of at least two disease-modifying antirheumatic drugs (DMARDs), including methotrexate, in adequate dosage regimens. Besides therapy with registrated TNF-α-blocking agents – infliximab, etanercept, and adalimumab – some patients were treated in clinical trials with lenercept, a soluble p55 TNF-α-receptor [18].
The number and nature of dermatological conditions that led patients in this cohort to consult a dermatologist during follow-up were investigated. The RA patients treated with TNA-α-blocking agents who experienced dermatological events was compared with a control group of patients who had RA but had never had TNF-α-blocking therapy. The control patients were selected from the Nijmegen inception cohort, in which 500 RA patients have been followed since 1985 [19]. Each control was paired with a TNF-α-treated patient for duration and season of the follow-up period, within a 2-month window.
Variables
Data collected at the start of TNF-α-blocking therapy were age, sex, duration of disease, presence or absence of rheumatoid factor (measured by ELISA; considered positive if results showed >10 IU/ml), antinuclear antibody (tested for by immunofluorescence on Hep-2 cells), number of DMARDs previously used, and start date of TNF-α-blocking therapy. Baseline information obtained included erythrocyte sedimentation rate (ESR), 28-joint counts for swelling and tenderness, and general wellbeing as indicated on a visual analogue scale, and the disease activity score (DAS28) was calculated [20].
Variables about which information was collected during TNF-α-blocking therapy were the use of concomitant DMARDs and prednisolone, dose and interval changes of TNF-α-blocking agents and, if appropriate, date and reason for discontinuation.
All patients who visited a dermatologist during follow-up were identified. Clinically important dermatological events were defined as any new manifestation or any exacerbation of pre-existing skin disease during follow-up. A standardized chart review form was used to record the following: start date of event, dermatological history, medication, morphological description, localization, histopathological and immunohistological information if available, working diagnosis, additional investigations, topical and systemic therapeutic actions, outcome of event, and any available information on rechallenge.
Drug-related eruptions were defined as skin reactions with a probable or definite relation to the use of TNF-α-blocking agents, based on a time relation with the administration of the agent, morphological pattern, and/or histological information. Drug-related eruptions were classified morphologically according to the criteria of Fitzpatrick and colleagues [21]. Events were also classified as major or minor, major events being any requiring hospitalization.
Patient-years of follow-up were calculated for total follow-up, time on active therapy, and time after discontinuation of therapy (time off therapy). The number of events per year of follow-up was calculated for each RA patient for total time of follow-up, time on active treatment, and time off treatment, if appropriate.
In the control group, the following baseline characteristics were collected: age, sex, disease duration, rheumatoid factor, antinuclear antibody, DAS28, the number of DMARDs previously used, and prednisolone use. All visits to a dermatologist during follow-up were identified. Events were not recorded in the control group.
Statistical analyses
The baseline characteristics of RA patients on TNF-α-blocking therapy were compared according to whether or not the patients experienced dermatological events. The chi-square test was applied for dichotomous variables and Student's t-test was used for continuous variables. Nonparametric tests were applied when appropriate. The Wilcoxon signed rank test was used to compare the number of events per patient-year of follow-up in patients receiving and patients not receiving active TNF-α-blocking therapy. Univariate and multivariate logistic regression analyses were performed to identify possible predictive factors for the occurrence of a dermatological visit (independent variable, dichotomous) in RA patients on TNF-α-blocking therapy. Dependent variables tested were sex, age at diagnosis, rheumatoid factor, antinuclear antibody, disease duration, DAS28 at baseline, prior number of DMARDs, use of prednisolone, and duration of follow-up. Odds ratios (ORs) and 95% confidence intervals (95% CIs) were calculated.
The number of patients who visited a dermatologist was compared between RA patients on TNF-α-blocking therapy and controls, using the chi-square test. P values and ORs were calculated.
All tests were two-sided, with P < 0.05 considered statistically significant. Statistical analyses were performed using SPSS statistical software (v 12.0.1, SPSS Inc, USA).
Results
Patients
A total of 289 RA patients started TNF-α-blocking therapy between June 1994 and December 2003. Their baseline characteristics are shown in Table 1.
The median follow-up time was 2.3 years (range 0.02 to 9.6). The total follow-up time was 911 patient-years, with 627 patient-years representing active therapy. Seventy of the 289 RA patients (24%) received more than one TNF-α-blocking agent and 8 (3%) received more than two agents. Infliximab was administered to 167 patients, adalimumab to 108, etanercept to 78, and lenercept to 31.
Dermatological events were recorded in 72 of the 289 RA patients (25%) receiving TNF-α-blocking therapy and in 37 (13%) of the control group (n = 289). The odds ratio (OR) of TNF-α-blocking therapy for a dermatological referral was 2.26 (95%CI 1.46 to 3.50, P < 0.0005). In patients on TNF-α-blocking therapy fifty-six instances of dermatological conditions were recorded in 34 patients (47%) and included, among others, 10 drug reactions – while the patient was receiving gold (7), nonsteroidal anti-inflammatory drugs (2), or methotrexate (1) – 10 cases of eczema, 9 of mycosis, 3 of other infections, and 5 of chronic venous insufficiency.
Predictive factors
In univariate analyses, duration of follow-up (OR 1.27, 95%CI 1.14 to 1.41, P < 0.0005) and of disease (OR 1.03, 95%CI 1.003 to 1.07, P < 0.05) were statistically significant predictive factors for a dermatological event. In a multivariate model, only duration of follow-up was a statistically significant predictive factor (OR 1.30, 95%CI 1.12 to 1.52, P < 0.001).
Dermatological events
One hundred and twenty-eight dermatological events were recorded during follow-up in RA patients on TNF-α-blocking therapy (0.14 event per patient-year), as listed in Table 2. The event per patient-year ratio was 0.16 during active treatment and 0.10 off treatment (P < 0.001). The number of events recorded during or after treatment was 56 for adalimumab (0.12 event per patient-year), 49 for infliximab (0.14 per patient-year), 16 for etanercept (0.13 per patient-year), and 13 for lenercept (0.07 per patient-year). TNF-α-blocking therapy was permanently withdrawn because of dermatological events 21 times in 19 patients.
Infections
Thirty-three infections were recorded in 27 patients, consisting of 20 fungal, 11 bacterial, and 2 viral infections (see Table 3). Two patients had had a previous episode of dermatomycosis. None of the patients required hospitalization. One patient, who temporarily discontinued adalimumab monotherapy twice because of elective surgery, developed a bacterial superinfection of pre-existing eczema after every restart.
Eczema
Eczema was diagnosed 20 times in 19 patients and appeared in various morphological patterns. Most events were described as erythematosquamous (n = 8) or erythematous (n = 3) lesions or plaques, localized on hands and feet (n = 3), arms and legs (n = 5), face (n = 1), neck (n = 1), and buttocks (n = 1). A vesicular rash on hands and feet was described five times. A papular rash was described in three cases, with localization around the eyes, on the back, and once on the back and lower legs. Diagnoses comprised dyshidrotic (n = 5), contact (n = 4), nummular (n = 1), atopic (n = 1), papular (n = 1), and nonspecific eczema (n = 8). Two patients had a prior history of dyshidrotic eczema.
Biopsies were performed in five events. Histology showed dermatitis and spongiosis in all cases, with high dermal perivascular infiltration in three. One biopsy also showed mild psoriasiform acanthosis and another showed additional keratinocyte necrosis.
Three patients stopped TNF-α-blocking therapy because of the dermatological event, after which the lesions resolved. Hospitalization was necessary for treatment of eczema in one patient. In another patient the eczematous lesions recurred after adalimumab therapy was restarted. Adalimumab was continued and topical steroids were applied with good effect. TNF-α-blocking therapy had already been stopped in 4 patients before the onset of eczema and was continued in 13 patients, of whom 7 had persisting or recurring lesions. Therapy consisted mostly of topical corticosteroids.
Drug-related eruptions
Drug-related eruptions occurred frequently during the first 5 months of TNF-α-blocking therapy and were caused by all four TNF-α-blocking agents (see Table 4). In two cases, a generalized drug-related eruption followed subcutaneous injection of etanercept. In two cases, the eruption developed during infusion (patients numbers 8 and 11, Table 4). In the other cases the time of onset ranged between 2 and 57 days after the most recent infusion.
Most drug-related eruptions consisted of a combination of morphological patterns, including exanthema, urticarial eruptions, lichenoid skin lesions, and purpura. In four patients, an eczematous drug-related eruption was seen. Classification as drug-related eruption was based on a time relation with administration of the TNF-α-blocking agent, the morphological pattern, and/or histological information. Two patients had experienced a previous drug-induced eruption (1 dermatitis in response to gold, 1 dermatitis after indomethacin).
The histological findings were compatible with the diagnosis in all cases. Perivascular infiltrations – predominantly lymphocytic – epidermal exocytosis, and hyperorthokeratosis were described. Interface dermatitis was described in three instances. One biopsy revealed focal infiltrations with marked vascular and endothelial proliferation.
Seven patients stopped and 8 patients continued therapy; 6 of them had a positive rechallenge and recurring lesions. One major event was recorded: an RA patient was hospitalized for an extensive eczematous eruption with urticaria on arms and legs (Fig. 1, and Patient no. 6 in Table 4). Treatment consisted mostly of topical application of corticosteroids and sometimes of systemic antihistamines.
Tumors and actinic keratosis
Events of skin malignancies were recorded five times, in four patients. One RA patient developed three basal cell carcinomas simultaneously on her left arm, right nostril, and right eyelid after 2.7 years of adalimumab therapy, which was subsequently stopped. One 74-year-old RA patient developed Bowen's disease on his right hand 2 years after adalimumab therapy had been stopped. The same patient later developed a squamous cell carcinoma on the left earlobe after the start of etanercept therapy. Other skin malignancies recorded were a squamous cell carcinoma (earlobe) after 1.5 months of adalimumab therapy and a low-grade basalioma (Pinkus epithelioma) on the leg after 6 months of adalimumab therapy. In all cases, histology confirmed the diagnosis and therapy consisted of excision. No recurrences were seen.
Actinic keratosis was recorded in five patients (three receiving adalimumab, one infliximab, and one lenercept). Excision or cryotherapy was successful in four. One patient had recurring actinic lesions on the scalp.
Benign tumors were recorded seven times during TNF-α-blocking therapy. One patient experienced an increased growth of a facial telangiectatic nevus, present since childhood, 2 months after starting etanercept therapy. Seborrheic keratosis (n = 3), oral hyperkeratosis (n = 1), histiocytoma (n = 1), and fibroma (n = 1) were also recorded.
Vasculitis
Vasculitis was recorded five times: four during and one after cessation of TNF-α-blocking therapy. The diagnosis was confirmed by biopsy in four cases. One patient developed a superficial necrotizing leukocytoclastic vasculitis with ulceration after 7 months of infliximab therapy, with complete recovery after discontinuation of infliximab. One patient developed a papular erythema in the groins after 5 years of adalimumab therapy. Histological examination was compatible with vasculitis with infiltration of mononuclear cells and presence of eosinophilic granulocytes. One patient developed a purpuric vasculitis on the legs after 1.5 months of lenercept therapy, improving spontaneously despite continuation of lenercept. One patient developed isolated digital vasculitis on his toes after one year of adalimumab therapy, which was continued. The lesions persisted. No biopsy was performed. One patient developed a generalized urticarial exanthema after therapy with etanercept 2 years earlier. Current therapy consisted of hydroxychloroquine and prednisolone. Histology showed a mild leukocytoclastic vasculitis.
Ulcers
The nine events with ulcers included four pressure ulcers, two ulcers due to dependency edema, one traumatic ulcer, one ulcer secondary to an unguis incarnatus, and one ulcer without further specification. Biopsies were taken in two patients, but no signs of vasculitis were found. A patient had a pressure ulcer with secondary infection and a fistula on his ankle, which contained osteosynthetic material. The patient was admitted to the hospital for intravenous antibiotic therapy and infliximab was stopped for several months. After recovery, the patient restarted infliximab without recurrence of his skin problems. TNF-α-blocking therapy was continued in the other eight patients, and in four of these the ulcers recovered; follow-up was missing in the other four.
Stasis dermatitis, edema, varices and chronic venous insufficiency
In 10 patients, a dermatological consultation was recorded for stasis dermatitis (n = 3), edema (n = 3), varices (n = 2), or chronic venous insufficiency (n = 2). In one patient with extensive varices, infliximab therapy was stopped temporarily because of a complicating thrombophlebitis. One patient had edema of both legs of unknown cause, with livid discoloration and induration. One patient had lymphedema secondary to RA. All other events were considered to be related to comorbidity, other than RA.
Psoriasis and psoriasiform eruptions
Psoriatic or psoriasiform eruptions were recorded in three RA patients. One developed a vesiculopustular erythematosquamous rash on hands and feet after 9 months of adalimumab therapy. Histology showed a mixed psoriasiform and spongiotic dermatitis. A second RA patient developed psoriasis guttata-like eruptions on her lower legs after 4 years of therapy with adalimumab. The lesions diminished after adalimumab was withdrawn. A third patient developed a psoriasiform eruption on arms and legs after 16 months of adalimumab therapy. Histology obtained in the latter two patients was consistent with psoriasis.
Other dermatological conditions
Other dermatological conditions that occurred during or after TNF-α-blocking therapy included, among others, dermatomyositis (1), drug-induced systemic lupus erythematosus (1), and lymphomatoid papulosis-like eruption (1). Details are shown in Table 5.
One RA patient developed a macular rash on the inner sides of the upper arms and legs after 2.5 months of lenercept monotherapy. A skin biopsy showed a nonspecific chronic dermatitis. A soft-tissue biopsy, including skin, fascia, and muscle, showed fascial and muscular infiltration, consistent with dermatomyositis.
One RA patient developed a drug-induced systemic lupus erythematosus after 20 months of infliximab therapy in combination with methotrexate, consisting of discoid lupus erythematosus lesions on her hands and scalp, aphthous lesions, conversion to antinuclear antibody positivity, and a positive anti-double stranded-DNA (titer 60 U/L). The skin lesions flared within one week after infusion and disappeared after discontinuation of infliximab.
A third RA patient developed macular erythematosquamous lesions on her lower arms, upper legs and trunk after 2.6 months of adalimumab monotherapy. Histology showed a dermal infiltration with CD30-positive atypical T cells. Although the lesions appeared to be lymphomatoid papulosis, they completely disappeared within 6 weeks. Adalimumab was not stopped. This patient developed a large-cell anaplastic non-Hodgkin lymphoma 2 years later.
Discussion
The present study is the first large prospective study focusing on dermatological conditions in RA patients on TNF-α-blocking therapy. Of the patients studied, 25% needed a dermatological consultation, compared with 13% in a RA control group, naive to TNF-α-blocking therapy. The number of dermatological events per patient-year was significantly higher during treatment than after treatment with TNF-α-blocking therapy. Dermatological events led to withdrawal of TNF-α-blocking therapy in 19 patients of 72 patients (26%). The events recorded most frequently were skin infections, eczema, and drug-related eruptions. Some other interesting events were recorded, such as psoriasis, drug-induced systemic lupus erythematosus, dermatomyositis, and a lymphomatoid-papulosis like eruption.
RA is known to be associated with dermatological conditions such as vasculitis, nodulosis, palmar erythema, and bullous pemphigoid, among others [22,23]. At present, information on the incidence and prevalence of dermatological conditions in RA mainly originates from cross-sectional or retrospective studies [24-26]. Few prospective studies have been conducted focusing on specific conditions affecting the skin [27,28].
In establishing a relation between the use of a drug and the occurrence of dermatological conditions, various factors must be considered. Information on clinical and histological patterns, time and dose relation, dechallenge and rechallenge, and analogy with previously reported cases can provide support in assessing the plausibility of such a relation [29]. The underlying disease and concomitant medication also need careful consideration, as they can provide alternative explanations.
In this study the largest group of dermatological events consisted of skin infections, mostly fungal infections and folliculitis. The use of TNF-α-blocking therapy has raised concerns regarding an increased susceptibility to infections, as TNF-α plays an important role in host-defence mechanisms [30]. An increased incidence of tuberculosis has been described [31], as well as a growing number of serious infections with fungal, mycobacterial, and intracellular bacterial pathogens [32-34]. Infections of the skin have not been the subject of report in clinical trials and observational studies with TNF-α-blocking therapy. Cases of severe necrotizing fasciitis have been described [35,36].
Skin infections have been reported frequently in the normal population and especially in RA patients [24-26]. Host-defence impairments resulting from the underlying disease might play a role in an increased susceptibility to skin infections in RA patients, as well as the use of corticosteroids and DMARDs such as methotrexate [28,37], which were recorded frequently in the present study (see Table 2). They could provide an alternative explanation for the occurrence of skin infections. However, most infections occurred during active treatment with TNF-α-blocking therapy, a finding that could suggest at least a relative contribution to an increased vulnerability to skin infections in the study population. In one patient, a bacterial superinfection of eczema occurred twice immediately after restart of adalimumab, showing a clear time relation.
For the description of the recorded drug-related eruptions, a clinico-morphological classification was chosen [21]. Four eruptions with a time relation and clinically or histological distinct drug-induced patterns also showed an eczematous appearance, both clinically and histologically. This is an unusual presentation for a drug-induced eruption and warrants further investigation.
Two drug-related eruptions occurred during infusion with infliximab or adalimumab, whereas all the others occurred after infusion. This will most likely not reflect the true ratio between acute and delayed reactions involving the skin, since acute reactions with skin involvement occur in 4% of the infusions and are usually treated by the rheumatologist without dermatological consultation [38].
Eczema was reported frequently in this study, even with various dermatitis conditions, such as xerosis cutis, stasis eczema, and seborrheic eczema, classified as separate entities. Previous studies have reported RA, in which Th1 (T helper cell type 1) immune responses dominate, to be negatively associated with Th2-cell-mediated atopic disorders, such as eczema [39-41], although a similar incidence of eczema in RA and non-RA patients has also been reported [42]. TNF-α-blocking therapy down-regulates Th1 immune responses [43], which might induce a shift of the Th1/Th2 balance towards Th2-dominated immune responses and which might promote an increased susceptibility to atopic disorders, such as eczema.
Although the time between the initiation of TNF-α-blocking therapy and the onset of dermatological conditions varied, a probable relation was seen in various events. These included, besides drug-related eruptions, events of cutaneous vasculitis, drug-induced systemic lupus erythematosus, dermatomyositis, and a lymphomatoid papulosis-like eruption.
An association between the use of TNF-α-blocking therapy and the induction of systemic lupus erythematosus and discoid lupus erythematosus is strongly suggested by the number of cases that have been published [10,11,13,44-46]. One case of discoid lupus erythematosus has been described on both etanercept and infliximab in the same RA patient [47].
Analogy with previous reports is also present for cutaneous vasculitis [13,47-49], although it is a known extra-articular manifestation of RA [22,23]. In the first case described, a probable relation with infliximab was present, based on the time relation and positive dechallenge. The other cases described were considered possibly related (Results section, Vaculitis, cases 2 and 3) and unlikely (cases 4 and 5). Almost all reported ulcers were considered secondary to other causes, as described.
Dermatomyositis has been reported previously, although the patient affected in that case had a different presentation, with raised creatinine phosphokinase, muscle atrophy, mechanic's hands, and vasculitis [17].
Another interesting finding was the occurrence of psoriasiform eruptions in three patients on TNF-α-blocking therapy. This observation is particularly interesting, since etanercept has received and infliximab is close to receiving FDA approval for treatment of psoriasis, after remarkable efficacy results in clinical trials [7,50,51]. The occurrence of guttate psoriasis has been reported after initiation of etanercept therapy for psoriasis in a placebo-controlled trial [51]. Another case report described the occurrence of psoriasiform eruptions with histologically a lichenoid dermatitis pattern in a patient with Crohn's disease [52].
An exacerbation of psoriasis was also seen in a patient with psoriatic arthritis receiving infliximab therapy. An additional analysis showed that 28 patients with various non-RA rheumatic diseases, including 12 juvenile idiopathic arthritis, 6 psoriatic arthritis, and 3 ankylosing spondylitis, had been treated with TNF-α-blocking therapy in the study centre. Five patients (18%) had visited a dermatologist for a dermatological condition during or after TNF-α-blocking therapy. The events included a drug-related eruption, eczema, and a facial mycosis in three patients with juvenile idiopathic arthritis and a superficial spreading melanoma in a patient with ankylosing spondylitis. This indicates that the occurrence of dermatological events during TNF-α-blocking therapy is not restricted to RA patients.
In the present study the control patients were matched for sartdate and duration of follow-up period in order to control for time-related effects. A statistically significant relation between the use of TNF-α-blocking therapy and the occurrence of dermatological visits was shown. The two groups studied differed for most baseline characteristics. These differences result from the indication for TNF-α-blocking agents, which were reserved for patients who fulfilled criteria for active disease and DMARD failure (see methods section; study design), had a longer disease duration, and whose disease was perhaps more refractory.
However, it is considered unlikely that these factors influenced the relation between the use of TNF-α-blocking therapy and dermatological visits. In a multivariate regression model, no baseline characteristic showed a predictive value for the occurrence of a dermatological event in RA patients on TNF-α-blocking therapy. Also, a statistically significantly higher number of dermatological events was recorded during active treatment with TNF-α-blocking therapy than after the therapy had been stopped.
Conclusion
This is the first prospective study showing a relation between TNF-α-blocking therapy and the occurrence of dermatological conditions. Future prospective studies are needed to investigate the incidence and the pathogenesis of the encountered events, because they are a clinically significant problem in RA patients receiving TNF-α-blocking therapy.
Abbreviations
CI = confidence interval; DAS28 = disease activity score including 28-joint counts; DMARD = disease-modifying antirheumatic drug; ELISA = enzyme-linked immunosorbent assay; pt-yr = patient-year; RA = rheumatoid arthritis; Th1/Th2 = T helper cell type 1/2; TNF = tumor necrosis factor.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MF participated in the study design, carried out the data collection and statistical analysis, and drafted the manuscript. WV participated in the study design, carried out the data collection, and helped to write the manuscript. MC participated in the study design and coordination and helped in the writing and revision of manuscript. EdJ participated in the study design and the data collection and helped to write the manuscript. PvdK and PvR helped to write and critically revise the manuscript and gave final approval of the manuscript. MF and WV contributed equally to the article. All authors read and approved the final manuscript.
Figures and Tables
Figure 1 Eczematous drug-related eruption a patient with rheumatoid arthritis after infliximab therapy: Eczematous eruptions on the left arm (top left) and right arm (top right) and erythematous eruptions with purpura on the left leg (bottom left) and right leg (bottom right).
Table 1 Baseline characteristics of patients with rheumatoid arthritis (RA) studied
Given TNF-α-blocking therapy Controlsa
Characteristic All patients N = 289 Patients with dermatological events N = 72 N = 289
Male sex, no. (%) 89 (31) 20 (28) 110 (38)
Age (yr) at diagnosis, mean (SD) 44.5 (14.7) 43.4 (12.7) 54.6 (14.1)**
RF-positive, no. (%) 249 (87) 68 (94) 205 (71)*
Disease duration (yr) at baseline, median (range) 9.2 (0.1–44.9) 10.3 (0.3–44.9)† 6.2 (0.0–12.6)**
DAS28 at baseline, mean (SD) 5.9 (1.1) 6.1 (1.1) 3.6 (1.4)**
ANA-positive at baseline, no. (%)b 112 (50) 33 (49) 118 (41)
Prior DMARDs, median (range) 4 (1–10) 5 (2–8) 1 (0–6)**
Prednisolone at baseline, no. (%) 112 (39) 34 (47) 21 (7)**
aNot given TNF-α-blocking therapy. bANA at start was present in respectively 261 and 67 patients on TNF-α-blocking therapy. *P < 0.001, **P < 0.0001, compared with RA patients on TNF-α-blocking therapy; †P < 0.001 compared with RA patients on TNF-α-blocking therapy who experienced no dermatological events. ANA, antinuclear antibody; DAS28, disease activity score based on 28 joints; DMARD, disease-modifying antirheumatic drug; RF, rheumatoid factor; SD, standard deviation; TNF, tumor necrosis factor.
Table 2 Dermatological events in patients with rheumatoid arthritis (RA) given TNF-α-blocking therapy
Nature of event Events Time to event (months) Events during treatment Major events Histology DMARDsb Prednisoloneb Permanent withdrawal of anti-TNF-αc
No. (%) Mediana Range No. (%) No. (%) No. (%) No. (%) No. (%) No. (%)
Infection 33 (25.8) 9.1 1.1–61.1 24 (73) 0 5 (15) 20 (61) 21 (64) 4 (12)
Eczema 20 (15.6) 7.1 0.2–49.9 16 (80) 1 (5) 4 (20) 8 (40) 7 (35) 3 (15)
Drug-related eruption 15 (11.7) 1.9 0.1–18.8 15 (100) 1 (7) 12 (80) 6 (40) 6 (40) 7 47)
Ulcers 9 (7.0) 13.6 0.3–52.5 3 (33) 1 (11) 2 (22) 7 (78) 4 (44) 1 (11)
Skin tumor, benign 7 (5.5) 12.9 2.0–18.1 7 (100) 0 2 (29) 5 (71) 4 (57) 0
Skin tumor, malignant 5 (3.9) 4.5 1.1–38.0 4 (80) 0 5 (100) 2 (40) 2 (40) 1 (20)
Xerosis cutis 6 (4.7) 8.9 4.2–26.3 6 (100) 0 1 (16) 4 (67) 1 (17) 1 (17)
Vasculitis 5 (3.9) 12.0 1.5–49.9 4 (80) 0 4 (80) 3 (60) 5 (100) 1 (20)
Actinic keratosis 5 (3.9) 26.3 4.5–112.9 2 (40) 0 3 (60) 5 (100) 2 (40) 0
CVI/varices 4 (3.0) 24.0 1.7–33.6 3 (75) 0 0 3 (75) 2 (50) 0
Psoriasis/psoriasiform 3 (2.3) 15.5 8.4–50.1 3 (100) 0 3 (100) 0 2 (67) 1 (33)
Edema 3 (2.2) 8.2 4.0–39.6 2 (67) 0 1 (33) 1 (33) 1 (33) 0
Stasis dermatitis 3 (2.2) 17.5 14.6–42.1 3 (100) 0 1 (33) 1 (33) 1 (33) 0
Seborrheic dermatitis 2 (1.5) 0.4, 19.8 – 2 (100) 0 0 0 0 0
Other event 8 (6.0) 5.0 1.9–25.9 6 (75) 0 4 (50) 4 (50) 2 (25) 2 (25)
Total 128 (100) 9.1 0.1–112.9 100 (78) 3 (2) 47 (37) 69 (54) 60 (47) 21 (16)
aMedian and range given for three cases or more; individual data given for two cases or fewer. bNumber of patients with concomitant DMARDs and prednisolone at the time of event. cPermanent discontinuation of TNF-α-blocking therapy because of the event. DMARD, disease-modifying anti-rheumatic drug; TNF, tumor necrosis factor.
Table 3 Skin infections in patients with rheumatoid arthritis (RA) given TNF-α-blocking therapy
Time to event
Infection No. of events Median Range Druga (no.) Active treatmentb (no.) Rechallenge (no.) Permanent withdrawal of anti-TNF-αc (no.) Biopsy (no.) Cultured species
Fungal
20 8.7 1.1–61.1
Dermatomycosis 9 A 3, I 4, E 2 7 0 1 Trichophyton verrucosum (1) T. rubrum (1)
Onychomycosis 3 A 3 3 0 0
Combination 5 A 3, I 1, L 1 4 0 1 Trichophyton rubrum (3) T. mentagrofytes (1)
Candidiasis 3 I 3 2 0 0 Candida spp. (2)
Bacterial
11 9.5 1.4–52.5
Folliculitis 5 A 3, E 2 4 yes, negative 1 2 Staphylococcus aureus (1)
Erysipelas 3 E 2, I 1 3 yes, negative 2 1
Bacterial superinfection of eczema 2 A 1, I 1 1 yes, positive 1 0
Furuncle 1 I 1 0 0 0
Viral – herpes zoster
2 17.3, 40.9d A 1, I 1 0 0 0
aA, adalimumab; I, infliximab; E, etanercept; L, lenercept. bDuring active treatment with TNF-α-blocking therapy. cPermanent discontinuation of TNF-α-blocking therapy due to the event.d Individual values
Table 4 Drug-related skin eruptions in patients with rheumatoid arthritis (RA) given TNF-α-blocking therapy
Patient no. Age (yr) Sex Druga Route Type of eruption Clinical description Localization Time to event (mo) Biopsy Comedicationb Therapy Rechallenge Permanent withdrawal of anti-TNF-α Course
1 62 f A i.v. Eczematous Erythematosquamous plaques and papules Neck/ axillary/ legs 4.5 Yes naproxen Local positive No Recurring
2 71 m A i.v. Exanthematous lichenoid Maculopapular exanthema Generalized 0.7 Yes prednisolone, naproxen, paracetamol Local positive Yes Recovery
3 77 m E s.c. Exanthematous Macular exanthema Generalized 6.8 Yes prednisolone, naproxen, omeprazole Local positive No Recurring
4 67 m E s.c. Lichenoid Macular exanthema, purpura Generalized 1.5 Yes diclofenac, omeprazole, triamterene, furosemide, candesartan Topical, systemic No Yes Recovery
5 69 f I i.v. Eczematous Erythematous plaque Right cheek 0.1 Yes MTX, pantoprazole, atenolol, calcium, hydrochlorothiazide Topical positive No Recurring
6 88 f I i.v. Eczematous urticarial Erythematosquamous macula, purpura Lower arms/legs 3.9 Yes leflunomide, carbasalate calcium, omeprazole, furosemide, simvastatin, paracetamol Topical No Yes Recovery
7 68 f I i.v. Eczematous urticarial Erythematosquamous plaques, urticaria, excoriations, lichenification, purpura Generalized 10.3 No AZA, furosemide, oxazepam, enalapril, spironolactone, metoprolol, flixotide, formoterol Topical, systemic negative No Recovery
8 60 f I i.v. Exanthematous Stippled exanthema Generalized 0.5 Yes naproxen, omeprazole Topical No Yes Recovery
9 53 f I i.v. Exanthematous Exanthema Upper arms/legs 0.2 No indomethacin Topical positive No Recurring
10 73 f I i.v. Exanthematous, with purpura Exanthema, purpura Lower legs 18.8 No MTX, folic acid, prednisolone, morphine, loperamide, latanoprost Topical No Yes Recovery
11 70 f I i.v. Exanthematous urticarial Exanthema, urticaria Arms/ trunk 16.6 Yes leflunomide None positive No Recurring
12 35 f I i.v. Exanthematous urticarial, with purpura Macular exanthema, uricaria, purpura Trunk/ axillary/ groins 1.9 Yes none Topical - Yes Recovery
13 58 f I i.v. Lichenoid Erythema, hyperpigmentation, atrophy Upper legs 15.5 Yes leflunomide, meloxicam, metoclopramide, acenocoumarol, digoxin None No Yes Recovery
14 58 f L i.v. Exanthematous Papular exanthema Generalized 0.4 Yes none Topical positive No Recurring
15 68 m L i.v. Exanthematous lichenoid Maculopapular exanthema Generalized 1.7 Yes prednisolone, paracetamol Topical negative No Recovery
Events numbers 5 and 11 occurred in the same patient, as did events numbers 2, 3, and 15. aA, adalimumab; Age = age ar event; I, infliximab; E, etanercept; L, lenercept. bMTX, methotrexate; AZA, azathioprine. f, female; i.v., intravenous; m, male; s.c., subcutaneous.
Table 5 Other dermatological events in patients with rheumatoid arthritis (RA) given TNF-α-blocking therapy
Patient no. Age (yr) Sex Diagnosis Druga Active treatment Event Clinical description Localization Time to event Biopsy Comedicationb Permanent withdrawal anti-TNF-α Therapy Course
1 56 f RA A Yes Lymphomatoid papulosis-like eruption Macular erythematosquamous lesions Lower arms, upper legs and trunk 2.6 Yes naproxen No None Recovery
2 53 f RA A Yes Rosacea Diffuse erythema, scaling, telangiectasias Head and face 1.9 Yes prednisolone, captopril, indomethacin, midazolam No Topical Persisting
3 74 f RA E Yes Pruritus Itch Trunk 2.5 No None No Topical Unknown
4 61 f RA I No Ecchymoses Ecchymoses Hands and feet 25.9 No AZA, prednisolone No Topical Partial recovery
5 58 f RA I Yes Drug-induced systemic lupus emythematosus Discoid erythematous lesions, aphthous lesions, ANA positive, anti-ds-DNA positive Hands, face, scalp 20.0 No MTX Yes Topical and systemic Recovery, no rechallenge
6 68 m RA I Yes Transient swelling of unknown cause Transient swelling 2 × 3 cm Scalp 20.0 No MTX, folic acid, naproxen No None Recovery
7 52 f RA L Yes Dermatomyositis Livid erythema, raised CPK, decreased proximal muscular strength Inner upper arms and legs 2.5 Yes None Yes None Recovery
8 53 m RA L No Erythema nodosum Painful erythematous nodules Lower legs 7.4 Yes AZA, naproxen, paracetamol No Topical Partial recovery
aA, adalimumab; I, infliximab; E, etanercept; L, lenercept. bMTX, methotrexate; AZA, azathioprine. CPK, creatinine phosphokinase; f, female, m, male.
==== Refs
Lipsky PE van der Heijde DM St Clair EW Furst DE Breedveld FC Kalden JR Smolen JS Weisman M Emery P Feldmann M Infliximab and methotrexate in the treatment of rheumatoid arthritis. Anti-Tumor Necrosis Factor Trial in Rheumatoid Arthritis with Concomitant Therapy Study Group N Engl J Med 2000 343 1594 1602 11096166 10.1056/NEJM200011303432202
Moreland LW Schiff MH Baumgartner SW Tindall EA Fleischmann RM Bulpitt KJ Weaver AL Keystone EC Furst DE Mease PJ Etanercept therapy in rheumatoid arthritis. A randomized, controlled trial Ann Intern Med 1999 130 478 486 10075615
Weinblatt ME Keystone EC Furst DE Moreland LW Weisman MH Birbara CA Teoh LA Fischkoff SA Chartash EK Adalimumab, a fully human anti-tumor necrosis factor alpha monoclonal antibody, for the treatment of rheumatoid arthritis in patients taking concomitant methotrexate: the ARMADA trial Arthritis Rheum 2003 48 35 45 12528101 10.1002/art.10697
Lovell DJ Giannini EH Reiff A Jones OY Schneider R Olson JC Stein LD Gedalia A Ilowite NT Wallace CA Long-term efficacy and safety of etanercept in children with polyarticular-course juvenile rheumatoid arthritis: interim results from an ongoing multicenter, open-label, extended-treatment trial Arthritis Rheum 2003 48 218 226 12528122 10.1002/art.10710
Brandt J Khariouzov A Listing J Haibel H Sorensen H Grassnickel L Rudwaleit M Sieper J Braun J Six-month results of a double-blind, placebo-controlled trial of etanercept treatment in patients with active ankylosing spondylitis Arthritis Rheum 2003 48 1667 1675 12794835 10.1002/art.11017
Braun J Brandt J Listing J Zink A Alten R Golder W Gromnica-Ihle E Kellner H Krause A Schneider M Treatment of active ankylosing spondylitis with infliximab: a randomised controlled multicentre trial Lancet 2002 359 1187 1193 11955536 10.1016/S0140-6736(02)08215-6
Mease PJ Goffe BS Metz J VanderStoep A Finck B Burge DJ Etanercept in the treatment of psoriatic arthritis and psoriasis: a randomised trial Lancet 2000 356 385 390 10972371 10.1016/S0140-6736(00)02530-7
Maini R St Clair EW Breedveld F Furst D Kalden J Weisman M Smolen J Emery P Harriman G Feldmann M Infliximab (chimeric anti-tumour necrosis factor alpha monoclonal antibody) versus placebo in rheumatoid arthritis patients receiving concomitant methotrexate: a randomised phase III trial. ATTRACT Study Group Lancet 1999 354 1932 1939 10622295 10.1016/S0140-6736(99)05246-0
Keystone EC Kavanaugh AF Sharp JT Tannenbaum H Hua Y Teoh LS Fischkoff SA Chartash EK Radiographic, clinical, and functional outcomes of treatment with adalimumab (a human anti-tumor necrosis factor monoclonal antibody) in patients with active rheumatoid arthritis receiving concomitant methotrexate therapy: a randomized, placebo-controlled, 52-week trial Arthritis Rheum 2004 50 1400 1411 15146409 10.1002/art.20217
Bleumink GS ter Borg EJ Ramselaar CG Stricker BHC Etanercept-induced subacute cutaneous lupus erythematosus Rheumatology 2001 40 1317 1319 11709621 10.1093/rheumatology/40.11.1317
Brion PH Mittal HA Kalunian KC Autoimmune skin rashes associated with etanercept for rheumatoid arthritis [letter] Ann Intern Med 1999 131 634 10523238
Kent PD Davis JM Davis MDP Matteson EL Bullous skin lesions following infliximab infusion in a patient with rheumatoid arthritis Arthritis Rheum 2002 46 2257 2258 12209539 10.1002/art.10348
Misery L Perrot JL Gentil PA Pallot PB Cambazard F Alexandre C Dermatological complications of etanercept therapy for rheumatoid arthritis Br J Dermatol 2002 146 334 335 11903255 10.1046/j.1365-2133.2002.4653_4.x
Vergara G Silvestre JF Betlloch I Vela P Albares MP Pascual JC Cutaneous drug eruption to infliximab: Report of 4 cases with an interface dermatitis pattern Arch Dermatol 2002 138 1258 1259 12225003 10.1001/archderm.138.9.1258
Wright RC Atopic dermatitis-like eruption precipitated by infliximab JAm Acad Dermatol 2003 49 160 161 12833036
Arnett FC Edworthy SM Bloch DA McShane DJ Fries JF Cooper NS Healey LA Kaplan SR Liang MH Luthra HS The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis Arthritis Rheum 1988 31 315 324 3358796
Flendrie M Creemers MC Welsing PM den Broeder AA van Riel PL Survival during treatment with tumour necrosis factor blocking agents in rheumatoid arthritis Ann Rheum Dis 2003 62 Suppl 2 ii30 ii33 14532145
Rau R Sander O van Riel PL van de Putte LB Hasler F Zaug M Kneer J van der AP Stevens RM Intravenous human recombinant tumor necrosis factor receptor p55-Fc IgG1 fusion protein Ro 45-2081 (lenercept): a double blind, placebo controlled dose-finding study in rheumatoid arthritis J Rheumatol 2003 30 680 690 12672184
Welsing PMJ van Riel PLCM The Nijmegen inception cohort of early rheumatoid arthritis J Rheumatol 2004 31 14 21 14705212
Prevoo ML van 't Hof MA Kuper HH van Leeuwen MA van de Putte LB van Riel PL Modified disease activity scores that include twenty-eight-joint counts. Development and validation in a prospective longitudinal study of patients with rheumatoid arthritis Arthritis Rheum 1995 38 44 48 7818570
Stern R Wintroub BU Freedberg IM, Eisen A, Wolff K, Austen KF, Goldsmith LA, Katz SI Cutaneous reactions to drugs Fitzpatrick's Dermatology in General Medicine 1999 5 New York: McGraw-Hill 1634 1642
Jorizzo JL Daniels JC Dermatologic conditions reported in patients with rheumatoid arthritis J Am Acad Dermatol 1983 8 439 457 6343438
Sibbitt WL Williams RC Cutaneous manifestations of rheumatoid arthritis Int J Dermatol 1982 21 563 572 6761281
Yamamoto T Ohkubo H Nishioka K Skin manifestations associated with rheumatoid arthritis J Dermatol 1995 22 324 329 7673550
Bicer A Tursen U Cimen OB Kaya TI Ozisik S Ikizoglu G Erdogan C Prevalence of dermatophytosis in patients with rheumatoid arthritis Rheumatol Int 2003 23 37 40 12548440
Doran MF Crowson CS Pond GR O'Fallon WM Gabriel SE Frequency of infection in patients with rheumatoid arthritis compared with controls: a population-based study Arthritis Rheum 2002 46 2287 2293 12355475 10.1002/art.10524
Wilkinson SM Smith AG Davis MJ Mattey DL Dawes PT Suspected cutaneous drug toxicity in rheumatoid arthritis – an evaluation Br J Rheumatol 1993 32 798 803 8103699
van der Veen MJ van der HA Kruize AA Bijlsma JW Infection rate and use of antibiotics in patients with rheumatoid arthritis treated with methotrexate Ann Rheum Dis 1994 53 224 228 8203949
Miller FW Hess EV Clauw DJ Hertzman PA Pincus T Silver RM Mayes MD Varga J Medsger TA Love LA Approaches for identifying and defining environmentally associated rheumatic disorders Arthritis Rheum 2000 43 243 249 10693862 10.1002/1529-0131(200002)43:2<243::AID-ANR2>3.0.CO;2-K
Bresnihan B Cunnane G Infection complications associated with the use of biologic agents Rheum Dis Clin North Am 2003 29 185 202 12635507 10.1016/S0889-857X(02)00101-1
Keane J Gershon S Wise RP Mirabile-Levens E Kasznica J Schwieterman WD Siegel JN Braun MM Tuberculosis associated with infliximab, a tumor necrosis factor alpha-neutralizing agent N Engl J Med 2001 345 1098 1104 11596589 10.1056/NEJMoa011110
Hyrich KL Silman AJ Watson KD Symmons DP Anti-tumour necrosis factor {alpha} therapy in rheumatoid arthritis: an update on safety Ann Rheum Dis 2004 63 1538 1543 15242866 10.1136/ard.2004.024737
Ellerin T Rubin RH Weinblatt ME Infections and anti-tumor necrosis factor alpha therapy Arthritis Rheum 2003 48 3013 3022 14613261 10.1002/art.11301
Netea MG Radstake T Joosten LA van der Meer JW Barrera P Kullberg BJ Salmonella septicemia in rheumatoid arthritis patients receiving anti-tumor necrosis factor therapy: association with decreased interferon-gamma production and Toll-like receptor 4 expression Arthritis Rheum 2003 48 1853 1857 12847679 10.1002/art.11151
Chan AT Cleeve V Daymond TJ Necrotising fasciitis in a patient receiving infliximab for rheumatoid arthritis Postgrad Med J 2002 78 47 48 11796874 10.1136/pmj.78.915.47
Baghai M Osmon DR Wolk DM Wold LE Haidukewych GJ Matteson EL Fatal sepsis in a patient with rheumatoid arthritis treated with etanercept Mayo Clin Proc 2001 76 653 656 11393506
Hernandez-Cruz B Cardiel MH Villa AR Alcocer-Varela J Development, recurrence, and severity of infections in Mexican patients with rheumatoid arthritis. A nested case-control study J Rheumatol 1998 25 1900 1907 9779842
Wasserman MJ Weber DA Guthrie JA Bykerk VP Lee P Keystone EC Infusion-related reactions to infliximab in patients with rheumatoid arthritis in a clinical practice setting: relationship to dose, antihistamine pretreatment, and infusion number J Rheumatol 2004 31 1912 1917 15468353
Hartung AD Bohnert A Hackstein H Ohly A Schmidt KL Bein G Th2-mediated atopic disease protection in Th1-mediated rheumatoid arthritis Clin Exp Rheumatol 2003 21 481 484 12942701
Hilliquin P Allanore Y Coste J Renoux M Kahan A Menkes CJ Reduced incidence and prevalence of atopy in rheumatoid arthritis. Results of a case-control study Rheumatology 2000 39 1020 1026 10986309 10.1093/rheumatology/39.9.1020
Rudwaleit M Andermann B Alten R Sorensen H Listing J Zink A Sieper J Braun J Atopic disorders in ankylosing spondylitis and rheumatoid arthritis Ann Rheum Dis 2002 61 968 974 12379517 10.1136/ard.61.11.968
Olsson AR Wingren G Skogh T Svernell O Ernerudh J Allergic manifestations in patients with rheumatoid arthritis APMIS 2003 111 940 944 14616545 10.1034/j.1600-0463.2003.1111004.x
Krueger JG The immunologic basis for the treatment of psoriasis with new biologic agents J Am Acad Dermatol 2002 46 1 23 11756941 10.1067/mjd.2002.120568
Carlson E Rothfield N Etanercept-induced lupus-like syndrome in a patient with rheumatoid arthritis Arthritis Rheum 2003 48 1165 1166 12687569 10.1002/art.11033
Favalli E Sinigaglia L Varenna M Arnoldi C Drug-induced lupus following treatment with infliximab in rheumatoid arthritis Lupus 2002 11 753 755 12475006 10.1191/0961203302lu236cr
Shakoor N Michalska M Harris CA Block JA Drug-induced systemic lupus erythematosus associated with etanercept therapy Lancet 2002 359 579 580 11867114 10.1016/S0140-6736(02)07714-0
Geborek P Crnkic M Petersson IF Saxne T Etanercept, infliximab, and leflunomide in established rheumatoid arthritis: clinical experience using a structured follow up programme in southern Sweden Ann Rheum Dis 2002 61 793 798 12176803 10.1136/ard.61.9.793
Jarrett SJ Cunnane G Conaghan PG Bingham SJ Buch MH Quinn MA Emery P Anti-tumor necrosis factor-alpha therapy-induced vasculitis: case series J Rheumatol 2003 30 2287 2291 14528531
McCain ME Quinet RJ Davis WE Etanercept and infliximab associated with cutaneous vasculitis Rheumatology 2002 41 116 117 11792895 10.1093/rheumatology/41.1.116
Chaudhari U Romano P Mulcahy LD Dooley LT Baker DG Gottlieb AB Efficacy and safety of infliximab monotherapy for plaque-type psoriasis: a randomised trial Lancet 2001 357 1842 1847 11410193 10.1016/S0140-6736(00)04954-0
Gottlieb AB Matheson RT Lowe N Krueger GG Kang S Goffe BS Gaspari AA Ling M Weinstein GD Nayak A A randomized trial of etanercept as monotherapy for psoriasis Arch Dermatol 2003 139 1627 1632 14676082 10.1001/archderm.139.12.1627
Verea MM Del Pozo J Yebra-Pimentel MT Porta A Fonseca E Psoriasiform eruption induced by infliximab Ann Pharmacother 2004 38 54 57 14742794 10.1345/aph.1C477
| 15899052 | PMC1174960 | CC BY | 2021-01-04 16:02:36 | no | Arthritis Res Ther. 2005 Apr 4; 7(3):R666-R676 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1724 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17251589905310.1186/ar1725Research ArticleIntra-articular injections of high-molecular-weight hyaluronic acid have biphasic effects on joint inflammation and destruction in rat antigen-induced arthritis Roth Andreas [email protected] Jürgen 12Wagner Andreas 1Fuhrmann Reneè [email protected] Albrecht 1Venbrocks Rudolf A 1Petrow Peter [email protected]äuer Rolf [email protected] Harald 4Ozegowski Jörg [email protected] Gundela 6Müller Peter J 6Kinne Raimund W [email protected] Department of Orthopaedics, 'Rudolf-Elle' Hospital, Friedrich Schiller University Jena, Eisenberg, Germany2 Department of Biochemistry, Rush Medical College Head, Chicago, Illinois, USA3 Institute of Pathology, Friedrich Schiller University Jena, Germany4 Institute of Animal Studies, Friedrich Schiller University Jena, Germany5 Institute of Biochemistry 2, Friedrich Schiller University Jena, Germany6 Hans Knoell Institute for Natural Products Research, Jena, Germany7 Experimental Rheumatology Unit, Friedrich Schiller University Jena, Germany2005 31 3 2005 7 3 R677 R686 6 12 2004 23 2 2005 23 2 2005 1 3 2005 Copyright © 2005 Roth et al, licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is cited.
To assess the potential use of hyaluronic acid (HA) as adjuvant therapy in rheumatoid arthritis, the anti-inflammatory and chondroprotective effects of HA were analysed in experimental rat antigen-induced arthritis (AIA). Lewis rats with AIA were subjected to short-term (days 1 and 8, n = 10) or long-term (days 1, 8, 15 and 22, n = 10) intra-articular treatment with microbially manufactured, high-molecular-weight HA (molecular weight, 1.7 × 106 Da; 0.5 mg/dose). In both tests, 10 buffer-treated AIA rats served as arthritic controls and six healthy animals served as normal controls. Arthritis was monitored by weekly assessment of joint swelling and histological evaluation in the short-term test (day 8) and in the long-term test (day 29). Safranin O staining was employed to detect proteoglycan loss from the epiphyseal growth plate and the articular cartilage of the arthritic knee joint. Serum levels of IL-6, tumour necrosis factor alpha and glycosaminoglycans were measured by ELISA/kit systems (days 8 and 29). HA treatment did not significantly influence AIA in the short-term test (days 1 and 8) but did suppress early chronic AIA (day 15, P < 0.05); however, HA treatment tended to aggravate chronic AIA in the long-term test (day 29). HA completely prevented proteoglycan loss from the epiphyseal growth plate and articular cartilage on day 8, but induced proteoglycan loss from the epiphyseal growth plate on day 29. Similarly, HA inhibited the histological signs of acute inflammation and cartilage damage in the short-term test, but augmented acute and chronic inflammation as well as cartilage damage in the long-term test. Serum levels of IL-6, tumour necrosis factor alpha, and glycosaminoglycans were not influenced by HA. Local therapeutic effects of HA in AIA are clearly biphasic, with inhibition of inflammation and cartilage damage in the early chronic phase but with promotion of joint swelling, inflammation and cartilage damage in the late chronic phase.
==== Body
Introduction
Rheumatoid arthritis (RA), a chronic systemic disease primarily affecting the joints, is characterised by progressive destruction of cartilage and bony structures of the joints [1,2]. Its social impact results from the personal suffering of patients as well as from medical and indirect costs [3].
Hyaluronic acid (HA) is a large linear glycosaminoglycan composed of repeating disaccharide units of glucuronic acid and N-acetylglucosamine, linked via the 1–4 position of the sugar rings [4]. The synovial fluid in the joint consists of ultrafiltrated plasma and HA, the latter being produced by type-B synoviocytes of the lining layer [5]. Inflammatory changes lead to depolymerisation of HA, resulting in a decrease of its molecular weight and its concentration [6]. Its lubricant properties decrease, contributing to the destruction of cartilage and bone [7].
HA protects cells and anatomical structures against mechanical overloading due to its viscoelastic characteristics [8]. The viscosity of the synovial fluid is reduced in patients with RA [9], a deficit that can be balanced by the supply of exogenous HA [10]. In addition, the production of endogenous synovial HA is stimulated via the supply of exogenous HA [11].
RA is characterised by a loss of proteoglycans in the affected joints [12,13]. HA possesses chondroprotective effects [10,14] and is reported to inhibit the loss of proteoglycans from the matrix of joint cartilage [15,16]. HA also blocks the loss of proteoglycans caused by the addition of catalytic cytokines to cultivated cartilage [17,18] and suppresses the degradation of cartilage matrix mediated by fibronectin fragments [19,20]. HA is also reported to protect the cartilage against proteoglycan loss, against chondrocyte cell death caused by free oxygen radicals, IL-1, or mononuclear-cell-enriched medium, and against other alterations [14,15,21-24]. Cartilage degradation induced by neutrophil leukocytes is also reduced by HA in vitro [25]. Injection of exogenous HA induces a decrease of inflammatory and proliferative processes within the synovium [26]. Also, HA inhibits the proliferation [27] and migration of white blood cells [28], and affects their adherence, chemotaxis, and phagocytosis properties [11,29,30]. Degradation of HA by reactive oxygen species, on the other hand, may reduce the protective properties of HA [14,31].
In spite of the known potential benefits of HA on a number of pathological features of RA, a general estimate of its validity for the treatment of RA is still lacking, particularly in terms of experimental studies in animal models of arthritis. The present study was therefore designed to examine the effects of HA in rat antigen-induced arthritis (AIA). This experimental monoarticular arthritis shares some characteristics of RA; for example, hyperplasia of the synovial membrane, inflammatory infiltration of the joints, and destruction of cartilage [32]. This model is also useful to characterise treatment responses; for example, the reduction of inflammation or changes in the synovial connective tissue [33].
Materials and methods
Animals
Female Lewis rats (10–12 weeks of age) were obtained from the Institute of Animal Studies, Friedrich Schiller University Jena, Germany. The rats were housed under standard conditions, in a 12-hour light/dark cycle. The animals were fed with standard rodent chow and water ad libitum. The rats were divided into two groups: non-arthritic animals (n = 6) and arthritic animals (n = 40). The latter were subdivided into the following groups (each n = 10): untreated AIA rats, short-term test (US); untreated AIA rats, long-term test (UL); HA-treated AIA rats, short-term test (HS); and HA-treated AIA rats, long-term test (HL). All animal studies were approved by the governmental committee for animal protection.
Hyaluronic acid
Pyrogen-free, sterile-filtered HA with a molecular weight of 1.7 × 106 Da was used, obtained by biotechnological fermentation from Streptococcus equisimilis ssp. zooepidemicus V 2541. This bacterial HA, also called non-animal-source hyaluronan, is completely identical to human HA. The content of pyrogen was minimised to less than 0.05 IE/ml HA by cleaning steps, therefore fulfilling the demands of the European Pharmacopeia (Supplement 2001, page 1472). The zero-viscosity of the purified 1.0% high-molecular-weight HA (molecular weight, 1.7 × 106 Da) in 0.9% NaCl solution amounted to h0 = 10.74 Pa s. The injection units contained 10 mg HA in 1 ml of 0.9% NaCl.
Induction of AIA
All experimental animals were immunised by two subcutaneous injections (days -21 and -14) of 0.5 g methylated bovine serum albumin (mBSA), dissolved in 0.5 ml saline and emulsified with 0.5 ml complete Freund's adjuvant [32,34]. Knee monoarticular arthritis was induced 2 weeks after the second immunisation via a single joint injection of 0.5 mg mBSA (50 μl of 10 mg/ml mBSA dissolved in 0.9% NaCl) into the right knee joint (day 0 of AIA). The left knee remained without injection.
Treatment with HA
On day 1 of AIA, all 40 arthritic animals received an intra-articular injection into the right inflamed knee joint. HA-treated AIA rats (groups HS and HL) received in each case 0.5 mg HA (50 μl of 10 mg/ml HA in 0.9% NaCl), whereas the untreated AIA rats (groups US and UL) received 50 μl PBS. The AIA rats of the long-term test received further injections at the beginning of each subsequent week (days 8, 15, and 22): the HL group received 50 μl HA, and the UL group received 50 μl PBS.
The short-term test (groups US and HS) was terminated 1 week after the first injection of HA or PBS (day 8). The long-term test (groups UL and HL) was terminated 1 week after the fourth injection (day 29).
In all cases, the contralateral (left) knee joint remained untreated. The group of six non-arthritic animals without AIA (12 weeks of age) served for the collection of normal values.
All injections (including those necessary to induce immunisation and knee AIA) were performed under ether anaesthesia. At the end of the experiment, the animals were sacrificed using an overdose of CO2 and cervical dislocation.
Collection of samples
Blood samples were collected by heart puncture after opening the thorax. The blood was centrifuged for 10 min at 3000 × g and ambient temperature. The serum was divided into three portions of at least 250 μl and was frozen at -80°C until analysis.
The knee joints were disconnected from the long bones and stored in 6% formaldehyde. In order to ensure an optimal impregnation with formaldehyde, the adhering remainders of the long bones were kept very short (approximately 1.0 cm above and below the joint space) and the dorsal joint capsule was opened.
Evaluation of arthritis
Joint swelling, body weight, and the general state of the animals were regularly monitored. The measurements of weight and mediolateral joint diameter took place on days 0, 1, 4, 8, 15, 22, and 29. The mediolateral joint diameter was measured using a vernier caliper [32,34].
Histological analyses
All preparations were stored in 6% formaldehyde for 24 hours. Decalcification in ethylenediamine tetraacetic acid subsequently took place and the preparations were embedded in paraffin. After the removal of paraffin, 5-μm thick sections were cut [35].
For the assessment of the histological arthritis scores, the sections were stained with haematoxylin and eosin. All slides were evaluated by an independent observer who was blinded to the design and details of the study. In all cases, three sections per knee joint were examined and scored using a semiquantitative scale.
The extent of acute joint inflammation – as defined by the degree of infiltration of the synovial membrane by polymorphonuclear leukocytes, and defined by the exudation of granulocytes in the joint space – was evaluated in each case with 0 = no changes, 1 = mild changes, 2 = moderate changes, and 3 = severe changes. In addition, the presence (score 1) or absence (score 0) of fibrin exudation in the joint space and periarticular inflammation was assessed, resulting in a maximum total score of 8 for acute inflammation.
Chronic joint inflammation – based on the parameters hyperplasia of synovial lining cells, infiltration by mononuclear cells, and fibrosis of synovial membrane or periarticular tissue – was evaluated with a score of 0–3, resulting in a maximum total score of 9.
The extent of the damage to articular cartilage and adjacent bone structures (cell necrosis, structural bone, and cartilage defects) was evaluated with score 0 = no damage, score 1 = <5% of the cartilage surface affected, score 2 = 5–10% of the cartilage surface affected, score 3 = 10–50% of the cartilage surface affected, and score 4 = >50% of the cartilage surface affected (maximal total score of 4).
Safranin O staining was performed to estimate the proteoglycan content in the cartilage [36-38]. In order to obtain comparable histological results, all slides were stained using exactly the same procedure [39]. The preparations were analysed under defined conditions using a Zeiss microscope Axiovert 200 M (20 × magnification) (Carl Zeiss, Göttingen, Germany)] and the results were stored as pixel pictures. The staining intensity was determined in 175 × 25 mm2 areas, using Scion Image software (Scion Corporation, Frederick, MD, USA).
First, the staining intensity (red) at the epiphyseal growth plate of the femoral condyle of non-arthritic and arthritic animals was measured (maximum value 255). The arithmetic mean obtained from these values was used as a reference value (232 [= 100%]). The measurements of articular cartilage took place at the most distal point of the curvature of the femoral condyle. In each case, values were obtained for the superficial layer, middle layer, and deep layer of the hyaline cartilage, as well as for the calcified cartilage layer (Fig. 1). Data were expressed as a percentage of the reference value. Subsequently, the values of the contralateral, non-arthritic knee joint (left) were subtracted from the arthritic knee (right), resulting in negative values in the case of proteoglycan loss.
Cytokine and serum glycosaminoglycan evaluation
The serum levels of IL-6, tumour necrosis factor alpha (TNF-α) and glycosaminoglycan (GAG) were determined at the end point of the short-term test (day 8) and at the endpoint of the long-term test (day 29).
The serum levels of IL-6 and TNF-α were determined using a commercial sandwich ELISA kits for rats according to the manufacturer's instructions (Biosource International, Camalliro, CA, USA). The detection limits were 8 pg/ml for IL-6 and 4 pg/ml for TNF-α. According to the manufacturer, there was no cross-reactivity with other rat cytokines.
The serum levels of total GAG were measured in non-diluted serum with a commercially available kit. The standard values for healthy rats were 10.8–17.4 mg/l (Glycane T Labor + Diagnostica, Freital, Germany).
Statistics
Statistical evaluations were carried out using the programme SigmaStat 2.0. Since nearly all data were not normally distributed, the non-parametric Mann–Whitney U test was used. Data were expressed as means and standard errors of the means. P ≤ 0.05 was considered statistically significant for α.
In cases in which P values for α were at the limits of significance (0.05 ≤ P ≤ 0.1; joint swelling day 29, cartilage damage day 8), the statistical power of the U test was determined using the actual difference at a given time point as delta.
Because for the time period from day 0 to day 8 the procedure and results did not differ between the US and UL groups or between the HS and HL groups, respectively, the values from the short-term test and the long-term test were pooled for statistical evaluation of this period in both cases.
Results
Body weight
At baseline, the body weight was 188 ± 29 g (untreated AIA rats) and 197 ± 22 g (HA-treated AIA rats). After a plateau between day 0 and day 8 in both untreated rats and HA-treated AIA rats, the body weight rose in concomitance with the decrease of arthritis severity. At the end of the long-term test (day 29), the animals weighed 213 ± 15 g (untreated AIA rats) and 230 ± 13 g (HA-treated AIA rats). The differences between the groups did not reach statistical significance at any time point.
Joint swelling
On day 1, AIA developed as a significant swelling of the right knee joint in all animals (Fig. 2). The swelling increased up to day 4 in untreated AIA rats (P < 0.001, n = 20), significantly decreasing on day 8 (P < 0.001, n = 20). The swelling then continued to slowly decrease until day 29 (P < 0.001, n = 10). At all time points after initiation of AIA, the swelling remained significantly higher compared with the baseline levels on day 0 (Fig. 2).
Intra-articular treatment with HA did not significantly affect the degree of joint swelling on days 1, 4, and 8 (Fig. 2). On day 15 (groups UL and HL, n = 10 each) there was a significant reduction of joint swelling in the HA-treated AIA group compared with the untreated AIA group (P < 0.05). On day 22 the swelling was no longer significantly different from the untreated AIA group (P = 0.37); in fact, it was even somewhat higher. On day 29 (end of the long-term test) the small increase of joint swelling in the HA-treated AIA group persisted (as compared with the untreated AIA group), although without reaching statistical significance (power 1 β = 0.851).
In general, therefore, HA seemed to positively affect the early chronic phase of AIA (day 15), but did not have an influence on the acute or late chronic phases of AIA, at least in terms of joint swelling.
Loss of proteoglycans from the epiphyseal growth plate of the femoral condyle and the articular cartilage
In the short-term test (day 8), the untreated AIA group was characterised by a significant decrease of the proteoglycan content in the epiphyseal growth plate of the arthritic right knee compared with the contralateral knee or with the right knee joint of non-arthritic animals (in both cases, P < 0.05; Fig. 3). Treatment with HA prevented this loss, maintaining proteoglycan levels close to those of non-arthritic animals.
In terms of individual zones of the articular cartilage, the untreated AIA rats underwent a change of -37% in the superficial layer, -26% in the middle layer, -13% in the deep layer, and -15% in the calcified cartilage layer (Figs 4b,d and 5). At this time point, treatment with HA was significantly effective in preventing the proteoglycan loss in the superficial layer (P < 0.01), the middle layer (P < 0.05), and the calcified cartilage layer (P < 0.05; Figs 4b,f and 5). In all layers, the proteoglycan content reached normal levels.
In the long-term test (day 29) there was no significant loss of proteoglycan content in the epiphyseal growth plate of untreated AIA rats (see Fig. 3). However, treatment with HA was characterised by a significant proteoglycan loss in the growth plate of the arthritic right knee compared with the contralateral knee or with the right knee joint of non-arthritic animals (P < 0.001; Fig. 3).
In the different layers of the articular cartilage, the untreated AIA rats no longer showed any significant proteoglycan loss; that is, there were no significant differences between the right knee joints and left knee joints of AIA rats, or between the right knee joint of AIA rats and the right knee joint of non-arthritic animals (Fig. 5; Safranin O staining data not shown). Treatment with HA did not significantly affect the proteoglycan content in any layer of the articular cartilage.
Histological scores of arthritis and cartilage damage
In the short-term test (day 8) there was a strong acute inflammation in untreated AIA rats (Fig. 6a). Treatment with HA significantly reduced the acute inflammation compared with the untreated AIA group (P < 0.05; Figs 4a,c,e and 6a). Notably, the untreated AIA group underwent a complete, spontaneous remission of the acute inflammation score from day 8 to day 29 (acute inflammation score nearly 0 in the long-term test; Figs 4c,g and 6a). The HA-treated AIA group in the long-term test clearly improved compared with the short-term test (P < 0.05), but it still showed significantly higher, residual acute inflammation than the untreated AIA group (P < 0.05; Figs 4e,g,h and 6a).
In the short-term test (day 8) a clear score for chronic inflammation was also observed (Figs 4a,c,e and 6b), without significant differences between untreated and HA-treated AIA groups. The chronic inflammation significantly decreased from day 8 to day 29 in untreated and HA-treated AIA (group US versus group UL, P < 0.001; group HS versus group HL, P < 0.01; Figs 4c,e,g,h and 6b). Unexpectedly, however, on day 29 the chronic inflammation score was more pronounced in the animals treated with HA compared with the untreated AIA group (P < 0.05; Figs 4g,h and 6b).
In terms of cartilage damage, the untreated AIA group was characterised by a maximum individual score of 3; that is, the maximal possible score of 4 was not observed (Fig. 6c). On day 8, the mean cartilage damage was somewhat more pronounced in the untreated AIA group, but without significant differences in comparison with the HA-treated AIA group (power 1 - β = 0.821). From day 8 to day 29, the cartilage damage decreased significantly in untreated rats and HA-treated AIA rats (group US versus group UL, P < 0.001; group HS versus group HL, P < 0.01). In the long-term test (day 29), however, the cartilage damage was significantly higher in the animals treated with HA than in the untreated AIA group (P < 0.05; Fig. 6c).
Systemic cytokine levels
In non-arthritic animals, the serum IL-6 levels were below the detection limit of the assay (Fig. 7a). Untreated AIA rats had significantly elevated IL-6 levels both in the short-term test and in the long-term test (P < 0.001 in both cases; significance not indicated in Fig. 7). Treatment with HA did not significantly influence IL-6 levels at either time point (Fig. 7a).
As for TNF-α, non-arthritic animals had mean serum levels of 5.45 ± 4.56 pg/ml (Fig. 7b). In the short-term test these values were increased both in the untreated and in the HA-treated AIA groups, but not to a significant degree. In the long-term test, the mean TNF-α levels were very similar to those of non-arthritic animals. Treatment with HA did not significantly influence TNF-α levels at either time point (Fig. 7b).
Serum GAG levels
In non-arthritic animals, the mean serum levels of GAG were 12.70 ± 3.30 μg/ml (Fig. 7c). Untreated AIA rats had significantly higher GAG levels than non-arthritic animals in the short-term test and in the long-term test (P < 0.05 and P < 0.001, respectively; significance not indicated in Fig. 7). Treatment with HA had no influence on this parameter at either time point (Fig. 7c).
Discussion
Clinical parameters of arthritis
The time course of AIA was similar to that described by other authors [32,34], confirming that the present results were representative of previous studies.
Treatment with HA did not reduce joint swelling in the acute phase, as significant reduction of joint swelling was found only on day 15 (i.e. in the early chronic phase of AIA). The temporary reduction of joint swelling may be a result of the reduced acute inflammation observed histologically at an earlier time point (day 8). This anti-inflammatory effect of HA is consistent with the effects previously reported in collagen-induced arthritis [22,40] and human RA [41,42].
Interestingly, however, while the joint swelling continued to progressively and spontaneously decrease in untreated AIA, it persisted in HA-treated animals after day 15, although the difference remained at the limits of significance (P = 0.06). The persistence of inflammation after prolonged application of HA (day 29) is further substantiated by significant proteoglycan loss in the epiphyseal growth plate, by significant persistence of both acute and chronic inflammation, and by significantly increased histological signs of articular cartilage damage. This biphasic course regarding joint inflammation and destruction after repeated application of high-molecular-weight HA has not been previously described. HA therefore probably shows only a limited temporal window of anti-inflammatory activity in arthritis.
Adverse reactions to HA have been described in human osteoarthritis, either after the first injection and subsequent intra-articular injections or at the beginning of a new treatment course. These adverse events consisted of pain and/or transient swelling of the injected joint, mostly mild or moderate in intensity [43-45]. The adverse reactions observed upon intra-articular treatment of human osteoarthritis are not comparable with the biphasic effects in the present study, since they were only short-lasting and limited to a period immediately following injection.
Proteoglycans in the epiphyseal growth plate of the femoral condyle and the articular cartilage
AIA was accompanied by a significant loss of proteoglycans in the epiphyseal growth plate of untreated AIA rats in the short-term test (significantly ameliorated by HA; day 8). This proteoglycan loss is probably caused by the inflammatory micromilieu in the arthritic joint [46] and the adjacent periarticular bone marrow [47], in analogy to the alterations of the primary and secondary spongiosa in AIA [48]. The significant prevention of proteoglycan loss from the epiphyseal growth plate by HA at this time point may be due to the clear anti-arthritic effects of HA, as also documented by the significant decrease of acute inflammation on day 8 (Fig. 6a). Whether intra-articular treatment with HA indirectly influences the inflammatory changes in the bone marrow, thereby preventing proteoglycan loss in the epiphyseal growth plate, remains to be investigated.
Unexpectedly, repeated intra-articular HA treatment induced proteoglycan loss in the epiphyseal growth plate in the long-term test (day 29). This late loss of proteoglycans in the epiphysis under prolonged HA treatment suggests late proinflammatory effects of HA, as also indicated by significantly elevated levels of acute and chronic inflammation, as well as cartilage damage (Fig. 6).
Regarding the proteoglycan content within the articular cartilage, HA completely prevented the proteoglycan loss in the short-term test. This applied to the severe loss of proteoglycans in the superficial layer — the layer most strongly affected in the present study (37%) and also that most strongly affected in the fibronectin-mediated arthritis model [19]. This prevention also applied to deeper layers of the calcified cartilage, however, indicating either deep-reaching effects of intra-articular HA (up to 100 μm; see Fig. 1) or an indirect effect on the micromilieu in the epiphyseal bone core. Such profound chondroprotective effects of HA have not previously been reported in an in vivo arthritis model. Previous in vitro studies, in which the proteoglycan release from the cell matrix of bovine chondrocyte cultures was inhibited by HA, have suggested a covering of the matrix of as a mechanism for the chondroprotective capacity of HA [19].
Acute and chronic inflammation, and cartilage damage
HA reduced the histological signs of acute inflammation in the short-term test (day 8). At this point, HA also showed a tendency (P = 0.06) to protect the joint again against cartilage damage (Fig. 6c). The tendency for decreased cartilage damage in the short-term test supports the assumption that HA forms a temporary protecting barrier over the cartilage, and thereby protects it against degradation [16,19,49]. The known reduction of free oxygen radicals [31], as well as the reduction of cytokines and other mediators of acute inflammation by HA [15,21], could contribute to both its chondroprotective capacity (Fig. 6c) and its anti-inflammatory effects (Fig. 6a) in early chronic AIA (day 8).
In the long-term test (day 29) HA treatment was associated with significantly higher histological signs of acute and chronic inflammation and, more importantly, with more severe cartilage damage. The late proinflammatory effects of repeated HA application, in parallel to the late loss of epiphysis proteoglycans and late damage of articular cartilage (see earlier), point to an interdependence of inflammation and damage in HA-treated AIA rats at this stage.
Significantly higher histological signs of articular cartilage damage after repeated HA application (day 29) are in contrast to a non-altered proteoglycan content of the articular cartilage at this time (Fig. 5). A seemingly normal proteoglycan content may therefore not be sufficient to exclude structural damage of the articular cartilage in arthritis. Consequently, more sensitive in vivo procedures will have to be established to reliably assess functional or structural cartilage alterations before irreversible damage occurs [50].
Serum levels of cytokines and GAG
In the present study, the systemic levels of IL-6, TNF-α and GAG were not significantly influenced by intra-articular administration of high-molecular-weight HA, indicating that both anti-inflammatory effects (day 8) and proinflammatory effects (day 29) are locally restricted.
Conclusion
In conclusion, HA appears to have a limited therapeutic window for local treatment of arthritis, as shown by amelioration of clinical signs (day 15), by prevention of proteoglycan loss in the articular cartilage and epiphyseal growth plate, and by prevention of structural cartilage damage, as well as by the reduction of acute inflammation in the arthritic joint (day 8).
Late aggravation of clinical signs (not significant, day 29), proteoglycan loss from the epiphyseal growth plate, and acute/chronic inflammation and structural cartilage damage at this time point strongly indicate biphasic effects of local HA treatment.
Whether these biphasic effects are due to accumulation of HA beyond pathological levels (which may be avoidable by single injection instead of repeated injection) or whether certain phases of the clinical course render the animals sensitive to the proinflammatory effects of HA remains the subject of future research, both in animal models and in human RA.
Abbreviations
AIA = antigen-induced arthritis; ELISA = enzyme-linked immunosorbent assay; GAG = glycosaminoglycan; HA = hyaluronic acid; HL = HA-treated AIA rats, long-term test; HS = HA-treated AIA rats, short-term test; IL = interleukin; mBSA = methylated bovine serum albumin; PBS = phosphate-buffered saline; RA = rheumatoid arthritis; TNF-α = tumour necrosis factor alpha; UL = untreated AIA rats, long-term test; US = untreated AIA rats, short-term test.
Competing interests
The co-authors J-H Ozegowski, P-J Müller, S Möller, G Peschel, A Roth, and RA Venbrocks published a patent for the use of sulfated hyaluronic acid for the prevention of inflammatory arthritis in 2000. This substance is not the same as that used for the experiments in the present study. (Ozegowski J-H, Müller P-J, Möller S, Peschel G, Roth A, Venbrocks RA: Pharmazeutische Formulierungen zur Hemmung von entzündlichen Arthritiden [Use of hyaluronic acid derivatives for the prevention of inflammatory arthritis], HA 00-52 2000).
A Roth has to publish papers in peer-reviewed journals as a part of the process to finish his thesis (habilitation in Germany). This is a non-financial academic interest.
Authors' contributions
AR carried out all experiments, the measurements and evaluation of the data, as well as the statistics, and drafted, revised, finalised, and submitted the manuscript. JM developed the experimental basis for the measurements of proteoglycans in the serum and supervised the analysis, trained and supervised AR in the analysis of proteoglycan loss from cartilage, actively participated in the analysis and evaluation of the data, and reviewed and contributed to the final version of the manuscript. AW participated in the preparation of the animals, fixation of the specimens, and blood collection, and reviewed the manuscript. RF actively participated in the experimental and organisational design of the study, gave valuable advice for the evaluation and interpretation of the experimental results, and reviewed the manuscript. AS reviewed all experimental data, gave valuable advice for the evaluation and interpretation of the experimental results and the subsequent conclusions, and reviewed the manuscript. RAV participated in the experimental and organisational design of the study, supervised the evaluation and interpretation of the experimental results, reviewed the manuscript, and organised financial support and laboratory space for the experiments. PP performed and summarised all histological analyses, and reviewed the manuscript. RB actively participated in the underlying animal experiments and the evaluation of the experimental results, and reviewed and contributed to the final version of the manuscript. HS actively participated in the underlying animal studies, provided and monitored all experimental animals, and reviewed the manuscript. JO, GP, and PJM developed the method for the production of HA, produced, purified, and quality-controlled the microbially manufactured, high-molecular-weight HA, and reviewed the manuscript. RWK participated in the experimental design of the study and the interpretation of the results, and reviewed and contributed to the initial and final version of the manuscript.
Acknowledgements
The authors are grateful to Dr Frank Brand, Mrs K Neumann, and Mrs K Axt (Department of Clinical Chemistry and Laboratory Diagnostics, Rudolf-Elle Hospital) for determination of laboratory values, and to Cordula Müller and Jana Schömburg (Research Department, Rudolf-Elle Hospital) for preparation of the histological sections. They are also grateful to Dr A Notni, Dr K Bergmann, and Dr R Winter (Department of Orthopaedics, Rudolf-Elle Hospital), as well as to Dr C Wicher and Mrs P Dobermann (Institute of Animal Studies, Friedrich-Schiller University Jena) for assistance in the animal experiments.
Figures and Tables
Figure 1 Measurement frames for Safranin O staining of the knee joint cartilage. After elimination of green tones and transformation of all red tones into grey tones, the staining intensity (a measure of the proteoglycan content) was determined in the following layers: S, superficial layer; M, middle layer; D, deep layer; and C, calcified cartilage.
Figure 2 Time course of knee joint swelling. Joint swelling (difference between the bilateral diameter of the right knee and the left knee) in untreated antigen-induced arthritis (AIA) rats and in hyaluronic acid (HA)-treated AIA rats. V, end of the short-term test (day 8) and end of the long-term test (day 29). The arrows indicate the days of intra-articular injection of HA (days 1, 8, 15, and 22). In the short-term test there was no significant difference between HA-treated rats and untreated AIA rats. In the long-term test HA-treated AIA rats showed significantly reduced values on day 15 (* P < 0.05). On day 29 there were no longer differences between the two groups; if at all, the swelling in the HA-treated group was somewhat higher than in untreated AIA group.
Figure 3 Safranin O staining intensity in the epiphyseal growth plate of the femoral condyle. The reference value of 232 (100%; continuous line) was obtained by computing all available values from both non-arthritic rats and antigen-induced arthritis (AIA) rats. In untreated AIA rats, the right (arthritic) joint showed a significant reduction of proteoglycan content of the epiphysis in the short-term test (day 8; *P < 0.05). This loss was not observed following hyaluronic acid (HA) treatment (day 8). The latter values were comparable with non-arthritic animals and with the contralateral joint (data not shown). Long-term treatment with HA (day 29) induced a significant loss of proteoglycans in the epiphyseal growth plate (*** P < 0.001). In contrast, the arthritic joints of untreated AIA rats showed values comparable with non-arthritic rats and contralateral joints (not shown).
Figure 4 Histological findings in synovial tissue and articular cartilage: haematoxylin and eosin (HE) staining (a, c, e, g, and h) for acute inflammation (arrowheads), chronic inflammation (*), and cartilage damage (arrows), as well as Safranin O staining (b, d, and f) for proteoglycan depletion (arrows). Images are shown for non-arthritic rats (a and b), untreated antigen-induced arthritis (AIA) rats (day 8, c and d; day 29, g), and hyaluronic acid hyaluronic acid (HA)-treated AIA rats (day 8, e and f; day 29, h). The bar indicates the distance in the histological section. SM, synovial membrane; P, patella; FE, femur. Safranin O staining: S, superficial layer; M, middle layer; D, deep layer; and C, calcified cartilage. N, non-arthritic rats; US, untreated AIA, short-term test; HS, HA-treated AIA, short-term test; UL, untreated AIA, long-term test; HL, HA-treated AIA, long-term test.
Figure 5 Safranin O staining intensity in different layers of the articular cartilage. Comparisons were made for different layers (S, superficial layer; M, middle layer; D, deep layer; and C, calcified cartilage) between non-arthritic rats, untreated antigen-induced arthritis (AIA) rats, and hyaluronic acid (HA)-treated AIA rats in terms of relative differences between right (arthritic) and left (contralateral) joints. In the short-term test (day 8) untreated AIA rats showed a reduced proteoglycan content in the superficial layer (** P = 0.01), middle layer (* P < 0.05), and calcified cartilage (* P < 0.05). HA-treated AIA rats showed proteoglycan contents comparable with those of non-arthritic rats in all layers. In the long-term test (day 29), untreated AIA rats also showed reduced proteoglycan contents in all layers compared with non-arthritic rats, but no statistical significance was attained. HA-treated AIA rats showed proteoglycan contents comparable with those of non-arthritic rats.
Figure 6 Histological scores. In the short-term test, hyaluronic acid (HA)-treated antigen-induced arthritis (AIA) rats showed a significant reduction of (a) the acute inflammation score compared with untreated AIA rats (* P < 0.05). The score of (b) chronic inflammation and (c) cartilage damage did not show significant differences between HA-treated rats and untreated AIA rats. In the long-term test, (a) the acute inflammation was reduced in both AIA groups compared with that in the short-term test (HA-treated AIA rats, P < 0.05; untreated AIA rats, P < 0.001; significance not indicated); nonetheless, the HA-treated AIA rats showed significantly higher scores than untreated AIA rats on day 29 (* P < 0.05). (b) The scores of chronic inflammation were reduced in both AIA groups compared with the short-term test (untreated AIA rats, P < 0.001; HA-treated AIA rats, P < 0.01; significance not indicated); nonetheless, the HA-treated rats showed significantly higher scores than untreated AIA rats on day 29 (* P < 0.05). (c) The cartilage damage was relatively low in both untreated rats and HA-treated AIA rats, but HA-treated AIA rats showed a significantly higher damage (* P < 0.05).
Figure 7 Serum levels of IL-6, tumour necrosis factor alpha (TNF-α) and glycosaminoglycan. During the course of antigen-induced arthritis (AIA), (a) IL-6 levels in non-arthritic rats showed values below the detection limit of the assay. Untreated and hyaluronic acid (HA)-treated AIA rats showed significantly increased values in the short-term test and in the long-term test compared with non-arthritic rats (all P < 0.001; data not shown), but HA treatment resulted in IL-6 levels comparable with those of untreated AIA. (b) Regarding TNF-α levels, the only change was a short-term, non-significant increase in all AIA rats, whether HA-treated or untreated. (c) Serum values of glycosaminoglycan increased significantly in all AIA rats compared with non-arthritic animals (non-arthritic animals versus untreated AIA rats [short-term test], P < 0.05, non-arthritic animals versus HA-treated AIA rats, untreated AIA rats [long-term test] and HA-treated AIA rats [long-term test], P < 0.001; significance not indicated). There were no differences between short-term tests and long-term tests, or between HA-treated rats and untreated AIA rats.
==== Refs
Firestein GS Zvaifler NJ Gallin J, Goldstein I, Willis K Rheumatoid arthritis: a disease of disordered immunity Inflammation: Basic Principles and Clincal Correlates 1992 New York: Raven Press 959 975
Gay S Gay RE Koopman WJ Molecular and cellular mechanisms of joint destruction in rheumatois arthritis: two cellular mechanisms explain joint destruction? Ann Rheum Dis 1993 52 Suppl 1 39 47
Mau W Bornmann M Weber H Weidemann HF Hecker H Raspe HH Prediction of permanent work disability in a follow-up study of early rheumatoid arthritis: results of a tree structured analysis using RECPAM Br J Rheumatol 1996 35 652 659 8670599
Balazs EA Iozzo RV Matrix proteoglycans: from molecular design to cellular function Annu Rev Biochem 1998 67 609 652 9759499 10.1146/annurev.biochem.67.1.609
Laurent TC Laurent UB Fraser JR The structure and function of hyaluronan: an overview Immunol Cell Biol 1996 74 1 7 8934648
Miossec P Dinarello CA Ziff M Interleukin-1 lymphocyte chemotactic activity in rheumatoid arthritis synovial fluid Arthritis Rheum 1986 29 461 470 3486662
Dahl LB Dahl IM Engström-Laurent A Granath K Concentration and molecular weight of sodium hyaluronate in synovial fluid from patients with rheumatoid arthritis and other arthropathies Ann Rheum Dis 1985 44 817 822 4083937
Balazs EA Denlinger JL Sodium hyaluronate and joint function J Equine Vet Sci 1985 5 217 228
Duffy JM McNally GM O'Gorman A Spedding PL Grimshaw J Grimshaw JT Walker B Guthrie DJS Mollan RAB Viscometic studies of pathological human synovial fluids J Orthop Rheumatol 1994 7 99 106
Namiki O Toyoshima H Morisaki N Therapeutic effect of intra-articular injection of high molecular weight hyaluronic acid on osteoarthritis of the knee Int J Clin Pharmacol Ther Toxicol 1982 20 501 507 7174151
Watterson JR Esdaile JM Viscosupplementation: therapeutic mechanisms and clinical potential in osteoarthritis of the knee J Am Acad Orthop Surg 2000 8 277 284 11029555
Saxne T Wollheim FA Pettersson H Heinegard D Proteoglycan concentration in synovial fluid: predictor of future cartilage destruction in rheumatoid arthritis? Br Med J (Clin Res Ed) 1987 295 1447 1448 3121058
Bensouyad A Hollander AP Dularay B Bedwell AE Cooper RA Hutton CW Dieppe PA Elson CJ Concentrations of glycosaminoglycans in synovial fluids and their relation with immunological and inflammatory mediators in rheumatoid arthritis Ann Rheum Dis 1990 49 301 307 2344209
Sato H Takahashi T Ide H Fukushima T Tabata M Sekine F Kobayashi K Negishi M Niwa Y Antioxidant activity of synovial fluid, hyaluronic acid, and two subcomponents of hyaluronic acid. Synovial fluid scavenging effect is enhanced in rheumatoid arthritis patients Arthritis Rheum 1988 31 63 71 3345232
Larsen NE Lombard KM Parent EG Balazs EA Effect of hylan on cartilage and chondrocyte cultures J Orthop Res 1992 10 23 32 1727935 10.1002/jor.1100100104
Shimazu A Jikko A Iwamoto M Koike T Yan W Okada Y Shinmei M Nakamura S Kato Y Effects of hyaluronic acid on the release of proteoglycan from the cell matrix in rabbit chondrocyte cultures in the presence and absence of cytokines Arthritis Rheum 1993 36 247 253 8431214
Morris ES Wilcon S Treadwell BV Inhibition of interleukin 1-mediated proteoglycan degradation in bovine articular cartilage explants by addition of sodium hyaluronate Am J Vet Res 1992 53 1977 1982 1466488
Akatsuka M Yamamoto Y Tobetto K Yasui T Ando T In vitro effects of hyaluronan on prostaglandin E2 induction by interleukin-1 in rabbit articular chondrocytes Agents Actions 1993 38 122 125 8480532
Homandberg GA Hui F Wen C Kuettner KE Williams JM Hyaluronic acid suppresses fibronectin fragment mediated cartilage chondrolysis: I. In vitro Osteoarthritis Cartilage 1997 5 309 319 9497938
Williams JM Plaza V Hui F Wen C Kuettner KE Homandberg GA Hyaluronic acid suppresses fibronectin fragment mediated cartilage chondrolysis: II. In vivo Osteoarthritis Cartilage 1997 5 235 240 9404468
Presti D Scott JE Hyaluronan-mediated protective effect against cell damage caused by enzymatically produced hydroxyl (OH.) radicals is dependent on hyaluronan molecular mass Cell Biochem Funct 1994 12 281 288 7834818
Campo GM Avenoso A Campo S Ferlazzo AM Altavilla D Calatroni Efficacy of treatment with glycosaminoglycans on experimental collagen-induced arthritis in rats Arthritis Res Ther 2003 5 R122 R131 12723984 10.1186/ar748
Abatangelo G Cortivo R Govoni E De Galateo A Brun P Hyaluronate reverses inhibition of proteoglycan synthesis by oxygen free radicals in cultured chick embryonic cartilage [abstract] Eur J Cell Biol 1989 49 Suppl 28 18
Cortivo R Brun P Cardarelli L O'Regan M Radice M Abatangelo G Antioxidant effects of hyaluronan and its alpha-methyl-prednisolone derivative in chondrocyte and cartilage cultures Semin Arthritis Rheum 1996 26 492 501 8870117 10.1016/S0049-0172(96)80030-8
Tobetto K Nakai K Akatsuka M Yasui T Ando T Hirano S Inhibitory effects of hyaluronan on neutrophil-mediated cartilage degradation Connect Tissue Res 1993 29 181 190 8222645
Goldberg RL Toole BP Hyaluronate inhibition of cell proliferation Arthritis Rheum 1987 30 769 778 3619961
Anastassiades T Robertson W Modulation of mitogen-dependent lymphocyte stimulation by hyaluronic acid J Rheumatol 1984 11 729 734 6520826
Corrado EM Peluso GF Gigliotti S Durante C Palmieri D The effects of intraarticular administration of hyaluronan on osteoarthritis of the knee: a clinical study with immunological and biochemical evaluations Eur J Rheum Inflamm 1993 15 547 576
Forrester JV Balazs EA Inhibition of phagocytosis by high molecular weight hyaluronate Immunology 1980 40 435 446 7429537
Håkansson L Hällgren R Venge P Regulation of granulocyte function by hyaluronic acid. In vitro and in vivo effects on phagocytosis, locomotion, and metabolism J Clin Invest 1980 66 298 305 7400316
Greenwald RA Oxygen radicals, inflammation, and arthritis: pathophysiological considerations and implications for treatment Semin Arthritis Rheum 1991 20 219 240 2042055 10.1016/0049-0172(91)90018-U
Bräuer R Kittlick PD Thoss K Henzgen S Different immunological mechanisms contribute to cartilage destruction in antigen-induced arthritis Exp Toxicol Pathol 1994 46 383 388 7894251
Griffiths RJ Henderson B, Edwards JCW, Pettpher ER The use of animals in the search for anti-inflammatory drugs Mechanisms and Models in Rheumatoid Arthritis 1995 London: Academic Press
Buchner E Bräuer R Schmidt C Emmrich F Kinne RW Induction of flare-up reactions in rat antigen-induced arthritis J Autoimmun 1995 8 61 74 7734037
Romeis B Böck P Mikroskopische Technik Mikroskopische Technik 1989 17 Urban-Schwarzenberg, München
Rosenberg L Chemical basis for the histological use of safranin O in the study of articular cartilage J Bone Joint Surg Am 1971 53 69 82 4250366
Kiviranta I Jurvelin J Tammi M Saamanen AM Helminen HJ Microspectrophotometric quantitation of glycosaminoglycans in articular cartilage sections stained with Safranin O Histochemistry 1985 82 249 255 2581923 10.1007/BF00501401
Arokoski J Kiviranta I Jurvelin J Tammi M Helminen HJ Long-distance running causes site-dependent decrease of cartilage glycosaminoglycan content in the knee joints of beagle dogs Arthritis Rheum 1993 36 1451 1459 7692860
Meyer P Burkhardt H Palombo-Kinne E Grunder W Bräuer R Stiller KJ Kalden JR Becker W Kinne RW 123I-antileukoproteinase scintigraphy reveals microscopic cartilage alterations in the contralateral knee joint of rats with 'monarticular' antigen-induced arthritis Arthritis Rheum 2000 43 298 310 10693869 10.1002/1529-0131(200002)43:2<298::AID-ANR9>3.0.CO;2-G
Campo GM Avenoso A Campo S Ferlazzo A Altavilla D Micali C Calatroni A Aromatic trap analysis of free radicals production in experimental collagen-induced arthritis in the rat: protective effect of glycosaminoglycans treatment Free Radic Res 2003 37 257 268 12688421 10.1080/1071576021000046640
Tanaka N Sakahashi H Sato E Hirose K Ishima T Ishii S Intra-articular injection of high molecular weight hyaluronan after arthrocentesis as treatment for rheumatoid knees with joint effusion Rheumatol Int 2002 22 151 154 12172954 10.1007/s00296-002-0214-y
Goto M Hanyu T Yoshio T Matsuno H Shimizu M Murata N Shiozawa S Matsubara T Yamana S Matsuda T Intra-articular injection of hyaluronate (SI-6601D) improves joint pain and synovial fluid prostaglandin E2 levels in rheumatoid arthritis: a multicenter clinical trial Clin Exp Rheumatol 2001 19 377 383 11491492
Puttick MP Wade JP Chalmers A Connell DG Rangno KK Acute local reactions after intraarticular hylan for osteoarthritis of the knee J Rheumatol 1995 22 1311 1314 7562764
Allen E Krohn K Adverse reaction to Hylan GF-20 [Letter] J Rheumatol 2000 27 1572 10852302
Martens PB Bilateral symmetric inflammatory reaction to hylan G-F 20 injection Arthritis Rheum 2001 44 978 979 11315938 10.1002/1529-0131(200104)44:4<978::AID-ANR156>3.0.CO;2-N
Simon J Surber R Kleinstauber G Petrow PK Henzgen S Kinne RW Brauer R Systemic macrophage activation in locally-induced experimental arthritis J Autoimmun 2001 17 127 136 11591121 10.1006/jaut.2001.0534
Kinne RW Brauer R Stuhlmuller B Palombo-Kinne E Burmester GR Macrophages in rheumatoid arthritis Arthritis Res 2000 2 189 202 11094428 10.1186/ar86
Oelzner P Bräuer R Henzgen S Thoss K Wunsche B Hersmann G Abendroth K Kinne RW Periarticular bone alterations in chronic antigen-induced arthritis: free and liposome-encapsulated clodronate prevent loss of bone mass in the secondary spongiosa Clin Immunol 1999 90 79 88 9884355 10.1006/clim.1998.4623
Bräuer R Kette H Henzgen S Thoss K Influence of cyclosporin A on cytokine levels in synovial fluid and serum of rats with antigen-induced arthritis Agents Actions 1994 41 96 98 8079831
Gray ML Eckstein F Peterfy C Dahlberg L Kim YJ Sorensen AG Smith RL Toward imaging biomarkers for osteoarthritis Clin Orthop 2004 427 Suppl 175 181
| 15899053 | PMC1174961 | CC BY | 2021-01-04 16:02:36 | no | Arthritis Res Ther. 2005 Mar 31; 7(3):R677-R686 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1725 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17261589905410.1186/ar1726Research ArticleThe determinants of change in tibial plateau bone area in osteoarthritic knees: a cohort study Wang Yuanyuan [email protected] Anita E [email protected] Flavia M [email protected] Department of Epidemiology and PreventiveMedicine, Monash UniversityMedical School, Alfred Hospital, Prahran, Vic 3181, Australia2 Graduate School of IntegrativeMedicine, Swinburne University of Technology, Hawthorn, Vic 3122, Australia2005 31 3 2005 7 3 R687 R693 8 9 2004 19 10 2004 27 1 2005 2 3 2005 Copyright © 2005 Wang et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Bone is integral to the pathogenesis of osteoarthritis (OA). Whether the bone area of the tibial plateau changes over time in subjects with knee OA is unknown. We performed a cohort study to describe this and identify factors that might influence the change. One hundred and twenty-six subjects with knee OA underwent baseline knee radiography and magnetic resonance imaging on their symptomatic knee. They were followed up with a repeatmagnetic resonance image of the same knee approximately 2 years later. The bone area of the tibial plateau was measured at baseline and follow-up. Risk factors assessed at baseline were tested for their association with change in tibial plateau bone area over time. One hundred and seventeen subjects completed the study. The medial and lateral tibial plateau bone areas increased by 2.2 ± 6.9% and 1.5 ± 4.3% per year, respectively. Being male (P = 0.001), having a higher body mass index (P = 0.002), and having a higher baseline grade of medial joint-space narrowing (P = 0.01) were all independently and positively associated with an increased rate of enlargement of bone area of the medial tibial plateau. A larger baseline bone area of the medial tibial plateau was inversely associated with the rate of increase of that area (P < 0.001). No factor examined affected the rate of increase of the bone area of the lateral tibial plateau. In subjects with established knee OA, tibial plateau bone area increases over time. The role of subchondral bone change in the pathogenesis of knee OA will need to be determined but may be one explanation for the mechanism of action of risk factors such as body mass index on knee OA.
==== Body
Introduction
Osteoarthritis (OA) is the most common form of joint disease, with a prevalence of 10 to 30% in persons over the age of 65 [1]. However, the pathogenesis of this degenerative joint disease is not fully understood. It affects articular cartilage, subchondral bone, and soft tissues including synovium and ligaments. However, the instigating lesion remains unclear and controversial, with proponents for both cartilage and bone abnormalities being the initiating factor [2].
Changes in subchondral bone are well described in established OA, in human [3-7] as well as animal [8-13] models. These changes include remodelling of the subchondral trabeculae [3,4,8-11], stiffening of the subchondral bone [5,12,13], thickening of the subchondral plate [4,6,11], a steep stiffness gradient [5], and a decrease of the ability to absorb energy [7]. These changes affect the mechanical properties of the subchondral bone and have been proposed to play a role in the initiation and progression of degeneration of the overlying articular cartilage [5,13,14]. It has been suggested that the changes in subchondral bone may be the initiating factor in the pathogenesis of OA rather than sequelae of cartilage damage [15]. Changes in the area of subchondral bone may reflect the changes in its architecture. The enlargement of the bone area of the tibial plateau may attenuate the tibial cartilage, and this attenuation may play a role in the process of OA. However, whether the size of subchondral bone is static or changes over time in subjects with OA is unknown.
We studied a cohort of symptomatic subjects with predominantly mild to moderate knee OA over the course of 2 years, to determine whether the bone area of the tibial plateau changes over time and to identify the factors that might influence this change.
Materials and methods
Patients were recruited by using a combined strategy including advertising through local newspapers and the Victorian branch of the Arthritis Foundation of Australia, as well as collaboration with general practitioners, rheumatologists, and orthopaedic surgeons, as previously described [16]. The study was approved by the ethics committee of the Alfred and Caulfield hospitals inMelbourne, Australia. All the patients gave their informed consent.
One hundred and thirty-two subjects aged over 40 years who fulfilled American College of Rheumatology (ACR) clinical and radiographic criteria for knee OA [17] entered the study. Subjects were excluded if any other form of arthritis was present, if there was any contraindication tomagnetic resonance imaging (MRI), if a total knee replacement was planned, or if they were unable to cooperate with study requirements. We have previously described this population in a study of the determinants of change in the volume of tibial cartilage in knee OA [16]. Weight was measured to the nearest 0.1 kg (shoes, socks, and bulky clothing removed) using a single pair of electronic scales. Height was measured to the nearest 0.1 cm (shoes and socks removed) using a stadiometer. Body mass index (BMI) (weight/height2 in kg/m2) was calculated. General health, pain, stiffness, and function were assessed using the SF-36 (36-Item Short-Form Health Survey) [18] and WOMAC (Western Ontario andMcMaster Universities Osteoarthritis Index) [19]. Participants were asked to complete a questionnaire regarding demographic data and current physical activity [20].
At baseline, each subject had a weight-bearing anteroposterior tibiofemoral radiograph taken of the symptomatic knee in full extension. All radiographs were independently scored by two trained observers using a published atlas to classify disease in the tibiofemoral joint [21]. The radiographic features of tibiofemoral OA in each compartment were graded on a 4-point scale (0–3) for individual features of osteophytes and joint-space narrowing [21]. Intraobserver reproducibility was 0.93 for osteophytes and 0.93 for joint-space narrowing. Interobserver reproducibility was 0.86 for osteophytes and 0.85 for joint-space narrowing (by κ statistic) [22]. Where both knees were symptomatic and showed changes of radiographic OA, the knee with the least severe disease was used.
Each subject had anMRI performed on the symptomatic knee at baseline and approximately 2 years later. Knees were imaged in a sagittal plane on the same 1.5-T whole-body magnetic resonance unit (Signa Advantage HiSpeed GEMedical Systems, Milwaukee, WI, USA) using a commercial receive-only extremity coil. The following sequence and parameters were used: a T1-weighted, fat-suppressed 3D gradient recall acquisition in the steady state; flip angle 55 degrees; repetition time 58 ms; echo time 12 ms; field of view 16 cm; 60 partitions; 512 (frequency direction, superior–inferior) × 512 (phase encoding direction, anterior–posterior) matrix; one acquisition, time 11 min 56s. Sagittal images were obtained at a partition thickness of 1.5 mm and an in-plane resolution of 0.31 mm × 0.31 mm (512 × 512 pixels). One trained reader made the measurements in duplicate. The bone areas of the medial and lateral tibial plateaux were determined by means of image processing on an independent work station using the software program Osiris (Digital Imaging Unit, University Hospital of Geneva, Geneva, Switzerland), by creating an isotropic volume from the input images, which were reformatted in the axial plane, and then areas were directly measured from these axial images, as previously described [22-24]. To measure the bone area of the tibial plateau, we selected the first image that showed both tibial cartilage and subchondral bone. The area of medial and lateral tibial plateau bone was measured manually on this image and the next distal image (Fig. 1). An average of the two areas was used as an estimate of the tibial plateau bone area. The coefficients of variation (for the repeated image analysis) for the medial and lateral tibial plateau bone area were 2.3% and 2.4%, respectively [22].
Descriptive statistics for characteristics of the subjects were tabulated. Independent t-tests were used for comparison of means. The chi-square test was used to compare nominal characteristics between the groups. Principal outcome measures in analyses were annual change and annual percentage change of tibial plateau bone area. Change in this area was obtained by subtracting the bone area at baseline from that at follow-up. The annual change was calculated by dividing this figure by the time betweenMRI scans. The annual percentage change was obtained by dividing annual change by the baseline bone area and multiplying by 100 to obtain a percentage. Stepwise multiple linear regression techniques were used to explore the possible factors affecting annual percentage change in tibial plateau bone area, including age, gender, BMI, WOMAC score, SF-36 score, physical activity, radiographic features (grades of osteophytes and joint-space narrowing in the studied compartment) and baseline tibial plateau bone area. A P value less than 0.05 (two-tailed) was regarded as statistically significant. All analyses were performed using the SPSS statistical package (standard version 11.5.0, SPSS, Chicago, IL, USA).
Results
One hundred and thirty-two subjects fulfilled the study criteria and entered this study. MR images were available for measurement of tibial plateau bone area in 126 subjects (Table 1); 6MRIs were unavailable for technical reasons. Most subjects had mild to moderate tibiofemoral OA. Compared with women, men were significantly taller (P < 0.001) and heavier (P = 0.01) and had lower BMI (P = 0.02), higher level of physical activity (P = 0.01), less severe medial compartment OA (grade 0 to 1 joint-space narrowing in medial compartment) (P = 0.04), and larger medial (P < 0.001) and lateral (P < 0.001) tibial plateau bone area (Table 1). One hundred and seventeen subjects completed the longitudinalMRI component of the study. There were no significant differences between subjects who completed the study and those who did not (results not shown).
The medial tibial bone area increased from 2054.6 ± 363.9 mm2 to 2128.3 ± 370.0 mm2 (P < 0.001) and the lateral tibial bone area, from 1407.2 ± 256.7 mm2 to 1442.8 ± 272.4 mm2 over the study period (P < 0.001). Medial and lateral tibial plateau bone areas increased by 36.8 mm2 (P < 0.001) and 19.1 mm2(P = 0.004) per year, respectively, representing an annual increase rate of 2.2% and 1.5% (Table 2). Although there were no significant differences between men and women in annual increase or percentage increase of medial or lateral tibial plateau area in univariate analyses, after adjustment for potential confounders (age, BMI, physical activity, grade of osteophytes, grade of joint-space narrowing, and baseline tibial plateau bone area), men showed a significantly greater annual increase of medial tibial plateau bone area than women (P = 0.002) and a significantly greater annual percentage increase of medial tibial plateau bone area (5.4% in men compared with 0.004% in women, P = 0.001) (Table 2). There were no significant differences between men and women in the change in lateral tibial plateau bone area.
In univariate analyses, a positive association was found between BMI and annual percentage increase in medial tibial plateau area (P = 0.002) (Table 3). There was an inverse association between physical activity (P = 0.02) and baseline medial tibial plateau area (P < 0.001) and annual percentage increase in medial tibial plateau area (Table 3).
In multivariate analyses, gender explained 7.6% of the variance in the annual percentage change in medial tibial bone area; being male was significantly associated with a higher annual percentage increase of medial tibial plateau area (P = 0.001), after adjustment for age, BMI, physical activity, grade of osteophytes, grade of joint-space narrowing, and baseline tibial plateau bone area (Table 3). In these analyses, BMI explained 7.3% of the variance in annual percentage change in medial tibial bone area, and grade of medial joint-space narrowing explained 4.6% of variance in annual percentage change in medial tibial bone area. BMI and grade of medial joint-space narrowing were independently and positively associated with the annual percentage increase of medial tibial plateau area (P = 0.002 and P = 0.01, respectively). Baseline medial tibial plateau area remained inversely associated with the rate of medial tibial plateau area increase (P < 0.001) with baseline medial tibial plateau area explaining 19.3% of the variance in annual percentage change in medial tibial bone area. However, the association with physical activity was no longer statistically significant (P = 0.40). No significant associations were found between the assessed factors and annual percentage increase of lateral tibial plateau area in multivariate analyses (Table 3).
Adjustment for WOMAC score and SF-36 score did not affect these results (results not shown).
Discussion
In this 2-year longitudinal study, we found that there was a significant increase in tibial plateau bone area in symptomatic subjects with predominantly mild to moderate knee OA. The rate of increase in medial tibial plateau bone area was greater in men than in women. Higher BMI and higher baseline grade of medial joint-space narrowing were positively associated with an increased rate of enlargement of medial tibial plateau bone area, while a larger baseline medial tibial plateau bone area was inversely associated with the rate of increase of that area. None of the assessed factors were associated with the rate of increase of lateral tibial plateau bone area.
No previous longitudinal studies have examined the change in tibial plateau bone area in subjects with knee OA. In this study, we showed that subjects with symptomatic, mild to moderate knee OA had a significant increase in medial tibial plateau bone area over the 2 years of follow-up. In a cross-sectional study, Jones and colleagues [25] showed that grade 1 medial osteophytosis was associated with a 10 to 16% increase in both medial and lateral tibial bone areas after adjustment for age, sex, and BMI. Similar results have been shown in a study of hip OA, where men with hip OA had larger femoral neck size as assessed byMRI than healthy controls matched for age and sex [26]. In that study, femoral neck size was greater in the hip with higher OA grade [26]. In this study, we found that increasing grade of medial joint-space narrowing was associated with an increased rate of medial tibial plateau bone area increase. However, baseline tibial bone area was inversely associated with the rate of tibial bone area increase. This suggests that the rate of increase of tibial plateau bone area may be more rapid early in knee OA, when the tibial plateau bone area is smaller, and that as the tibial plateau bone area enlarges, the rate of increase slows down. All these associations were demonstrated in the medial compartment of tibial plateau, which is a more common site for knee OA than the lateral compartment [27]. The causes for the differences between the compartments are unknown.
In general, the factors affecting the change in tibial plateau bone area over the period of our observations were consistent with those factors previously described to be associated with tibial plateau bone size in cross-sectional studies [23,25,28,29]. However, in most studies tibial bone size has been measured using different tools [23,25,28-30]. A number of cross-sectional studies have shown that men have larger tibial bone size than women as measured by tibial plateau area [23], bone area at 4% of the tibial length [28], or articular surface area [29]. Although in our longitudinal study we found that BMI was associated with an increased rate of increase in tibial plateau bone area, Dacre and colleagues [30] did not show that BMI was significantly correlated with tibial plateau width measured on radiographs. These differences may be attributable to the different measures used for assessing tibial plateau bone size.
Our measurement of tibial plateau bone area byMRI is averaged on a two-dimensional projection of the tibia. This potential source of systematic error may become especially important when individuals of different body size are compared. The measurement has high reproducibility [22-25]. However, small positional changes in the longitudinal study may have resulted in overestimation of the measurement error and underestimation of longitudinal change. Our results cannot be simply explained by positing an increase in tibial bone size due to the presence of osteophytes, because adjustment for grade of osteophytes did not alter the findings. Indeed, the rate of tibial bone expansion was associated with an increased grade of joint-space narrowing, not of osteophytes. In this study, we measured only bone size and not bone mineral density, which may be important in the initiation and progression of OA [31-33]. It has been well known that normal bone metabolism depends on the presence of vitamin D. Low intake and low serum levels of vitamin D have been shown to be associated with an increased risk for progression of knee OA [34], or increased incidence of radiographically identified hip OA characterized by joint-space narrowing [35]. However, we did not take serum levels or intake of vitamin D into account in this study.
Here we have shown an expansion in medial tibial plateau bone over 2 years in subjects with mild to moderate knee OA. Consistent with this finding, previous human studies have shown changes in density and architecture of the subchondral bone in established OA [3-7]. Animal models support these findings [8-13]. In a guinea pig model of OA, trabecular remodelling was detected deep within the femoral head when only mild cartilage abnormalities were present [8,9]. In the cruciate-deficient-dog model of OA, by the time the dogs were 3 months old the articular cartilage in the unstable knee showed both the histologic changes typical of early OA and loss of trabecular bone [11]. These changes progressed over a period of 18 to 54 months, with subchondral sclerosis and increased thickness of the subchondral plate. By then, the differences in trabecular thickness and in surface-to-volume ratio were greater than at 3 months [11]. These changes affect the mechanical properties of the subchondral bone and may affect the initiation and progression of cartilage degeneration [5,13,14]. However, at this stage it is unclear which abnormality occurs first. In addition, factors, such as knee adduction moment, which have been shown to affect the progression of knee OA, are also associated with increased medial tibial plateau area [36]. This raises the possibility that changes in tibial bone size may mediate the effect of biomechanical factors on the pathogenesis of knee OA. Similarly, it may be that the effect of obesity on risk of knee OA is mediated via an effect on tibial bone.
We have previously reported that bone size was an independent predictor of knee cartilage volume [23]. However, average tibial plateau bone area was not associated with the rate of tibial cartilage loss [16]. Subjects with knee OA experience a loss of tibial cartilage volume, independent of tibial bone size, and an increase of tibial plateau bone area in the knee. This would tend to result in attenuation of articular cartilage over time. This may result in biomechanical changes at the knee as OA progresses, which may further contribute to the pathogenetic process. The mechanism for the bone changes is not known but may be a combination of biomechanical and systemic factors. For example, higher levels of transforming growth factor β and of insulin-like growth factors 1 and 2 may provide a 'bone-forming' stimulus in subjects with OA [37-39], which could explain the larger bone size in OA.
Conclusion
In subjects with established knee OA, tibial plateau bone area increases over time. The role of subchondral bone change in the pathogenesis of knee OA will need to be determined but may be one explanation for the mechanism of action of risk factors such as BMI on knee OA.
Abbreviations
BMI = body mass index; MRI = magnetic resonance imaging/image; OA = osteoarthritis; SF-36 = 36-Item Short-Form Health Survey; WOMAC = Western Ontario and McMaster Universities Osteoarthritis Index.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
All authors participated in the design of the study. YW carried out the measurement of the tibial plateau bone areas, performed the statistical analysis, and drafted the manuscript. AW and FC reviewed the manuscript. All authors read and approved the final manuscript.
Figures and Tables
Figure 1 Axial T1-weighted fat-saturated 3DMRI image showing measurement of tibial plateau bone area. The area of medial (Roi 2) and lateral (Roi 1) tibial plateau bone is measured manually on the first image that shows both tibial cartilage and subchondral bone (left), and on the next distal image (right). An average of the two areas is used as an estimate of the tibial plateau bone area. MRI, magnetic resonance image; Roi, region of interest.
Table 1 Characteristics of the study population
Total (n = 126) Men (n = 58) Women (n = 68) P
a
Age, years 63.6 ± 10.1 64.2 ± 10.0 63.2 ± 10.3 0.58
Height, cm 167.7 ± 9.1 174.6 ± 6.8 161.8 ± 6.1 <0.001
Weight, kg 81.4 ± 15.5 85.1 ± 14.9 78.3 ± 15.3 0.01
BMI, kg/m2 28.9 ± 5.1 27.8 ± 3.8 29.9 ± 5.8 0.02
WOMAC score 436 ± 224 411 ± 214 457 ± 232 0.25
SF-36 score 98 ± 7 98 ± 6 99 ± 8 0.25
Physical activity 6.2 ± 1.8 6.7 ± 1.7 5.8 ± 1.8 0.01
Time between scans, years 1.95 ± 0.21 1.94 ± 0.19 1.96 ± 0.23 0.58
OA gradeb
Grade of medial osteophytes
<2 107 50 57 0.50
≥ 2 18 7 11 0.35
Grade of medial JSN
<2 90 35 55 0.04
≥ 2 35 22 13 0.13
Grade of lateral osteophytes
<2 103 48 55 0.49
≥ 2 22 9 13 0.39
Grade of lateral JSN
<2 116 55 61 0.58
≥ 2 9 2 7 0.10
Tibial plateau area, mm2
Medial 2054.6 ± 363.9 2331.1 ± 304.3 1860.2 ± 263.2 <0.001
Lateral 1407.2 ± 256.7 1533.4 ± 244.9 1328.2 ± 232.8 <0.001
Except where indicated otherwise, values are means ± standard deviations. aMen vs women, by independent t-tests or chi-square test, as appropriate. bValues are the number of subjects in each grade category. One radiograph was lost after the study commenced. BMI, body mass index; JSN, joint-space narrowing.; MRI, magnetic resonance imaging; OA, osteoarthritis; SF-36, 36-Item Short-Form Health Survey; WOMAC, Western Ontario and McMaster Universities Osteoarthritis Index.
Table 2 Annual change and rate in tibial plateau bone area
Crude analysis Adjusted analysis
Annual change Total P Men Women Pa Men Women Pb
Increase, mm2
Medial 36.8 ± 70.7 <0.001 38.3 ± 62.6 35.7 ± 76.1 0.85 70.8 ± 12.3 13.5 ± 9.6 0.002
Lateral 19.1 ± 60.1 0.004 16.5 ± 83.9 20.4 ± 45.1 0.77 17.2 ± 13.3 20.1 ± 8.5 0.87
Change %
Medial 2.2 ± 6.9 <0.001 1.6 ± 2.8 2.6 ± 8.7 0.49 5.4 ± 1.1 0.004 ± 0.9 0.001
Lateral 1.5 ± 4.3 0.004 1.3 ± 5.8 1.6 ± 3.4 0.79 1.6 ± 0.9 1.5 ± 0.6 0.95
Values are mean ± standard deviations, or means ± standard errors for adjusted analysis. aMen vs women. bMen vs women, adjusted for age, body mass index, physical activity, grade of osteophytes, grade of joint-space narrowing, and baseline tibial plateau bone area.
Table 3 Factors affecting annual percentage change in tibial plateau bone area
Tibial plateau area factorsa Univariate analysis Regression coefficient (95% CI) P Multivariate analysisb Regression coefficient (95% CI) P
Medial tibial plateau area
Age -0.10 (-0.23, 0.03) 0.13 -0.07 (-0.18, 0.05) 0.26
Gender 0.94 (-1.78, 3.66) 0.49 -5.42 (-8.67, -2.18) 0.001
BMI 0.43 (0.17, 0.70) 0.002 0.42 (0.16, 0.67) 0.002
Physical activity -0.85 (-1.58, -0.11) 0.02 -0.30 (-1.00, 0.40) 0.40
Grade of medial osteophytes 0.02 (-1.78, 1.83) 0.98 0.27 (-1.53, 2.07) 0.77
Grade of medial JSN 0.35 (-1.10, 1.80) 0.63 1.88 (0.43, 3.33) 0.01
Baseline medial tibial area -0.006 (-0.010, -0.003) <0.001 -0.012 (-0.017, -0.008) <0.001
Lateral tibial plateau area
Age -0.02 (-0.10, 0.07) 0.67 -0.01 (-0.11, 0.08) 0.80
Gender 0.25 (-1.62, 2.12) 0.79 -0.08 (-2.51, 2.36) 0.95
BMI -0.02 (-0.19, 0.15) 0.82 -0.04 (-0.23, 0.16) 0.72
Physical activity -0.05 (-0.57, 0.48) 0.85 -0.16 (-0.74, 0.41) 0.57
Grade of lateral osteophytes -0.63 (-1.68, 0.42) 0.24 -0.18 (-1.70, 1.35) 0.82
Grade of lateral JSN -0.87 (-2.35, 0.61) 0.25 -0.42 (-2.31, 1.48) 0.66
Baseline lateral tibial area -0.003 (-0.007, 0.001) 0.11 -0.002 (-0.008, 0.003) 0.38
aFactors were defined as follows: age = percentage change per 1-year increase in age; gender = men compared with women; BMI = percentage change per unit increase in body mass index; physical activity = percentage change per unit increase in physical activity level; grade of osteophytes = percentage change per increase 1 in grade of osteophytes; grade of JSN=percentage change per increase 1 in grade of joint-space narrowing; baseline tibial area = percentage change per 1 mm2 increase in baseline tibial plateau bone area. bIncludes age, gender, BMI, physical activity, grade of osteophytes, grade of JSN, and baseline tibial plateau bone area in regression equation. BMI, body mass index; JSN, joint-space narrowing.
==== Refs
Felson DT Osteoarthritis Rheum Dis Clin North Am 1990 16 499 512 2217954
Burr DB The importance of subchondral bone in osteoarthrosis Curr Opin Rheumatol 1998 10 256 262 9608330
Matsui H Shimizu M Tsuji H Cartilage and subchondral bone interaction in osteoarthrosis of human knee joint: a histological and histomorphometric study Microsc Res Tech 1997 37 333 342 9185155 10.1002/(SICI)1097-0029(19970515)37:4<333::AID-JEMT8>3.0.CO;2-L
Buckland-Wright C Subchondral bone changes in hand and knee osteoarthritis detected by radiography Osteoarthritis Cartilage 2004 S10 19 14698636 10.1016/j.joca.2003.09.007
Radin EL Rose RM Role of subchondral bone in the initiation and progression of cartilage damage Clin Orthop Relat Res 1986 213 34 40 3780104
Grynpas MD Alpert B Katz I Lieberman I Pritzker KP Subchondral bone in osteoarthritis Calcif Tissue Int 1991 49 20 26 1893292
Radin EL Paul IL Tolkoff MJ Subchondral bone changes in patients with early degenerative joint disease Arthritis Rheum 1970 13 400 405 4246869
Layton MW Goldstein SA Goulet RW Feldkamp LA Kubinski DJ Bole GG Examination of subchondral bone architecture in experimental osteoarthritis by microscopic computed axial tomography Arthritis Rheum 1988 31 1400 1405 3190784
Dedrick DK Goulet R Huston L Goldstein SA Bole GG Early bone changes in experimental osteoarthritis using microscopic computed tomography J Rheumatol Suppl 1991 27 44 45 2027128
Brandt KD Myers SL Burr D Albrecht M Osteoarthritic changes in canine articular cartilage, subchondral bone, and synovium fifty-four months after transection of the anterior cruciate ligament Arthritis Rheum 1991 34 1560 1570 1747141
Dedrick DK Goldstein SA Brandt KD O'Connor BL Goulet RW Albrecht M A longitudinal study of subchondral plate and trabecular bone in cruciate-deficient dogs with osteoarthritis followed up for 54 months Arthritis Rheum 1993 36 1460 1467 8216405
Brown TD Radin EL Martin RB Burr DB Finite element studies of some juxtarticular stress changes due to localized subchondral stiffening J Biomech 1984 17 11 24 6715384 10.1016/0021-9290(84)90075-7
Hayami T Pickarski M Wesolowski GA McLane J Bone A Destefano J Rodan GA Duong le T The role of subchondral bone remodeling in osteoarthritis: reduction of cartilage degeneration and prevention of osteophyte formation by alendronate in the rat anterior cruciate ligament transection model Arthritis Rheum 2004 50 1193 1206 15077302 10.1002/art.20124
Radin EL Paul IL Rose RM Role of mechanical factors in pathogenesis of primary osteoarthritis Lancet 1972 1 519 522 4110024 10.1016/S0140-6736(72)90179-1
Dieppe P Subchondral bone should be the main target for the treatment of pain and disease progression in osteoarthritis Osteoarthritis Cartilage 1999 7 325 326 10329316 10.1053/joca.1998.0182
Wluka AE Stuckey S Snaddon J Cicuttini FM The determinants of change in tibial cartilage volume in osteoarthritic knees Arthritis Rheum 2002 46 2065 2072 12209510 10.1002/art.10460
Altman R Asch E Bloch D Bole G Borenstein D Brandt K Christy W Cooke TD Greenwald R Hochberg M Development of criteria for the classification and reporting of osteoarthritis. Classification of osteoarthritis of the knee. Diagnostic and Therapeutic Criteria Committee of the American Rheumatism Association Arthritis Rheum 1986 29 1039 1049 3741515
Ware JJ Snow K Kosinski M Grandek B SF-36 Health Survey: Manual and Interpretation Guide 1993 Boston: The Health Institute, New England Medical Center
Bellamy N Buchanan WW Goldsmith CH Campbell J Stitt LW Validation study of WOMAC: a health status instrument for measuring clinically important patient relevant outcomes to antirheumatic drug therapy in patients with osteoarthritis of the hip or knee J Rheumatol 1988 15 1833 1840 3068365
Spector TD Harris PA Hart DJ Cicuttini FM Nandra D Etherington J Wolman RL Doyle DV Risk of osteoarthritis associated with long-term weight-bearing sports: a radiologic survey of the hips and knees in female ex-athletes and population controls Arthritis Rheum 1996 39 988 995 8651993
Altman RD Hochberg M Murphy WA JrWolfe F Lequesne M Atlas of individual radiographic features in osteoarthritis Osteoarthritis Cartilage 1995 3 70 8581752
Wluka AE Davis SR Bailey M Stuckey SL Cicuttini FM Users of oestrogen replacement therapy have more knee cartilage than non-users Ann Rheum Dis 2001 60 332 336 11247861 10.1136/ard.60.4.332
Cicuttini F Forbes A Morris K Darling S Bailey M Stuckey S Gender differences in knee cartilage volume as measured by magnetic resonance imaging Osteoarthritis Cartilage 1999 7 265 271 10329301 10.1053/joca.1998.0200
Jones G Glisson M Hynes K Cicuttini F Sex and site differences in cartilage development: a possible explanation for variations in knee osteoarthritis in later life Arthritis Rheum 2000 43 2543 2549 11083279 10.1002/1529-0131(200011)43:11<2543::AID-ANR23>3.0.CO;2-K
Jones G Ding C Scott F Glisson M Cicuttini F Early radiographic osteoarthritis is associated with substantial changes in cartilage volume and tibial bone surface area in both males and females Osteoarthritis Cartilage 2004 12 169 174 14723876 10.1016/j.joca.2003.08.010
Arokoski JP Arokoski MH Jurvelin JS Helminen HJ Niemitukia LH Kroger H Increased bone mineral content and bone size in the femoral neck of men with hip osteoarthritis Ann Rheum Dis 2002 61 145 150 11796401 10.1136/ard.61.2.145
Ledingham J Regan M Jones A Doherty M Radiographic patterns and associations of osteoarthritis of the knee in patients referred to hospital Ann Rheum Dis 1993 52 520 526 8346979
Russo CR Lauretani F Bandinelli S Bartali B Di Iorio A Volpato S Guralnik JM Harris T Ferrucci L Aging bone in men and women: beyond changes in bone mineral density Osteoporos Int 2003 14 531 538 12827220 10.1007/s00198-002-1322-y
Faber SC Eckstein F Lukasz S Muhlbauer R Hohe J Englmeier KH Reiser M Gender differences in knee joint cartilage thickness, volume and articular surface areas: assessment with quantitative three-dimensional MR imaging Skeletal Radiol 2001 30 144 150 11357452 10.1007/s002560000320
Dacre JE Scott DL Da Silva JA Welsh G Huskisson EC Joint space in radiologically normal knees Br J Rheumatol 1991 30 426 428 1747696
Zhang Y Hannan MT Chaisson CE McAlindon TE Evans SR Aliabadi P Levy D Felson DT Bone mineral density and risk of incident and progressive radiographic knee osteoarthritis in women: the Framingham Study J Rheumatol 2000 27 1032 1037 10782833
Hart DJ Cronin C Daniels M Worthy T Doyle DV Spector TD The relationship of bone density and fracture to incident and progressive radiographic osteoarthritis of the knee: the Chingford Study Arthritis Rheum 2002 46 92 99 11817613 10.1002/1529-0131(200201)46:1<92::AID-ART10057>3.0.CO;2-#
Bruyere O Dardenne C Lejeune E Zegels B Pahaut A Richy F Seidel L Ethgen O Henrotin Y Reginster JY Subchondral tibial bone mineral density predicts future joint space narrowing at the medial femoro-tibial compartment in patients with knee osteoarthritis Bone 2003 32 541 545 12753870 10.1016/S8756-3282(03)00059-0
McAlindon TE Felson DT Zhang Y Hannan MT Aliabadi P Weissman B Rush D Wilson PW Jacques P Relation of dietary intake and serum levels of vitamin D to progression of osteoarthritis of the knee among participants in the Framingham Study Ann Intern Med 1996 125 353 359 8702085
Lane NE Gore LR Cummings SR Hochberg MC Scott JC Williams EN Nevitt MC Serum vitamin D levels and incident changes of radiographic hip osteoarthritis: a longitudinal study. Study of Osteoporotic Fractures Research Group Arthritis Rheum 1999 42 854 860 10323440 10.1002/1529-0131(199905)42:5<854::AID-ANR3>3.0.CO;2-I
Jackson BD Teichtahl AJ Morris ME Wluka AE Davis SR Cicuttini FM The effect of the knee adduction moment on tibial cartilage volume and bone size in healthy women Rheumatology (Oxford) 2004 43 311 314 14679292
Uchino M Izumi T Tominaga T Wakita R Minehara H Sekiguchi M Itoman M Growth factor expression in the osteophytes of the human femoral head in osteoarthritis Clin Orthop Relat Res 2000 377 119 125 10943193 10.1097/00003086-200008000-00017
Pujol JP TGF-beta and osteoarthritis: in vivo veritas? Osteoarthritis Cartilage 1999 7 439 440 10489315 10.1053/joca.1999.0248
Dequeker J Mohan S Finkelman RD Aerssens J Baylink DJ Generalized osteoarthritis associated with increased insulin-like growth factor types I and II and transforming growth factor beta in cortical bone from the iliac crest. Possible mechanism of increased bone density and protection against osteoporosis Arthritis Rheum 1993 36 1702 1708 8250990
| 15899054 | PMC1174962 | CC BY | 2021-01-04 16:02:37 | no | Arthritis Res Ther. 2005 Mar 31; 7(3):R687-R693 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1726 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17271589905510.1186/ar1727Research ArticleThe active metabolite of leflunomide, A77 1726, interferes with dendritic cell function Kirsch Bernhard M [email protected] Maximilian [email protected] Karl [email protected] Johannes [email protected] Josef S [email protected] Bruno [email protected] Thomas M [email protected]örl Walter H [email protected] Gerhard J [email protected]äemann Marcus D [email protected] Department of Internal Medicine III/Clinical Divisions of Nephrology and Dialysis, Medical University of Vienna, Vienna, Austria2 Department of Internal Medicine III/Clinical Divisions of Endocrinology and Metabolism, Medical University of Vienna, Vienna, Austria3 Ludwig Boltzmann Institute of Rheumatology, Vienna, Austria4 Department of Internal Medicine III/Clinical Division of Rheumatology, Medical University of Vienna, Vienna, Austria5 CeMM – Center of Molecular Medicine, Austrian Academy of Sciences, Vienna, Austria6 Institute of Immunology, Medical University of Vienna, Vienna, Austria2005 1 4 2005 7 3 R694 R703 16 12 2004 21 1 2005 23 2 2005 1 3 2005 Copyright © 2005 Kirsch et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Leflunomide, a potent disease-modifying antirheumatic drug used in the treatment of rheumatoid arthritis (RA), exhibits anti-inflammatory, antiproliferative and immunosuppressive effects. Although most of the beneficial effects of leflunomide have been attributed to its antimetabolite activity, mainly in T cells, other targets accounting for its potency might still exist. Because of mounting evidence for a prominent role of dendritic cells (DCs) in the initiation and maintenance of the immune response in RA, we analyzed the effect of the active metabolite of leflunomide (A77 1726; LEF-M) on phenotype and function of human myleloid DCs at several stages in their life cycle. Importantly, DCs differentiated in the presence of LEF-M exhibited an altered phenotype, with largely reduced surface expression of the critical co-stimulatory molecules CD40 and CD80. Furthermore, treatment of DCs during the differentiation or maturation phase with LEF-M aborted successful DC maturation. Exogenous addition of uridine revealed that DC modulation by LEF-M was independent of its proposed ability as an antimetabolite. In addition, the ability of DCs to initiate T-cell proliferation and to produce the proinflammatory cytokines IL-12 and tumour necrosis factor-α was markedly impaired by LEF-M treatment. As a molecular mechanism, transactivation of nuclear factor-κB, an transcription factor essential for proper DC function, was completely suppressed in DCs treated with LEF-M. These data indicate that interference with several aspects of DC function could significantly contribute to the beneficial effects of leflunomide in inflammatory diseases, including RA.
==== Body
Introduction
Dendritic cells (DCs) are the most potent antigen-presenting cells in the immune system [1,2]. They represent a heterogeneous population of bone-marrow-derived cells located in lymphoid as well as in nonlymphoid organs. In peripheral tissues these antigen-presenting cells are immature and are functionally equipped to capture and process antigens. DCs are activated by pathogen-associated microbial patterns such as lipopolysaccharide (LPS) or by proinflammatory cytokines such as tumour necrosis factor (TNF)-α, and via the interaction of CD40 with its ligand (CD154), which is expressed on activated T cells [3]. Mature DCs possess optimal immunostimulatory properties because of maximal expression of their antigen-presenting and co-stimulatory molecules (i.e. CD40, CD80 and CD86) and their increased production of proinflammatory cytokines, including IL-12 and TNF-α. In contrast to the central role played by mature DCs in the initiation of primary immune responses, immature DCs stimulate T-cell responses only weakly or they may even induce tolerance to potential autoantigens [4].
Pharmacological modulation of DC activation has been demonstrated to prevent disease progression in several T-cell-mediated diseases [5], and it may therefore represent a promising approach to specific treatment of immunological disorders [6,7]. Notably, corticosteroids and another well known antirheumatic drug, namely gold thiomalate, significantly inhibit DC function, which may contribute to their clinical effectiveness [8,9].
Leflunomide is a novel disease-modifying antirheumatic drug that exerts its effects after metabolic opening of the isoxazole ring via its active metabolite A77 1726 (LEF-M). Its major target is supposed to be dihydro-orotate-dehydrogenase (DHODH) [10], which is a key enzyme in de novo pyrimidine synthesis. Leflunomide reversibly inhibits DHODH activity with subsequent depletion of nucleotides, leading to cell cycle arrest in proliferating lymphocytes [11]. This effect can be reversed to a certain degree by supplying the product of DHODH activity (i.e. uridine) to target cells. Other targets of LEF-M are tyrosine kinases such as Lck or JAK3 in activated T and B cells [12]. Immunosuppressive effects of leflunomide have been described including, inhibition of T cells and antibody production [13]. Furthermore, it was demonstrated that leflunomide blocks activation of nuclear factor-κB (NF-κB), which is a central proinflammatory transcription factor in several cell lines [14], and impairs transendothelial migration of peripheral blood mononuclear cells [15]. Apart from its well established beneficial effects in the treatment of rheumatoid arthritis (RA) [16,17], leflunomide is also effective in treatment against chronic allograft rejection [18,19].
DCs were postulated to play an important role in RA pathogenesis because they may perpetuate the disease by presenting self-antigen(s) [20,21]. Thus, DCs could represent an interesting target for dampening the disease process in RA. Moreover, DCs also play a fundamental role in allograft rejection [22].
Because the effect of leflunomide on DC function has not yet been investigated, we analyzed the influence of leflunomide on the complete DC life cycle in vitro. We found that LEF-M potently altered the phenotype and function of DCs, independent of its well known antimetabolite activity, revealing a novel immunomodulatory activity of this agent with potential clinical implications for the treatment of RA and other immune cell mediated disorders.
Materials and methods
Media and reagents
RPMI 1640 (GIBCO BRL, Grand Island, NY, USA) supplemented with 2 mmol/l L-glutamine, 100 μg/ml streptomycin, 100 U/ml penicillin and 10% foetal calf serum (FCS; Hyclone, Logan, UT, USA) was used as culture medium. LPS (Escherichia coli 0111:B4) and uridine were purchased from Sigma Chemie GmbH Co. (Deisenhofen, Germany). Recombinant human (rh) granulocyte–macrophage colony-stimulating factor (GM-CSF) was obtained from Schering-Plough (Kenilworth, NJ, USA) and rh-IL-4 was from Strathmann Biotech GmbH (Hannover, Germany). Plasma concentrations in RA patients of A77 1726 (the active metabolite of leflunomide) achieved with a leflunomide maintenance dose of 20 mg/day are 46 ± 31 μg/ml (approximately 150 ± 100 μmol/l [23]). Therefore, we chose concentrations from 75 to 150 μmol/l of A77 1726 (kindly provided by Aventis, Strasbourg, France) for DC treatment. A77 1726 is referred to as 'LEF-M' throughout the report. In some experiments uridine was added to test the reversibility of the observed effects of LEF-M.
Cell preparation and culture
Peripheral blood mononuclear cells were obtained from buffy coats of healthy blood donors (courtesy of the Austrian Red Cross) by density gradient centrifugation over Ficoll-Paque PLUS (Amersham Biosciences, Uppsala, Sweden). For isolation of monocytes, peripheral blood mononuclear cells were depleted of T cells by sheep erythrocyte-rosetting overnight.
Monocytes (>85% CD14+) were cultured in six-well plates (Costar, Cambridge, MA, USA) at a cell density of 5 × 105 cells/ml in RPMI 1640/10% FCS medium in a humidified atmosphere containing 5% carbon dioxide. For induction of DC differentiation, the culture medium was supplemented for 5 days with 50 ng/ml rh-GM-CSF and 10 ng/ml rh-IL-4. For initation of maturation, LPS (100 ng/ml) was added for an additional 48 hours. For the DC differentiation and maturation experiments, different concentrations of LEF-M, or medium as control, were added either at the beginning of the culture or 6 hours before the addition of LPS. Cell viability was assessed by staining with propidium iodide (PI; Sigma, Saint Louis, MO, USA) and subsequent flow cytometric analysis of the cells.
Surface marker expression
For evaluation of surface marker expression, cells (50 μl at 5 × 106 cells/ml) were incubated with fluorescein isothiocyanate (FITC)-conjugated or phycoerythrin (PE)-conjugated mAbs for 45 min at 4°C. For control purposes, nonbinding isotype-matched FITC-conjugated and PE-conjugated mouse IgG (An der Grub, Kaumberg, Austria) were employed. After extensive washing cells were analyzed on a COULTER EPICS XL-MLC flowcytometer (Beckman Coulter, Fullerton, CA, USA) using EXPO32 software. All measurements were done using a three-colour setup, which was established using standard compensation procedures. FITC-labelled mAbs to CD1a (IgG1; clone HI149), CD14 (IgG2b; clone MΦP9), CD83 (IgG1; clone HB15e) and HLA-DR (IgG2a; L243), and R-PE-labelled mAbs to CD80 (IgG1; L307.4), CD86 (IgG2b; clone IT2.2) and CD206 (mannose receptor; IgG1; clone 19.2) were obtained from Becton Dickinson (San Diego, CA, USA). FITC-conjugated anti-CD40 (IgG1; clone LOB7/6) was purchased from Serotec (Oxford, UK). R-PE-labelled anti-major histocompatibility complex (MHC) class I antibody (IgG2a; clone 3F10) was obtained from Ancell (Bayport, MN, USA).
Morphological cell analysis
Microscopy was performed in parallel to all other analyses to assess cell morphology by using a light optical microscope (Olympus Corporation, Tokyo, Japan).
Assessment of T-cell stimulatory capability
Stimulator cells were irradiated (3000 rad, 137Cs source) and added at increasing cell numbers to 1 × 105 allogeneic T cells in 96-well culture plates in RPMI 1640 medium supplemented with 10% FCS (total volume 200 μl/well). After 4–5 days, cells were pulsed with 1 μCi [3H]thymidine (ICN Pharmaceuticals, Irvine, CA, USA). After another 18 hours the cells were harvested on glass-fibre filters (Packard, Meriden, CT, USA) and DNA-associated radioactivity was determined using a microplate scintillation counter (Packard, Meriden, CT, USA). DNA synthesis was expressed as mean counts/min of triplicate cultures.
Measurement of cytokine production
DCs were differentiated and subsequently activated (100 ng/ml LPS) in the presence or absence of different concentrations of LEF-M. Cell-free supernatants were harvested 48 hours after cell activation. Cytokines were measured by sandwich enzyme-linked immunosorbent assays using matched pair antibodies. Capture as well as detection antibodies to human IL-12p40 were obtained from R&D Systems (Minneapolis, MN, USA). Antibodies to human TNF-α were from PharMingen (San Diego, CA, USA). Standards consisted of human recombinant material from R&D Systems. Assays were set up in duplicate and were performed in accordance with recommendations from the manufacturers. The lower limit of detection was 20 pg/ml for all cytokines.
Analysis of nuclear factor-κB activation
NF-κB activation was assessed using an electrophoretic mobility shift assay (EMSA).
Nuclear extracts from DCs were prepared as described perviously [24]. Oligonucleotides resembling the consensus binding site for NF-κB (5'-AGTTGAGGGGACTTTCCCAGGC-3') and activator protein-1 (5'-CGCTTGATGACTCAGCCGGAA-3') were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). The double-stranded oligonucleotides used in all experiments were end-labelled using T4 polynucleotide kinase and [γ-32P]-ATP. After labelling, 5 μg nuclear extract was incubated with 120,000 counts/min labelled probe in the presence of 3 μg poly(dI-DCs) at room temperature for 30 min. This mixture was separated on a 6% polyacrylamide gel in Tris/glycine/EDTA buffer at pH 8.5. Control experiments were performed as described previously [25]. The specificity of NF-κB binding was proven using excess, unlabelled NF-κB probe that competed successfully for NF-κB binding, whereas an unrelated competitor (activator protein-1 oligonucleotide) did not (data not shown).
Statistical analysis
Comparisons were performed using two-tailed paired Student's t-tests. P < 0.05 was considered statistically significant.
Results
LEF-M impairs differentiation of monocyte-derived dendritic cells
In the first set of experiments we sought to determine whether leflunomide influences the differentation of freshly isolated monocytes into immature DCs. Therefore, we added GM-CSF and IL-4 to freshly isolated monocytes for 5 days to differentiate them to immature DCs in the presence or absence of LEF-M. Subsequently, we assessed surface marker expression using fluorescence-activated cell sorting analysis and found profound phenotypical differences between these differentiated cells. In the absence of LEF-M we found the typical immature DC phenotype, including high levels of MHC class II and high levels of CD1a, and a distinct profile of co-stimulatory molecules (Fig. 1a); neither the monocyte lineage marker CD14 nor the typical DC maturation marker CD83 was expressed. In contrast, LEF-M-treated DCs exhibited a different phenotype, with profoundly suppressed surface expression of CD40 and CD80 (Fig. 1a). Importantly, LEF-M markedly prevented the induction of the Langerhans cell-associated marker CD1a, which is a marker of successful DC differentiation, whereas expression of CD86, mannose receptor and MHC class I and II molecules remained unaffected by LEF-M (Fig. 1a). Of note, LEF-M did not interfere with the characteristic disappearance of the monocyte marker CD14. Furthermore, we found no difference in cell viability between LEF-M-treated and control cells, as determined by PI staining. As calculated from eight independent experiments, the percentage PI positivity was 13.8 ± 4.5% in untreated cells versus 14.3 ± 1.4% in cells treated with 150 μmol/l LEF-M (mean percentage ± standard error of the mean).
Addition of uridine did not rescue DC differentiation from the effects of LEF-M, indicating that inhibition of DHODH did not underlie the observed effects (Fig. 1c).
Finally, LEF-M-modulated DCs were assessed for maturation sensitivity. Although immature control DCs exposed to LPS exhibited typical features of mature DCs (Fig. 1b), including upregulation of CD40, CD80, CD86, MHC class I and II and neo-expression of CD83, the maturation program was arrested in cells that were differentiated and subsequently activated in the presence of LEF-M. As shown in Fig. 1b, LEF-M-treated DCs, despite LPS stimulation, continued to exhibit profoundly inhibited expression of CD40 and CD80, whereas LEF-M only marginally affected CD86 and MHC expression. Importantly, CD83 expression was abolished in LEF-M pretreated cells (Fig. 1b). Again, addition of uridine did not reverse the inhibitory effects of LEF-M on DC maturation (Fig. 1d). Again, the effects of LEF-M on DC phenotype were not simply a consequence of cellular cytotoxicity, as indicated by unchanged cell morphology and viability (percentage PI positivity was 9.0 ± 2.9% in untreated cells versus 15.3 ± 0.5% in cells treated with 150 μmol/l LEF-M; data expressed as mean percentage ± standard error of the mean, calculated from eight independent experiments).
LEF-M impaires cytokine production and the allostimulatory capacity of monocyte-derived dendritic cells
DCs are typically characterized by their ability to produce large amounts of predominantly T-cell modulatory cytokines [26]. Analyzing cytokine production of cells that were differentiated and subsequently maturated in the presence of LEF-M, we found dose-dependent inhibition of IL-12p40 and TNF-α and of IL-10 production (Fig. 2).
In addition to the observed distortion in DC phenotype after differentiation and maturation, we found profound impairment of the allo-stimulatory function of LEF-M pretreated DCs. As shown in Fig. 3a, immature control DCs exhibited poor stimulatory capacity of allogeneic T-cells. DCs differentiated in the presence of LEF-M were even less potent stimulators in the mixed leukocyte culture (Fig. 3a). LPS exposure dramatically increased the stimulatory capability of control DCs, but DCs differentiated in the presence of LEF-M and subsequently exposed to an activation stimulus were as ineffective as immature control DCs in supporting T-cell proliferation (Fig. 3b).
LEF-M interferes directly with maturation of dendritic cells
We then analyzed whether LEF-M affects DC maturation when the drug was added to immature DCs (i.e. after completion of DC differentiation). Although immature control DCs responded readily, with increased expression of co-stimulatory and antigen-presenting molecules, LEF-M markedly interfered with the activation-induced upregulation of CD40 and CD86 but not that of CD80 (Fig. 4a). Importantly, neo-expression of CD83 – an indicator of proper DC maturation [27] – was significantly impaired in LEF-M-treated DCs (Fig. 4a). A further striking feature of mature DCs is the development of prominent cell clusters a few hours after addition of the maturation stimulus. On analyzing LEF-M-treated DCs, we detected complete abrogation of this clustering response (Fig. 4b,c). Another typical hallmark of mature DCs is their exceptional T-cell stimulatory capacity. As shown in Fig. 5, mature DCs exhibited optimal T-cell stimulatory capability. In contrast, the presence LEF-M solely during the maturation period of already differentiated DCs abrogated their stimulatory capacity in a concentration-dependent manner.
Effect of LEF-M on nuclear factor-κB activation in dendritic cells
Activation of the transcription factor NF-κB is essential for DC function [28,29]. DCs readily respond to diverse stimuli such as microbial products, cytokines and tissue damage, all of which converge on the NF-κB pathway [30]. Our findings of an impaired DC function in LEF-M-treated cells prompted us to analyze the effect of LEF-M on the activation of this central transcription factor in DCs. As shown in Fig. 6, employment of electrophoretic mobility shift assays revealed a clear time-dependent increase in nuclear binding of the NF-κB consensus site upon LPS stimulation in DCs. The specificity of NF-κB binding was indicated by competition with unlabelled probe and an unrelated competitor (activator protein-1 oligonucleotide; data not shown). Strikingly, treatment of immature DCs with LEF-M profoundly suppressed nuclear translocation of NF-κB in LPS-stimulated DCs after both 40 and 70 min (Fig. 6).
Discussion
This study reveals a novel aspect of the immunomodulatory action of leflunomide, namely the profound interference of LEF-M (A77 1726) with DC function. Using human monocyte-derived DCs as a model system, we demonstrated that LEF-M disrupts differentiation of DCs from uncommitted monocytic precursor cells, resulting in maturation-insensitive DCs. Furthermore, we showed that the maturation process of uncommitted immature DCs was markedly impaired by LEF-M. The metabolite LEF-M differentially affected the expression of critical surface molecules, inhibited the production of proinflammatory cytokines and, at the functional level, profoundly impaired the T-cell stimulatory capacity of DCs. As a molecular basis for the ability of LEF-M to interfere with several aspects of DC function, the activation-driven nuclear transmigration of the essential transcription factor NF-κB was markedly impaired by LEF-M. These findings have substantial implications for our understanding of the effects of leflunomide as a disease-modifying antirheumatic drug, because the initiation of an immune response critically depends on proper DC function. Furthermore, interference with DC maturation and function could also be involved in the beneficial effects of leflunomide on chronic allograft rejection [18], which is not shared by most other currently used immunosuppressive drugs such as calcineurin inhibitors.
The observation that DCs could play a pivotal role in the formation and maintenance of joint inflammation in RA [31] was confirmed by the finding reported by Balanescu and coworkers [32] of a correlation between co-stimulatory molecule expression of synovial DCs and disease activity in RA patients. Moreover, mature DCs might be central in the development of perivascular aggregates in synovial inflammation areas, the formation of organized lymphoid structures, and in the perpetuation of inflammatory and erosive activity [20,21]. Although there is sufficient evidence for an impact of leflunomide on synoviocytes, chondrocytes and osteoclasts [33-36], our data suggest that the potent inhibition of DC function by LEF-M might contribute to the beneficial effects of leflunomide treatment in patients with RA.
Exposure of DCs to LEF-M led to an alteration in the surface marker profile. Our findings concerning the impact of LEF-M on critical co-stimulatory molecules might be especially important in RA because the expression level of co-stimulatory molecules on DCs correlates with disease activity in patients with RA [32]. Another important finding in the present study was the observed disruption by LEF-M of the DC differentiation process. Interestingly, neo-expression of CD1a – the classic Langerhans cell-associated marker – was strongly inhibited in LEF-M-treated DCs. This finding is accordance with observations of significant efficacy of leflunomide in psoriasis [37], in which CD1a is highly overexpressed in involved skin [38]. Importantly, CD14 – a classic monocyte/macrophage marker – was downregulated, indicating that LEF-M does not subvert the DC differentiation programme toward macrophages as has been shown for IL-6, IL-10 and corticosteroids [39,40].
A central observation in our study was the functional alteration of DCs differentiated in the presence of LEF-M; these cells exhibited a marked reduction in their T-cell stimulatory capacity upon activation. These data indicate that LEF-M, by blocking the differentiation of monocytic precursors into mature DCs, potentially impairs proper DC function and might therefore modulate immune responsiveness against potential autoantigens and other antigens. Our finding of markedly decreased production of TNF-α and IL-12 by LEF-M-treated DCs, in conjunction with insufficient co-stimulatory molecule expression of DCs, may be of interest for further DC studies with LEF-M, because recent reports demonstrated this phenotype to be potentially tolerogenic [41,42].
Interestingly, we found the effects of LEF-M on DCs to be mediated independent of its inhibition of DHODH. As shown for several other leflunomide-mediated effects on other cell types, such as osteoclasts in the RA joint [43], memory T-cell lines in an autoimmune encephalomyelitis model [44] and in articular chondrocytes [34], or on functional effects such as repression of viral replication [45,46], the inhibitory effects of LEF-M in the present study are clearly independent of pyrimidine synthesis.
The transcription factor NF-κB plays a decisive role in proper DC function. NF-κB translocation is essential to the ability of mature DC to present antigen to naïve T cells [28,29]. Recently reported data demonstrate that LEF-M inhibits TNF-α-induced NF-κB activation in several cell lines [14,47]. Interestingly, we found a profound suppression of NF-κB transactivation in activated DCs by LEF-M. These results are in accordance with our findings showing impaired expression of maturation markers and reduced allo-stimulatory capacity of leflunomide-treated DCs, because selective inhibition of NF-κB activity has been shown to impair maturation of DCs [48]. Our findings concerning cytokine production are also consistent with NF-κB inhibition, because the human IL-12 promoter contains crucial NF-κB binding sites and TNF-α production is also NF-κB dependent [49]. Although the mechanisms underlying this profound NF-κB inhibitory activity of LEF-M on DCs are currently unknown, it is tempting to speculate that leflunomide may interfere with phosphorylation/dephosphorylation events in the LPS-triggered signalling program. Apart from the possibility that LEF-M might directly induce the transcription of distinct IκB family members, LEF-M could also induce particular phosphatases to inhibit the IκB-inactivating kinase IKK. Furthermore, recent studies have shown that leflunomide acts at the level of IκB-α phosphorylation via interference with IKK-α activation, ultimately leading to defective IκB-α phosphorylaton. Although further studies are required to unravel the detailed molecular mechanisms of suppressed NF-κB transactivation in LEF-M-treated DCs, our findings indicate that NF-κB inhibition is a central feature of the molecular actions of LEF-M on DCs.
Importantly, the results from the present study were obtained with monocyte-derived DCs generated from healthy volunteers. Hence, further studies will be necessary to clarify the effects of LEF-M on peripheral and synovial DCs in experimental models of arthritis and on DCs obtained from RA patients. Nevertheless, our finding of DC inhibition induced by LEF-M reveals a novel view of the disease-modifying effects of this drug, which appear to act on both T cells and DCs. In fact, the involvement of DC–T cell interactions in the pathways leading to and perpetuating RA and the effects of inhibiting this process are supported by recent findings on the significant clinical effects of interference with CD80/86–CD28 co-stimulation [50].
Conclusion
The present study shows that monocyte-derived DCs are sensitive targets of LEF-M, possibly by inhibitory effects on NF-κB. DCs are affected by LEF-M at all major stages in their life cycle, ultimately leading to an impairment in DC function. In addition to a direct inhibitory action on specific T-cell responses, modulation of the immune system may therefore also be explained through the effects of leflunomide on DCs rendering these cells less able to support immunoinflammatory responses. Thus, the versatile role played by leflunomide as an immunomodulatory agent in vitro and in vivo is further supported by its effect on DCs. These findings reveal a novel mode of action of the active leflunomide metabolite during induction of cellular immune responses, which may contribute to the clinical effectiveness of leflunomide in diseases that involve exaggerated immune responsiveness.
Abbreviations
DC = dendritic cell; DHODH = dihydro-orotate-dehydrogenase; FCS = fetal calf serum; FITC = fluorescein isothiocyanate; GM-CSF = granulocyte–macrophage colony-stimulating factor; IL = interleukin; LEF-M = active metabolite of leflunomide; LPS = lipopolysaccharide; mAb = monoclonal antibody; MHC = major histocompatibility complex; NF-κB = nuclear factor-κB; PE = phycoerythrin; PI = propidium iodide; RA = rheumatoid arthritis; rh = recombinant human; TNF = tumour necrosis factor.
Authors' contributions
BK performed all flow cytometric and proliferation experiments, wrote the draft version of the manuscript and compiled the figures. MZ performed the uridine experiments. KS performed the electrophoretic mobility shift assays. JG analyzed the statistical data. JSS provided substantial input into the study design and helped in writing the manuscript. BW helped with statistical analysis and with finalizing the manuscript. WHH provided substantial input into the study design and helped with finalizing the manuscript. TMS was involved in all phases of the experimental process. GJZ performed the cytokine measurements. MDS designed the experiments, controlled all experimental steps and finalized the manuscript. All authors read and approved the final manuscipt.
Acknowledgements
We thank Bianca Weissenhorn and Margarethe Merio for expert technical assistance.
This study was supported in part by grants of the Austrian Jubilee Fund (ÖNB 10282; to MDS), the Austrian Science Fund (P16788-B13; to TMS) and the Center of Molecular Medicine, a basic research institute within the companies of the Austrian Academy of Sciences (to TMS and JSS).
Figures and Tables
Figure 1 LEF-M interferes with DC differentiation. (a) Monocytes were cultured for 5 days with granulocyte–macrophage colony-stimulating factor (GM-CSF; 50 ng/ml) plus IL-4 (10 ng/ml) in the absence or presence of 150 μmol/l of the active metabolite of leflunomide (LEF-M). Subsequently, surface marker expression was determined using fluorescence-activated cell sorting (FACS) analysis. Open profiles with dotted line represent staining pattern with an isotype control antibody, open profiles with fine line indicate the staining pattern of differentiated control dendritic cells (DCs) stained with the indicated mAbs, whereas solid grey profiles show staining of DCs differentiated in the presence of LEF-M. (b) Myeloid precursor cells differentiated in the presence of LEF-M are resistant to maturation. Cells were treated as described above and then stimulated with lipopolysaccharide (LPS; 100 ng/ml) for 48 hours. Open profiles with dotted line represent staining pattern with an isotype control antibody, open profiles with fine line indicate staining of activated control DCs, and solid grey profiles show staining of DCs differentiated in the presence of LEF-M and subsequently exposed to LPS. Data are representative of at least four independent experiments. (c,d) The effects of LEF-M on DC differentiation are independent of pyrimidine depletion. The respective change in mean flourescence intensity (MFI) are shown (c) after the differentiation phase for CD40 and CD1a and (d) after subsequent maturation with 100 ng/ml LPS for CD40 and CD83 with and without 50 μmol/l uridine. White bars represent control DCs, and black bars indicate LEF-M-treated cells. Shown are mean percentage control responses ± standard error of the mean, calculated from five to eight independent experiments. Student's t-tests were calculated for control versus LEF-M-treated DCs and for LEF-M-treated DCs versus without uridine addition, as indicated. *P < 0.05, **P < 0.01.
Figure 2 LEF-M abrogates cytokine production in DCs. Dendritic cells (DCs) were differentiated and subsequently activated (100 ng/ml lipopolysaccharide [LPS]) in the presence or absence of the indicated concentrations of the active metabolite of leflunomide (LEF-M). Cell-free supernatants were collected 48 hours after addition of LPS and then analyzed using enzyme-linked immunosorbent assay. Shown are mean percentage of control responses ± standard error of the mean for IL-12, tumour necrosis factor (TNF)-α and IL-10, calculated from at least 10 independent experiments. Student's t-tests were calculated for control DCs versus LEF-M-treated DCs. *P < 0.05, **P < 0.01. Mean cytokine levels (± standard deviation) in stimulated control cultures were 793 ± 343 pg/ml (IL-10), 23.6 ± 7.6 ng/ml (IL-12) and 2.9 ± 1.1 ng/ml (TNF-α).
Figure 3 DCs differentiated in the presence of LEF-M exhibit reduced T-cell stimulatory capacity. (a) Monocytes were cultured for 5 days with granulocyte–macrophage colony-stimulating factor (GM-CSF; 50 ng/ml) plus IL-4 (10 ng/ml) in the presence or absence of the indicated concentrations of LEF-M. Dendritic cells (DCs) differentiated in the presence of the active metabolite of leflunomide (LEF-M) are labelled 'LEF-M DCs' in the figure. The cells were extensively washed, irradiated (3000 rad) and subsequently co-cultured with 1 × 105 purified allogeneic T cells at the indicated ratios. (b) To determine maturation sensitivity, DCs differentiated in the presence or absence of LEF-M were exposed to 100 ng/ml lipopolysaccharide for an additional 48 hours. Then, the cells were employed as allogeneic stimulators, as described above. DNA synthesis was assessed at day 5. The standard deviation of the counts/min (cpm) for the respective triplicates was generally below 20%. Shown are the means of at least eight independent experiments.
Figure 4 Treatment with LEF-M during maturation of immature DCs leads to a differentially affected phenotype. Monocytes were cultured for 5 days with granulocyte–macrophage colony-stimulating factor (GM-CSF; 50 ng/ml) plus IL-4 (10 ng/ml). (a) On day 5 these immature dendritic cells (DCs; 5 × 105/ml) were activated with lipopolysaccharide (LPS; 100 ng/ml) in the absence or presence of 150 μmol/l of the active metabolite of leflunomide (LEF-M) for 48 hours. Surface marker expression was determined by fluorescence-activated cell sorting analysis. Open profiles with dotted line represent the staining pattern with an isotype control, open profiles with fine line indicate the staining pattern of DC exposed to LPS with the indicated monoclonal antibodies, and solid grey profiles show staining of DCs matured in the presence of LEF-M. The results shown are representative of five independent experiments. (b,c) Effect of LEF-M on maturation-associated clustering of DCs; immature DCs were stimulated with LPS in the (panel b) absence or (panel c) presence of 150 μmol/l LEF-M. After 8 hours of cultivation, cells were analyzed by inspecting photomicrographs obtained by light microscopy. Similar results were obtained in four additional experiments. MHC, major histocompatibility complex.
Figure 5 Functional impairment of DCs matured in the presence of LEF-M. Immature dendritic cells (iDCs) were exposed to lipopolysaccharide (LPS; 100 ng/ml) in the absence or presence of the indicated concentrations of the active metabolite of leflunomide (LEF-M). Then, the cells were extensively washed, irradiated (3000 rad) and subsequently co-cultured with 1 × 105 purified allogeneic T cells at the indicated ratios. DNA synthesis was assessed after 5 days and was measured in triplicate. The standard deviation of triplicates was generally below 20%. The data shown are expressed as mean counts/min (cpm) of four independent experiments.
Figure 6 LEF-M suppresses LPS-induced NF-κB activation in DCs. Immature dendritic cells (DCs) were cultured for 2 hours with or without the active metabolite of leflunomide (LEF-M; 150 μmol/l), followed by addition of lipopolysaccharide (LPS; 100 ng/ml) or medium as control. After 40 and 70 min total nucleoprotein was extracted and nuclear factor-κB (NF-κB) activity was detected using electrophoretic mobility shoft assay. Similar results were obtained in two independent experiments. (Nonspecific bands are labelled NS.)
==== Refs
Steinman RM The dendritic cell system and its role in immunogenicity Annu Rev Immunol 1991 9 271 296 1910679 10.1146/annurev.iy.09.040191.001415
Banchereau J Steinman RM Dendritic cells and the control of immunity Nature 1998 392 245 252 9521319 10.1038/32588
Banchereau J Bazan F Blanchard D Briere F Galizzi JP van Kooten C Liu YJ Rousset F Saeland S The CD40 antigen and its ligand Annu Rev Immunol 1994 12 881 922 7516669 10.1146/annurev.iy.12.040194.004313
Jonuleit H Schmitt E Steinbrink K Enk AH Dendritic cells as a tool to induce anergic and regulatory T cells Trends Immunol 2001 22 394 400 11429324 10.1016/S1471-4906(01)01952-4
Thompson AG Thomas R Induction of immune tolerance by dendritic cells: implications for preventative and therapeutic immunotherapy of autoimmune disease Immunol Cell Biol 2002 80 509 519 12406384 10.1046/j.1440-1711.2002.01114.x
Roelen DL Schuurhuis DH van den Boogaardt DE Koekkoek K van Miert PP van Schip JJ Laban S Rea D Melief CJ Offringa R Prolongation of skin graft survival by modulation of the alloimmune response with alternatively activated dendritic cells Transplantation 2003 76 1608 1615 14702533 10.1097/01.TP.0000086340.30817.BA
Menges M Rossner S Voigtlander C Schindler H Kukutsch NA Bogdan C Erb K Schuler G Lutz MB Repetitive injections of dendritic cells matured with tumor necrosis factor alpha induce antigen-specific protection of mice from autoimmunity J Exp Med 2002 195 15 21 11781361 10.1084/jem.20011341
Wang ZY Morinobu A Kawano S Saegusa J Wang B Kumagai S Gold sodium thiomalate suppresses the differentiation and function of human dendritic cells from peripheral blood monocytes Clin Exp Rheumatol 2002 20 683 688 12412200
Piemonti L Monti P Allavena P Sironi M Soldini L Leone BE Socci C Di Carlo V Glucocorticoids affect human dendritic cell differentiation and maturation J Immunol 1999 162 6473 6481 10352262
Davis JP Cain GA Pitts WJ Magolda RL Copeland RA The immunosuppressive metabolite of leflunomide is a potent inhibitor of human dihydroorotate dehydrogenase Biochemistry 1996 35 1270 1273 8573583 10.1021/bi952168g
Fairbanks LD Bofill M Ruckemann K Simmonds HA Importance of ribonucleotide availability to proliferating T-lymphocytes from healthy humans. Disproportionate expansion of pyrimidine pools and contrasting effects of de novo synthesis inhibitors J Biol Chem 1995 270 29682 29689 8530356 10.1074/jbc.270.50.29690
Xu X Williams JW Bremer EG Finnegan A Chong AS Inhibition of protein tyrosine phosphorylation in T cells by a novel immunosuppressive agent, leflunomide J Biol Chem 1995 270 12398 12403 7759480 10.1074/jbc.270.21.12398
Siemasko KF Chong AS Williams JW Bremer EG Finnegan A Regulation of B cell function by the immunosuppressive agent leflunomide Transplantation 1996 61 635 642 8610393 10.1097/00007890-199602270-00020
Manna SK Aggarwal BB Immunosuppressive leflunomide metabolite (A77 1726) blocks TNF-dependent nuclear factor-kappa B activation and gene expression J Immunol 1999 162 2095 2102 9973483
Grisar J Aringer M Koller MD Stummvoll GH Eselbock D Zwolfer B Steiner CW Zierhut B Wagner L Pietschmann P Leflunomide inhibits transendothelial migration of peripheral blood mononuclear cells Ann Rheum Dis 2004 63 1632 1637 15547088 10.1136/ard.2003.018440
Smolen JS Kalden JR Scott DL Rozman B Kvien TK Larsen A Loew-Friedrich I Oed C Rosenburg R Efficacy and safety of leflunomide compared with placebo and sulphasalazine in active rheumatoid arthritis: a double-blind, randomised, multicentre trial. European Leflunomide Study Group Lancet 1999 353 259 266 9929017 10.1016/S0140-6736(98)09403-3
Emery P Breedveld FC Lemmel EM Kaltwasser JP Dawes PT Gomor B Van Den Bosch F Nordstrom D Bjorneboe O Dahl R A comparison of the efficacy and safety of leflunomide and methotrexate for the treatment of rheumatoid arthritis Rheumatology (Oxford) 2000 39 655 665 10888712
Hardinger KL Wang CD Schnitzler MA Miller BW Jendrisak MD Shenoy S Lowell JA Brennan DC Prospective, pilot, open-label, short-term study of conversion to leflunomide reverses chronic renal allograft dysfunction Am J Transplant 2002 2 867 871 12392293 10.1034/j.1600-6143.2002.20909.x
Williams JW Mital D Chong A Kottayil A Millis M Longstreth J Huang W Brady L Jensik S Experiences with leflunomide in solid organ transplantation Transplantation 2002 73 358 366 11884931 10.1097/00007890-200202150-00008
Thomas R MacDonald KP Pettit AR Cavanagh LL Padmanabha J Zehntner S Dendritic cells and the pathogenesis of rheumatoid arthritis J Leukoc Biol 1999 66 286 292 10449169
Santiago-Schwarz F Anand P Liu S Carsons SE Dendritic cells (DCs) in rheumatoid arthritis (RA): progenitor cells and soluble factors contained in RA synovial fluid yield a subset of myeloid DCs that preferentially activate Th1 inflammatory-type responses J Immunol 2001 167 1758 1768 11466401
Morelli AE Thomson AW Dendritic cells: regulators of alloimmunity and opportunities for tolerance induction Immunol Rev 2003 196 125 146 14617202 10.1046/j.1600-065X.2003.00079.x
Schmidt A Schwind B Gillich M Brune K Hinz B Simultaneous determination of leflunomide and its active metabolite, A77 1726, in human plasma by high-performance liquid chromatography Biomed Chromatogr 2003 17 276 281 12833393 10.1002/bmc.244
Johnson DR Levanat S Bale AE Direct molecular analysis of archival tumor tissue for loss of heterozygosity Biotechniques 1995 19 190 192 8527133
Stuhlmeier KM Tarn C Csizmadia V Bach FH Selective suppression of endothelial cell activation by arachidonic acid Eur J Immunol 1996 26 1417 1423 8766541
Everson MP McDuffie DS Lemak DG Koopman WJ McGhee JR Beagley KW Dendritic cells from different tissues induce production of different T cell cytokine profiles J Leukoc Biol 1996 59 494 498 8613695
Zhou LJ Tedder TF Human blood dendritic cells selectively express CD83, a member of the immunoglobulin superfamily J Immunol 1995 154 3821 3835 7706722
Pettit AR Quinn C MacDonald KP Cavanagh LL Thomas G Townsend W Handel M Thomas R Nuclear localization of RelB is associated with effective antigen-presenting cell function J Immunol 1997 159 3681 3691 9378953
Yoshimura S Bondeson J Foxwell BM Brennan FM Feldmann M Effective antigen presentation by dendritic cells is NF-kappaB dependent: coordinate regulation of MHC, co-stimulatory molecules and cytokines Int Immunol 2001 13 675 683 11312255 10.1093/intimm/13.5.675
Koski GK Lyakh LA Cohen PA Rice NR CD14+ monocytes as dendritic cell precursors: diverse maturation-inducing pathways lead to common activation of NF-kappab/RelB Crit Rev Immunol 2001 21 179 189 11642603
Pettit AR Thomas R Dendritic cells: the driving force behind autoimmunity in rheumatoid arthritis? Immunol Cell Biol 1999 77 420 427 10540208 10.1046/j.1440-1711.1999.00855.x
Balanescu A Nat R Regalia T Radu E Bojinca V Ionescu R Predescu V Popescu E Predeteanu D Correlation between the immunophenotypical presentation of dendritic cells and the clinical response to anti-rheumatic treatment in rheumatoid arthritis Rom J Intern Med 2003 41 255 267 15526509
Burger D Begue-Pastor N Benavent S Gruaz L Kaufmann MT Chicheportiche R Dayer JM The active metabolite of leflunomide, A77 1726, inhibits the production of prostaglandin E(2), matrix metalloproteinase 1 and interleukin 6 in human fibroblast-like synoviocytes Rheumatology (Oxford) 2003 42 89 96 12509619
Palmer G Burger D Mezin F Magne D Gabay C Dayer JM Guerne PA The active metabolite of leflunomide, A77 1726, increases the production of IL-1 receptor antagonist in human synovial fibroblasts and articular chondrocytes Arthritis Res Ther 2004 6 R181 R189 15142263 10.1186/ar1157
Cutolo M Sulli A Ghiorzo P Pizzorni C Craviotto C Villaggio B Anti-inflammatory effects of leflunomide on cultured synovial macrophages from patients with rheumatoid arthritis Ann Rheum Dis 2003 62 297 302 12634225 10.1136/ard.62.4.297
Urushibara M Takayanagi H Koga T Kim S Isobe M Morishita Y Nakagawa T Loeffler M Kodama T Kurosawa H The antirheumatic drug leflunomide inhibits osteoclastogenesis by interfering with receptor activator of NF-kappa B ligand-stimulated induction of nuclear factor of activated T cells c1 Arthritis Rheum 2004 50 794 804 15022321 10.1002/art.20206
Kaltwasser JP Nash P Gladman D Rosen CF Behrens F Jones P Wollenhaupt J Falk FG Mease P Efficacy and safety of leflunomide in the treatment of psoriatic arthritis and psoriasis: a multinational, double-blind, randomized, placebo-controlled clinical trial Arthritis Rheum 2004 50 1939 1950 15188371 10.1002/art.20253
Cauli A Pitzalis C Yanni G Awad M Panayi GS CD1 expression in psoriatic and rheumatoid arthritis Rheumatology (Oxford) 2000 39 666 673 10888713
Buelens C Verhasselt V De Groote D Thielemans K Goldman M Willems F Interleukin-10 prevents the generation of dendritic cells from human peripheral blood mononuclear cells cultured with interleukin-4 and granulocyte/macrophage-colony-stimulating factor Eur J Immunol 1997 27 756 762 9079819
Chomarat P Banchereau J Davoust J Palucka AK IL-6 switches the differentiation of monocytes from dendritic cells to macrophages Nat Immunol 2000 1 510 514 11101873 10.1038/82763
Van Parijs L Perez VL Biuckians A Maki RG London CA Abbas AK Role of interleukin 12 and costimulators in T cell anergy in vivo J Exp Med 1997 186 1119 1128 9314560 10.1084/jem.186.7.1119
Penna G Adorini L 1 Alpha, 25-dihydroxyvitamin D3 inhibits differentiation, maturation, activation, and survival of dendritic cells leading to impaired alloreactive T cell activation J Immunol 2000 164 2405 2411 10679076
Kobayashi Y Ueyama S Arai Y Yoshida Y Kaneda T Sato T Shin K Kumegawa M Hakeda Y The active metabolite of leflunomide, A77 1726, inhibits both the generation of and the bone-resorbing activity of osteoclasts by acting directly on cells of the osteoclast lineage J Bone Miner Metab 2004 22 318 328 15221489 10.1007/s00774-003-0489-4
Korn T Magnus T Toyka K Jung S Modulation of effector cell functions in experimental autoimmune encephalomyelitis by leflunomide: mechanisms independent of pyrimidine depletion J Leukoc Biol 2004 76 950 960 15328336 10.1189/jlb.0504308
Schlapfer E Fischer M Ott P Speck RF Anti-HIV-1 activity of leflunomide: a comparison with mycophenolic acid and hydroxyurea Aids 2003 17 1613 1620 12853743 10.1097/00002030-200307250-00005
Waldman WJ Knight DA Lurain NS Miller DM Sedmak DD Williams JW Chong AS Novel mechanism of inhibition of cytomegalovirus by the experimental immunosuppressive agent leflunomide Transplantation 1999 68 814 825 10515382 10.1097/00007890-199909270-00014
Manna SK Mukhopadhyay A Aggarwal BB Leflunomide suppresses TNF-induced cellular responses: effects on NF-kappa B, activator protein-1, c-Jun N-terminal protein kinase, and apoptosis J Immunol 2000 165 5962 5969 11067959
Yoshimura S Bondeson J Brennan FM Foxwell BM Feldmann M Role of NFkappaB in antigen presentation and development of regulatory T cells elucidated by treatment of dendritic cells with the proteasome inhibitor PSI Eur J Immunol 2001 31 1883 1893 11433385 10.1002/1521-4141(200106)31:6<1883::AID-IMMU1883>3.0.CO;2-V
Murphy TL Cleveland MG Kulesza P Magram J Murphy KM Regulation of interleukin 12 p40 expression through an NF-kappa B half-site Mol Cell Biol 1995 15 5258 5267 7565674
Kremer JM Westhovens R Leon M Di Giorgio E Alten R Steinfeld S Russell A Dougados M Emery P Nuamah IF Treatment of rheumatoid arthritis by selective inhibition of T-cell activation with fusion protein CTLA4Ig N Engl J Med 2003 349 1907 1915 14614165 10.1056/NEJMoa035075
| 15899055 | PMC1174963 | CC BY | 2021-01-04 16:02:36 | no | Arthritis Res Ther. 2005 Apr 1; 7(3):R694-R703 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1727 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17291589905610.1186/ar1729Research ArticleClinical evaluation of autoantibodies to a novel PM/Scl peptide antigen Mahler Michael [email protected] Reinout [email protected]ähnrich Cornelia [email protected]üthner Martin [email protected] Marvin J [email protected] Dr Fooke Laboratorien GmbH, Neuss, Germany2 Radboud University Nijmegen, The Netherlands3 Euroimmun GmbH, Lübeck, Germany4 Labor Seelig und Kollegen, Karlsruhe, Germany5 Faculty of Medicine, University of Calgary, Canada2005 1 4 2005 7 3 R704 R713 8 1 2005 16 2 2005 22 2 2005 4 3 2005 Copyright © 2005 Mahler et al, licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is cited.
Anti-PM/Scl antibodies represent a specific serological marker for a subset of patients with scleroderma (Scl) and polymyositis (PM), and especially with the PM/Scl overlap syndrome (PM/Scl). Anti-PM/Scl reactivity is found in 24% of PM/Scl patients and is found in 3–10% of Scl and PM patients. The PM/Scl autoantigen complex comprises 11–16 different polypeptides. Many of those proteins can serve as targets of the anti-PM/Scl B-cell response, but most frequently the PM/Scl-100 and PM/Scl-75 polypeptides are targeted. In the present study we investigated the clinical relevance of a major alpha helical PM/Scl-100 epitope (PM1-α) using a newly developed peptide-based immunoassay and compared the immunological properties of this peptide with native and recombinant PM/Scl antigens. In a technical comparison, we showed that an ELISA based on the PM1-α peptide is more sensitive than common techniques to detect anti-PM/Scl antibodies such as immunoblot, indirect immunofluorescence on HEp-2 cells and ELISA with recombinant PM/Scl polypeptides. We found no statistical evidence of a positive association between anti-PM1-α and other antibodies, with the exception of known PM/Scl components. In our cohort a negative correlation could be found with anti-Scl-70 (topoisomerase I), anti-Jo-1 (histidyl tRNA synthetase) and anti-centromere proteins. In a multicenter evaluation we demonstrated that the PM1-α peptide represents a sensitive and reliable substrate for the detection of a subclass of anti-PM/Scl antibodies. In total, 22/40 (55%) PM/Scl patients, 27/205 (13.2%) Scl patients and 3/40 (7.5%) PM patients, but only 5/288 (1.7%) unrelated controls, tested positive for the anti-PM1-α peptide antibodies. These data indicate that anti-PM1-α antibodies appear to be exclusively present in sera from PM/Scl patients, from Scl patients and, to a lesser extent, from PM patients. The anti-PM1-α ELISA thus offers a new serological marker to diagnose and discriminate different systemic autoimmune disorders.
==== Body
Introduction
Systemic autoimmune diseases such as scleroderma (Scl), polymyositis (PM), rheumatoid arthritis, systemic lupus erythematosus (SLE) and mixed connective tissue disease are characterized by the occurrence of circulating antibodies to defined intracellular targets [1]. Some of these autoantibodies represent useful diagnostic markers for a variety of systemic autoimmune diseases [1,2].
Antibodies targeting the PM/Scl complex serve as a marker for the PM/Scl overlap syndrome, where they are found in 24% of sera, but they are also seen in 8% of PM patients and in 3% of Scl patients [3-6]. The PM/Scl complex was identified as the human counterpart of the yeast exosome and consists of 11–16 polypeptides with molecular masses ranging from 20 to 110 kDa [7-11]. PM/Scl-100, the human equivalent of the yeast Rrp6p, has been cloned by two independent groups and its key function during the 5.8 S rRNA end formation has been described [12-14].
In previous studies, the human immune response targeting the PM/Scl complex has been reported to be predominantly directed against two polypeptides with apparent molecular masses of 100 kDa and 75 kDa [15]. In the past it has been shown that nearly all PM/Scl-positive sera contain autoantibodies to the 100 kDa protein and that only about 50–60% react with the 75 kDa protein [7,8,15-17]. A more recent study has shown that the PM/Scl-75 protein contains a previously unidentified N-terminal region that is important for the antigenicity of the protein [18]. The reactivity of sera with this new isoform of PM/Scl-75c is similar to the conventional PM/Scl-100 protein [18]. Several other components of the human exosome, including hRrp4p, hRrp40p, hRrp41, hRrp42p, hRrp46p and hCsl4p, are also recognized by anti-PM/Scl antibodies, but to a lesser extent [10,19].
In several studies during the past decade, we and others have attempted to identify the epitopes on PM/Scl-100 that are recognized by the cognate autoantibodies [12,20-23]. The prime reactivity of anti-PM/Scl-100 sera was localized to a domain of the protein represented by amino acids 231–245 using membrane-bound peptide arrays [22,23]. The amino acids contributing to the antibody binding were identified by mutational analysis [22,23]. Based on these observations and on secondary structure predictions, a local alpha-helical structure has been proposed for this major PM/Scl-100 epitope [22,23].
The aim of this study was to develop an ELISA with a 15-mer peptide comprising the PM/Scl-100 major epitope as a substrate, and to evaluate its sensitivity and specificity for the detection of anti-PM/Scl antibodies.
Materials and methods
Serum samples
In the present study three different serum panels were used to analyze the accuracy of the alpha helical PM/Scl-100 epitope (PM1-α) peptide in the ELISA. For the technical comparative study, 33 sera with anti-PM/Scl reactivity were preselected by indirect immunofluorescence on HEp-2 cells and cryopreserved monkey liver sections (Euroimmun, Lübeck, Germany) and by immunoblot with total cell extracts (Panel I). Panel II consisted of sera from a previous study and included patients with PM/Scl, patients with PM, patients with Scl, patients with dermatomyositis (DM) patients with melanoma and normal donors [18]. For the multicenter evaluation, serum samples were collected from patients with PM/Scl overlap syndrome (n = 40), from patients with Scl (n = 50), from patients with PM (n = 40) and from patients with various control diseases including rheumatoid arthritis (n = 69), SLE (n = 114), undifferentiated connective tissue disease (n = 10), mixed connective tissue disease (n = 6), Hashimoto thyroiditis (n = 11), Grave's disease (n = 12), other autoimmune disorders (n = 8), and hepatitis C virus infection (HCV) (n = 48) (Panel III).
PM/Scl patients were diagnosed based on the official PM and Scl criteria and were only considered true overlap patients if they fulfilled both the criteria for PM and for Scl [24,25]. All other patients with autoimmune disorders were classified according to the official criteria for each disease as also applied in a recent investigation [26]. Sera were stored in aliquots at -80°C until use and were shipped on dry ice. Collection of patient samples was carried out according to local ethics committee regulations.
Antigens for ELISA
The identified sequence LDVPPALADFIHQQR of the PM/Scl-100 (accession number JH0796) major B-cell epitope covering amino acids 231–245 was used to synthesize the PM1-α peptide with an additional cysteine residue at the C-terminus using Fmoc chemistry [22]. Crude peptide obtained from peptide synthesis was purified by high-performance liquid chromatography. The quality and purity of the peptide was assessed by mass spectrometry and analytical high-performance liquid chromatography. The molecular mass was found at 1824.1167 Da (average; monoisotopic mass = 1822.9274 Da) and a purity of 100% was determined. The isoelectric point of the peptide was 4.0. Recombinant PM/Scl-100 (Diarect AG, Freiburg, Germany), was expressed in Escherichia coli and purified via a His-tag, and the quality was ensured by immunoblot and checkerboard analysis of positive and negative sera in the ELISA [27].
Indirect immunofluorescence
Indirect immunofluorescence (IIF) was carried out using BioChip-mosaics with HEp-2 cells and primate liver as substrates (lot number 10116D; Euroimmun GmbH). Antibody titers were determined using 10-fold serial dilutions in PBS and the assay was performed according the manufacturer's instructions.
Immunoblotting
Total cell extracts from HEp-2 cells that were separated by SDS-PAGE and transferred onto nitrocellulose were used as substrate for immunoblotting (lot numbers 01011a-88 and 01011a-89; Euroimmun GmbH). The identity of the PM/Scl antigens was ensured using PM/Scl index sera, which were previously characterized by several methods. Sera were diluted and incubated according to the manufacturer's instruction.
ELISA
The PM1-α peptide was absorbed onto 96-well polystyrene plates (maxisorb; Nunc, Rosilke, Denmark) by overnight incubation at 4°C in 0.1 M carbonate buffer (pH 9.5). Different coating concentrations and different blocking, washing and incubation conditions were compared to optimize the assay conditions. Finally, the evaluation of antibody binding to the PM1-α peptide was performed as follows. Serum samples diluted 1:100 in dilution buffer at a volume of 100 μl/well were incubated for 30 min. After washing three times with washing buffer, anti-human IgG conjugate was added to the wells (100 μl/well) and incubated for 30 min. Surplus conjugate was removed by three washing cycles. The substrate was finally added to each well (100 μl/well) and incubated for 15 min. After stopping the color reaction with stop solution, the absorbance was measured at 450 nm. All steps were carried out at room temperature.
A highly positive index patient serum that was available in larger quantities was used to generate a calibrator. The sample was diluted 1:200 to yield an optical density of about 2.0 in the ELISA. The optical density of each patient sample was divided by the optical density of the calibrator and the result was multiplied by 10. For the technical comparison, the cut-off value of the prototype kits was based on the mean ± three standard deviations of 12 healthy blood donors. During the multicenter study the cut-off was validated and optimized by receiver operating characteristic analysis (see later).
All ELISAs using recombinant proteins were performed as already described, using recombinant proteins expressed in E. coli and purified using either a His-tag or ion-exchange chromatography [17,18].
Addressable laser bead immunoassay
Microspheres embedded with laser reactive dyes (Luminex Corporation, Austin, TX, USA) that were coupled with autoantigens were part of a commercial kit (QUANTA Plex 8 TM; INOVA Diagnostics Inc., San Diego, CA, USA). This profile test allows for the semiquantitative detection of autoantibodies to chromatin, Jo-1, Rib-P, RNP, Scl-70, Sm, SS-A (Ro) and SS-B (La). The assay was performed according to the manufacturer's instructions. Briefly, each test serum was diluted to 1/1000 and 50 μl was added to a well of a microtiter plate, mixed with the antigen-coated beads that were preserved in the well, and incubated for 30 min. Then 50 μl phycoerythrin-conjugated goat anti-human IgG (Jackson ImmunoResearch Laboratories Inc., West Grove, PA, USA) was added to each well and incubated for an additional 30 min. The reactivity of the antigen-coated beads was determined on a Luminex 100™ dual laser flow cytometer (Luminex Corporation). The cut-off for a positive test result was based on the reactivity of control samples. The control samples were titrated to provide high, medium, low and negative values. Further information is available online .
Statistical evaluation of the results
The results obtained from the comparative study were evaluated using Analyse-it software (Version 1.62; Analyse-it Software, Ltd, Leeds, UK). Receiver operating characteristic curves, positive predictive values and negative predictive values, as well as the test efficiency, were calculated. Furthermore, the correlation coefficients between the immunoassays based on the different antigens were calculated.
Results
Technical comparison of IIF, immunoblot and ELISA for the detection of anti-PM/Scl antibodies
To compare the different techniques, 33 anti-PM/Scl sera preselected on the basis of their IIF pattern and/or immunoblot result were tested in prototype ELISA kits based on the full-length recombinant PM/Scl-100 polypeptide expressed in E. coli and on the synthetic PM1-α peptide. In total, 26/33 (78.8%) were positive in the ELISA with the recombinant protein and 32/33 (97.0%) were positive in the ELISA with the synthetic peptide. Results are summarized in Table 1. Based on the high sensitivity of the peptide-based ELISA in this technical comparison, we evaluated the clinical accuracy of the assay in an extended multicenter study using clinically defined sera from various centers.
Correlation of anti-PM1-α with anti-PM/Scl-75a, PM/Scl-75c and PM/Scl-100 reactivity in ELISA
A panel of sera (n = 81) tested previously for reactivity to recombinant PM/Scl proteins (Panel II) was assayed for anti-PM1-α peptide reactivity in the ELISA. The results were compared with the known reactivity of these sera with the recombinant proteins [18]. When all assays were adjusted to the same specificity (91.1%), the clinical sensitivity for the PM/Scl overlap syndrome was 36.1% for PM1-α, was 27.8% for PM/Scl-75c and was 25.0% for PM/Scl-100.
There was a clear correlation between the peptide reactivity and the reactivity of the sera with the recombinant proteins. Not surprisingly, the strongest correlation was observed with the anti-PM/Scl-100 reactivity (Fig. 1). Whereas the majority of the sera showed comparable reactivity in all four assays, some individual samples showed a higher reactivity to the recombinant proteins than to the synthetic peptide, and vice versa. Overall, only one sample (from a patient with DM) was found that tested positive for the recombinant proteins but negative for the synthetic peptide. However, 11 sera that tested positive in the peptide ELISA remained undetected using the recombinant proteins.
When analyzing only the PM/Scl patients from this panel (36/81), the correlation between the reactivity of the peptide and the recombinant proteins was even higher (PM/Scl-100, R2 = 0.82). Very importantly, no sera were found positive for PM/Scl-75c and/or PM/Scl-100 but negative for the PM/Scl peptide in the PM/Scl patient group. Two samples were PM/Scl-75c-positive (new isoform), PM1-α-positive and PM/Scl-100-negative. One sample reacted with PM/Scl-100 and PM1-α but not with the PM/Scl-75 proteins. Of the PM/Scl sera, 27.8% (10/36) was positive for the peptide but was negative for all recombinant polypeptides.
Correlation with other autoantibodies
A statistical evaluation was performed using a patient cohort of 70 clinically defined PM/Scl sera and PM sera to evaluate correlations between anti-PM1-α peptide antibodies and other autoantibodies in ELISA assays using recombinant proteins. No significant correlation was found with Ro-52, Ro-60, La or Mi-2 antibodies (Table 2). Anti-PM1-α antibodies and anti-Jo-1 reactivity were negatively correlated. In addition, a reduced number of samples from 28 patients with clinically defined PM/Scl were also tested in an addressable laser bead immunoassay for autoantibodies to chromatin, Rib-P, RNP, Scl-70 and Sm, SS-A (Ro) and SS-B (La) (QUANTA Plex 8 TM; INOVA Diagnostics Inc.). Although antibodies to chromatin, Rib-P and RNP were detected in some patients, none of these antibodies appeared to be coincident with anti-PM1-α reactivity (Table 2).
Multicenter evaluation of the PM1-α ELISA
Sera from 40 clinically defined but serologically unselected patients with PM/Scl overlap syndrome, as well as from 205 Scl patients, 40 PM patients and various other controls (Panel III), were analyzed in the PM/Scl peptide ELISA (see Table 3). The results from all patients were used to calculate a receiver operating characteristic curve, which showed a clear discrimination between PM/Scl patients and various controls (Fig. 2). At a selected cut-off value of 1.5 RU, 22/40 (55%) PM/Scl patients tested positive for anti-PM1-α antibodies displaying a reactivity of up to 11.6 RU with a mean value of 3.1 ± 3.2 RU (Table 3). Patients from related disorders including Scl and PM showed a lower mean reactivity compared with the overlap patients but a higher reactivity than more unrelated controls. In total, 27/205 (13.2%) scleroderma patients (mean 0.7 ± 1.3 RU) and 3/40 (7.5%) PM patients (mean 1.0 ± 1.1 RU) tested positive, while 3/114 (2.6%) patients with SLE and 2/48 (4.2%) patients with HCV infection had anti-PM1-α antibodies. None of the remaining controls showed reactivity to the PM1-α peptide in the ELISA (Table 3, Fig. 3).
In total, 6.6% control sera tested positive for anti-PM1-α antibodies. This resulted in a diagnostic sensitivity of 55% and a specificity of 93.4% of the peptide ELISA (positive predictive value = 38.6%, negative predictive value = 96.5%, test efficiency = 90.7%). When Scl patients and PM patients were excluded from the group of controls 5/288 (1.7%) patients were positive, resulting in a specificity of 98.2% (positive predictive value = 81.5%, negative predictive value = 94.0%, test efficiency = 92.9%). These data indicate that, within the assay parameters used here, anti-PM1-α antibodies appear to be mainly present in sera from PM/Scl patients, from Scl patients, and to a lesser extent, PM patients.
Discussion
The aim of this study was to compare the autoantigenicity of the PM1-α peptide that we have described previously [22,23] with that of native and recombinant PM/Scl-75 and PM/Scl-100 polypeptides. The results of the technical comparison showed that the PM1-α peptide ELISA is more sensitive than the ELISA tests based on the recombinant proteins, and than immunoblot and IIF experiments. Also, our results suggest that increased titers of autoantibodies directed to PM1-α might be more prevalent in patients with the PM/Scl overlap syndrome and related diseases than autoantibodies to the full-length proteins, which up to now were considered the most frequently present.
In the past, the presence of these antibodies in serum was generally monitored by IIF with HEp-2 cells, by immunodiffusion assays with calf thymus extract and/or by immunoblot using extractable nuclear antigens [4,5,15]. All these techniques allow the detection of a wide variety of autoantibodies present in patient serum [2]. The detection of anti-PM/Scl antibodies by immunoblotting, however, is difficult, because the reactivity of the antibodies with particularly PM/Scl-75 in cell extracts is notoriously weak in immunoblot, which may be due to the importance of conformational epitopes [15]. This observation could be confirmed in the technical comparison of IIF, immunoblot and ELISA in the present study. In recent years, ELISA using recombinant PM/Scl-100 has become a common method to detect anti-PM/Scl reactivity because it can easily be applied in an automated setting.
Since anti-PM/Scl-75 reactivity was previously detected only in patient sera that also contained anti-PM/Scl-100 autoantibodies [15], this protein is usually not included in such assays. A recent investigation has shown that also the use of an incomplete recombinant PM/Scl-75 polypeptide may have led to an underestimation of the diagnostic value of the PM/Scl-75 antigen [18].
We recently characterized the antibody response to a major PM/Scl epitope and found that 14/14 (100%) samples with PM/Scl antibodies demonstrated reactivity to the major epitope in a membrane-based peptide array [22,23]. We have characterized the major PM/Scl-100 B-cell epitope at the amino acid level and identified the key amino acids involved in antibody binding [22,23]. Using this peptide as an antigen, we developed a highly sensitive and specific ELISA system that detects a subpopulation of anti-PM/Scl antibodies present in 55% of PM/Scl patients, in 13.2% of Scl patients and in 7.5% of PM patients. Interestingly, this peptide also contains a generalized T-cell epitope pattern (ALADFIHQQR; amino acids 236–245) as well as several major histocompatibility complex epitopes [28-30].
Synthetic peptides represent ideal antigenic targets for immunoassays because they can easily be produced in high quality and quantity. Furthermore, less lot-to-lot variation will be observed since the production is not dependent on the biological variation of native sources of antigens. More and more synthetic peptides are being used in immunological assay systems to detect autoantibodies. Some of them show higher specificities and sensitivities than the corresponding assay with recombinant protein or native protein as substrate [23].
The combined use of different PM/Scl antigens, including the recombinant PM/Scl-100 and the recently identified isoform of PM/Scl-75, as well as the PM1-α peptide, may represent the most sensitive and specific method to detect antibodies to the human exosome. Advances in multi-analyte technologies such as line assays, multiplex systems and micro-arrays allow for the development of sophisticated profile assays containing multiple different antigens. This may improve the diagnosis of a variety of disorders, especially of autoimmune diseases since for most of those disorders no highly sensitive marker is available. The diagnosis of PM/Scl, Scl and PM might be improved by providing an antigen array that includes different PM/Scl antigens in combination with Scl-70 (topoisomerase I), Ku70/86, centromere proteins, RNA polymerase, NOR-90, Jo-1, Mi-2, PL-7, PL-12 and fibrillarin.
Taken together, the use of the PM/Scl-100 synthetic peptide in an ELISA remarkably improves the clinical identification of patients with the PM/Scl overlap syndrome. Although the prevalence of autoantibodies recognizing most other exosome subunits is relatively low [8,10,11], the co-occurrence of antibodies targeting different exosome subunits in patient sera might be indicative of intermolecular epitope spreading and might be a marker for the overlap syndrome. The co-occurrence of anti-PM/Scl-100 and anti-PM/Scl-75 seems to be particularly associated with the PM/Scl overlap syndrome [18], but whether the use of even more components of the human exosome will further increase the sensitivity of these assays remains to be investigated.
Apart from patients with PM/Scl overlap syndrome and patients with Scl or PM alone, two HCV-positive and three SLE patients displayed reactivity to the PM1-α peptide in ELISA. HCV infection has been associated with a plethora of immune and autoimmune perturbations [31]. Although the cause and effect remain to be proved, there are reports of HCV infection preceding or coincident with polyarthritis, rheumatoid arthritis, SLE, and PM/DM. The role of anti-PM1-α antibodies in HCV patients and SLE patients remain a matter for further investigation. In the present study we found antibodies to the PM1-α peptide present in 13.2% of unselected scleroderma patients. In only a few of those patients was a history of myositis documented. It is possible that the myositis was mild in the majority of patients and was completely overlooked by the examining clinician or that the antibody precedes the associated clinical features [32]. We therefore conclude that the complete autoantibody profile is important for a careful examination of patients with rheumatic diseases and to access all their clinical features.
Frank and colleagues analyzed sera from 216 patients with idiopathic inflammatory myopathies to assess putative associations between anti-SS-A/Ro-52 and other autoantibodies. These included sera containing antibodies that recognize Jo-1, Mi-2, PM/Scl, signal recognition particle, as well as the scleroderma-related antibodies anti-topoisomerase I (Scl-70) and anti-centromere. A high proportion of sera that contain anti-Jo-1 antibodies, anti-signal recognition particle or anti-PM/Scl antibodies were found to contain antibodies to the Ro52 protein [33]. The reported association between anti-Ro-52 and anti-PM/Scl antibodies is not found in our cohort. Although our correlation study is based on a limited number of samples, we found no correlation between anti-PM/Scl antibodies and anti-Ro52. In contrast, Yamanishi and colleagues reported an association of PM/Scl syndrome with anti-Ku antibody and rimmed vacuole formation [34]. Similar to this observation, we found that two out of 29 (6.9%) anti-PM/Scl-positive samples also were positive for anti-Ku86 antibodies. In a previous study it became evident that the PM/Scl-100 major epitope shares some sequence homology to an amino acid stretch (amino acids 58–72) of the heterochromatin protein p25β, which is frequently the target of anti-chromo antibodies from a subpopulation of patients also having anti-centromere antibodies [22]. Although none of 14 PM1-α-positive samples showed reactivity to the corresponding region of p25β, a more complex immunological relationship between the major PM/Scl-100 epitope and the corresponding p25β peptide cannot be excluded. More samples with anti-PM/Scl and anti-chromo antibodies have to be tested for cross-reactivity. Further studies are required to analyze the association of anti-PM1-α antibodies with other known autoantibodies.
Today's sophisticated epitope mapping methods will probably lead to the identification of additional peptides, which can be used as specific targets in diagnostic and therapeutic approaches to patient management. This may lead to a new scientific research area with high impact for the development of diagnostic and therapeutic products – to the area of peptide engineering.
Conclusion
In the present study, we showed that the detection of anti-PM/Scl antibodies using an ELISA system based on a major PM/Scl-100 epitope is remarkably improved compared with conventional detection methods. It could be shown that a subpopulation of PM/Scl antibodies directed against the PM1-α peptide is present in 55% of PM/Scl patients, in 13.2% of Scl patients and in 7.5% of PM patients. In rare cases anti-PM1-α reactivity was also found in patients suffering from HCV, SLE or melanoma. Within our patient cohorts we found no statistical evidence of a positive association between anti-PM1-α and antibodies other than to PM/Scl components. Based on the results of the present study we conclude that anti-PM1-α antibodies are exclusively present in sera from patients suffering from Scl or PM and most frequently in patients with the PM/Scl overlap syndrome. We therefore conclude that the new anti-PM1-α ELISA test offers a new serological test that will improve the diagnosis of complex connective tissue disorders.
Abbreviations
DM = dermatomyositis; ELISA = enzyme-linked immunosorbent assay; HCV = hepatitis C virus; IIF = indirect immunofluorescence; PBS = phosphate-buffered saline; PM = polymyositis; PM1-α = alpha helical PM/Scl-100 epitope; RU = relative units; Scl = scleroderma; SLE = systemic lupus erythematosus.
Competing interests
MM is employed at Dr Fooke Laboratorien GmbH, which may commercialize the assay.
Authors' contributions
MM developed and validated the ELISA system, planned the experiments and filed the manuscript. MJF and RR delivered clinically defined sera, advised MM in evaluating the clinical part of this study and contributed to the preparation of the manuscript. CD organized the analysis of anti-PM/Scl samples in IIF and immunoblot. MB advised MM during the characterization of the PM1-α peptide.
Acknowledgements
The authors thank Dr R Mierau and Prof. E Genth (Rheumaklinik Aachen, Germany) for providing clinically defined sera, Mark L Fritzler (University of Calgary, Canada) for technical assistance with the addressable laser bead immunoassay and Wilma Vree Egberts (Radboud University Nijmegen, The Netherlands) for technical assistance.
Figures and Tables
Figure 1 Correlation diagrams of PM1-α, PM/Scl-75a, PM/Scl-75c and PM/Scl-100. A panel of sera tested previously for reactivity to recombinant polymyositis/scleroderma (PM/Scl) components (PM/Scl-75a, PM/Scl-75c and PM/Scl-100) was assayed for anti-PM1-α peptide reactivity in an ELISA [18]. Correlation diagrams are shown comparing the peptide ELISA with the recombinant proteins (a)–(c) for all sera (n = 81) and (b)–(f) for only the sera of PM/Scl patients (n = 36).
Figure 2 Receiver operating characteristic analysis of the PM1-α ELISA. Results obtained from three centers and based on 567 patients including polymyositis/scleroderma (PM/Scl) patients (n = 40), Scl patients (n = 205) and PM patients (n = 40) as well as other controls were used to calculate a receiver operating characteristic analysis (a) for all control samples and (b) for unrelated controls (without Scl and PM). The curve shows a clear discrimination between PM/Scl patient samples and various controls as emphasized by an area under the curve value of 0.901 (all controls) and 0.958 (unrelated controls). The differentiation between PM/Scl patients and controls was significantly improved when Scl patients and PM patients were excluded from the control group (b). SE, standard error.
Figure 3 Reactivity of polymyositis/scleroderma (PM/Scl) patients and controls in the PM1-α ELISA. Results obtained from three centers and based on 567 patients including PM/Scl patients (n = 40), Scl patients (n = 205) and PM patients (n = 40) as well as other controls were used to calculate comparative descriptive analysis. The diagram shows a significantly increased reactivity of the PM/Scl sera compared with the control groups. Comparative descriptives show vertical box-plots for each sample, side by side for comparison. The blue line series shows parametric statistics: diamond, mean and the requested confidence interval around the mean; notched line, requested parametric percentile range. The notched box and whiskers show non-parametric statistics: notched box, median, lower and upper quartiles, and confidence interval around the median; dotted line, connects the nearest observations within 1.5 interquartile ranges (IQR) of the lower and upper quartiles. + and ○, possible outliers – observations more than 1.5 IQR (near outliers) and more than 3.0 IQR (far outliers) from the quartiles. Vertical lines, requested nonparametric percentile range. SLE, systemic lupus erythematosus; HCV, hepatitis C virus; RA, rheumatoid arthritis.
Table 1 Results of the technical comparison of indirect immunofluorescence (IIF), immunoblot and ELISA for the detection of anti-polymyositis/scleroderma (anti-PM/Scl) antibodies
Sample number IIF Immunoblot ELISA
Titer Pattern PM/Scl-75 PM/Scl-100 Other PM/Scl-100 PM1-α
1 1:10000 FG, N ++ +++ 15.3 17.7
2 1:1000 FG, N + + 2.9 7.6
3 1:3200 FG, N ++ ++ 14.4 29.2
1:320 AMA
4 1:320 FG, N - + 1.1 4.7
5 1:1000 FG, N + + 2.2 12.1
6 1:3200 FG, N - - 0.6 2.3
7 1:1000 FG, N ++ +++ 5.2 13.3
8 1:1000 FG, N + + 2.3 12.7
9 1:3200 FG, N + +++ P38 15.5 36.1
10 - - + 1.7 4.0
11 1:1000 FG, N - ++ 19.2 2.4
12 1:1000 FG, N ++ + 1.2 3.2
13 1:3200 FG, N +++ +++ Ku86, Jo-1 6.5 10.6
14 1:3200 FG, N - ++ 2.1 5.9
15 1:3200 FG, N ++ +++ 11.4 22.5
16 1:1000 FG, N - + 3.6 8.5
17 1:1000 FG, N + ++ 3.2 15.0
18 1:3200 FG, N ++ ++ 6.1 11.8
19 n.d. n.d. - ++ 7.1 19.4
20 1:320 N - + 2.1 5.6
21 n.d. n.d. +++ ++ 4.1 12.8
22 1:10000 FG - + P38 0.7 1.5
1:10000 SPA
23 1:1000 N n.d. n.d. 0.5 2.7
24 n.d. n.d. ++ +++ 18.2 34.7
25 1:10000 N n.d. n.d. P38 0.5 2.0
1:1000 Rib
26 1:1000 N n.d. n.d. 1.9 8.4
27 n.d. n.d. + ++ 5.9 19.4
28 n.d. n.d. ++ +++ 20.9 37.3
29 1:1000 Hom n.d. n.d. Ku86, Cen, M2 0.6 0.8
1:100 N*
30 1:1000 N - + 0.6 1.4
31 1:3200 Hom, N - - 0.5 2.9
32 1:3200 N - - 0.5 1.3
33 1:320 FG, N + ++ 7.2 25.7
Number positive/tested 26/29 17/29 27/29 26/33 32/33
-, negative; +, weak positive; ++, positive; +++, strong positive; n.d., not determined; FG, fine granular; Hom, homogenous; SPA, spindle apparatus; N, nucleoli; AMA, anti-mitochondrial antibodies; Cen, centromere; Rib, ribosomal. * Primate liver.
Table 2 Correlation of anti-PM1-α and other known autoantibodies
Number positive/all sera (% positive) Number positive/PM1-α-positives (% positive) Number positive/PM1-α-negatives (% positive) P
Scl-70 0/28 (0) 0/15 (0) 0/13 (0) *
Sm 0/28 (0) 0/15 (0) 0/13 (0) *
Rib-P 1/28 (3.6) 1/15 (6.7) 0/13 (0) *
RNP 4/28 (14.4) 1/15 (6.7) 3/13 (23.1) *
Chromatin 1/28 (3.6) 1/15 (6.7) 0/13 (0%) *
Ro-52 18/70 (25.7) 6/19 (31.6) 12/51 (23.5) 0.7056
Ro-60 6/70 (8.6) 2/19 (10.6) 4/51 (7.8) 0.9018
La 3/70 (4.3) 1/19 (5.3) 2/51 (3.9) 0.6766
Mi-2 10/70 (14.3) 3/19 (15.8) 7/51 (13.7) 0.8693
Jo-1 16/70 (22.9) 1/19 (5.2) 15/51 (29.4) 0.0688
* Not calculated due to the limited number of samples.
Table 3 Results of ELISA using PM1-α peptide with polymyositis/scleroderma and various control sera
Number (%) of anti-PM1-α-positive sera Mean value/standard deviation Top value
Polymyositis/scleroderma (n = 40) 22 (55) 3.1/3.2 11.6
Rheumatic disease controls (n = 452) 33 (7.3) 0.6/0.9 7.7
Polymyositis (n = 40) 3 (7.5) 1.0/1.1 7.4
Scleroderma (n = 205) 27 (13.2) 0.9/1.2 7.5
Rheumatoid arthritis (n = 69) 0 (0) 0.3/0.2 1.1
Mixed connective tissue disease (n = 6) 0 (0) 0.4/0.1 0.6
Undifferentiated connective tissue disease (n = 10) 0 (0) 0.3/0.0 0.4
Systemic lupus erythematosus (n = 114) 3 (2.6) 0.5/0.7 7.7
Other rheumatic diseases (n = 8) 0 (0) 0.3/0.1 0.6
Hepatitis C virus (n = 48) 2 (4.2) 0.5/0.5 2.6
Organ specific disorders (n = 23) 0 (0) 0.4/0.2 0.8
Hashimoto thyroiditis (n = 11) 0 (0) 0.3/0.2 0.8
Grave's disease (n = 12) 0 (0) 0.4/0.2 0.8
Healthy individuals (n = 4) 0 (0) 0.6/0.2 0.7
==== Refs
Tan EM Antinuclear antibodies: diagnostic markers for autoimmune diseases and probes for cell biology Adv Immunol 1989 44 93 151 2646863
von Muhlen CA Tan EM Autoantibodies in the diagnosis of systemic rheumatic diseases Semin Arthritis Rheum 1995 24 323 358 7604300 10.1016/S0049-0172(95)80004-2
Wolfe JF Adelstein E Sharp GC Antinuclear antibody with distinct specificity for polymyositis J Clin Invest 1977 59 176 178 318657
Reimer G Steen VD Penning CA Medsger TA JrTan EM Correlates between autoantibodies to nucleolar antigens and clinical features in patients with systemic sclerosis (scleroderma) Arthritis Rheum 1988 31 525 532 2451921
Reichlin M Maddison PJ Targoff I Bunch T Arnett F Sharp G Treadwell E Tan EM Antibodies to a nuclear/nucleolar antigen in patients with polymyositis overlap syndromes J Clin Immunol 1984 4 40 44 6699138 10.1007/BF00915286
Oddis CV Okano Y Rudert WA Trucco M Duquesnoy RJ Medsger TA Jr Serum autoantibody to the nucleolar antigen PM-Scl. Clinical and immunogenetic associations Arthritis Rheum 1992 35 1211 1217 1418007
Reimer G Scheer U Peters JM Tan EM Immunolocalization and partial characterization of a nucleolar autoantigen (PM-Scl) associated with polymyositis/scleroderma overlap syndromes J Immunol 1986 137 3802 3808 3537125
Gelpi C Alguero A Angeles Martinez M Vidal S Juarez C Rodriguez-Sanchez JL Identification of protein components reactive with anti-PM/Scl autoantibodies Clin Exp Immunol 1990 81 59 64 2199097
Allmang C Petfalski E Podtelejnikov A Mann M Tollervey D Mitchell P The yeast exosome and human PM-Scl are related complexes of 3' → 5' exonucleases Genes Dev 1999 13 2148 2158 10465791
Brouwer R Pruijn GJ van Venrooij WJ The human exosome: an autoantigenic complex of exoribonucleases in myositis and scleroderma Arthritis Res 2001 3 102 106 11178117 10.1186/ar147
Brouwer R Allmang C Raijmakers R van Aarssen Y Egberts WV Petfalski E van Venrooij WJ Tollervey D Pruijn GJ Three novel components of the human exosome J Biol Chem 2001 276 6177 6184 11110791 10.1074/jbc.M007603200
Bluthner M Bautz FA Cloning and characterization of the cDNA coding for a polymyositis-scleroderma overlap syndrome-related nucleolar 100-kD protein J Exp Med 1992 176 973 980 1383382 10.1084/jem.176.4.973
Ge Q Frank MB O'Brien C Targoff IN Cloning of a complementary DNA coding for the 100-kD antigenic protein of the PM-Scl autoantigen J Clin Invest 1992 90 559 570 1644924
Briggs MW Burkard KT Butler JS Rrp6p, the yeast homologue of the human PM-Scl 100-kDa autoantigen, is essential for efficient 5.8 S rRNA 3' end formation J Biol Chem 1998 273 13255 13263 9582370 10.1074/jbc.273.21.13255
Ge Q Wu Y Trieu EP Targoff IN Analysis of the specificity of anti-PM-Scl autoantibodies Arthritis Rheum 1994 37 1445 1452 7945469
Targoff IN Reichlin M Nucleolar localization of the PM-Scl antigen Arthritis Rheum 1985 28 226 230 3918546
Brouwer R Hengstman GJ Vree Egberts W Ehrfeld H Bozic B Ghirardello A Grondal G Hietarinta M Isenberg D Kalden JR Autoantibody profiles in the sera of European patients with myositis Ann Rheum Dis 2001 60 116 123 11156543 10.1136/ard.60.2.116
Raijmakers R Renz M Wiemann C Egberts WV Seelig HP van Venrooij WJ Pruijn GJ PM-Scl-75 is the main autoantigen in patients with the polymyositis/scleroderma overlap syndrome Arthritis Rheum 2004 50 565 569 14872500 10.1002/art.20056
Brouwer R Vree Egberts WT Hengstman GJ Raijmakers R van Engelen BG Seelig HP Renz M Mierau R Genth E Pruijn GJ van Venrooij WJ Autoantibodies directed to novel components of the PM/Scl complex, the human exosome Arthritis Res 2002 4 134 138 11879549 10.1186/ar389
Ge Q Wu Y James JA Targoff IN Epitope analysis of the major reactive region of the 100-kd protein of PM-Scl autoantigen Arthritis Rheum 1996 39 1588 1595 8814071
Bluthner M Bautz EK Bautz FA Mapping of epitopes recognized by PM/Scl autoantibodies with gene-fragment phage display libraries J Immunol Methods 1996 198 187 198 8946014 10.1016/S0022-1759(96)00160-3
Bluthner M Mahler M Muller DB Dunzl H Bautz FA Identification of an alpha-helical epitope region on the PM/Scl-100 autoantigen with structural homology to a region on the heterochromatin p25beta autoantigen using immobilized overlapping synthetic peptides J Mol Med 2000 78 47 54 10759029 10.1007/s001090050381
Mahler M Bluthner M Pollard KM Advances in B-cell epitope analysis of autoantigens in connective tissue diseases Clin Immunol 2003 107 65 79 12763475 10.1016/S1521-6616(03)00037-8
Anonymous Preliminary criteria for the classification of systemic sclerosis (scleroderma). Subcommittee for scleroderma criteria of the American Rheumatism Association Diagnostic and Therapeutic Criteria Committee Arthritis Rheum 1980 23 581 590 7378088
Bohan A Peter JB Polymyositis and dermatomyositis N Engl J Med 1975 292 344 347 1090839
Mahler M Fritzler MJ Bluthner M Identification of a SmD3 epitope with a single symmetrical dimethylation of an arginine residue as a specific target of a subpopulation of anti-Sm antibodies Arthritis Res Ther 2005 7 R19 R29 15642139 10.1186/ar1455
Datasheet Recombinant PM/Scl-100 Autoantigen
Falk K Rotzschke O Stevanovic S Jung G Rammensee HG Allele-specific motifs revealed by sequencing of self-peptides eluted from MHC molecules Nature 1991 351 290 296 1709722 10.1038/351290a0
Rothbard JB Taylor WR A sequence pattern common to T cell epitopes EMBO J 1988 7 93 100 2452085
Sette A Buus S Appella E Smith JA Chesnut R Miles C Colon SM Grey HM Prediction of major histocompatibility complex binding regions of protein antigens by sequence pattern analysis Proc Natl Acad Sci USA 1989 86 3296 3300 2717617
McMurray RW Elbourne K Hepatitis C virus infection and autoimmunity Semin Arthritis Rheum 1997 26 689 701 9062950 10.1016/S0049-0172(97)80005-4
Arbuckle MR MT McClain MV Rubertone RH Scofield GJ Dennis JA James JB Harley Development of autoantibodies before the clinical onset of systemic lupus erythematosus N Engl J Med 2003 349 1526 1533 14561795 10.1056/NEJMoa021933
Frank MB McCubbin V Trieu E Wu Y Isenberg DA Targoff IN The association of anti-Ro52 autoantibodies with myositis and scleroderma autoantibodies J Autoimmun 1999 12 137 142 10047434 10.1006/jaut.1998.0265
Yamanishi Y Maeda H Katayama S Ishioka S Yamakido M Scleroderma-polymyositis overlap syndrome associated with anti-Ku antibody and rimmed vacuole formation J Rheumatol 1996 23 1991 1994 8923381
| 15899056 | PMC1174964 | CC BY | 2021-01-04 16:02:37 | no | Arthritis Res Ther. 2005 Apr 1; 7(3):R704-R713 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1729 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17231598747610.1186/ar1723Research ArticleSerum protein profile in systemic-onset juvenile idiopathic arthritis differentiates response versus nonresponse to therapy Miyamae Takako [email protected] David E [email protected] Bonnie [email protected] Masaaki [email protected] Tomoyuki 3Yokota Shumpei [email protected] William L [email protected] Manda [email protected] Klaus [email protected] Norihiro 5Vallejo Abbe N [email protected] Raphael [email protected] Division of Rheumatology, Children's Hospital of Pittsburgh, Department of Pediatrics, University of Pittsburgh School of Medicine, Pittsburgh, PA 152132 University of Pittsburgh Cancer Institute, University of Pittsburgh School of Medicine, Pittsburgh, PA 152133 Department of Pediatrics, Yokohama City University School of Medicine, Yokohama, Japan4 Départment de Phamacologie, Faculté de Medicine, Université de Sherbrooke, Québec, Canada5 Osaka University, Osaka, Japan2005 4 4 2005 7 4 R746 R755 20 1 2005 22 2 2005 26 2 2005 28 2 2005 Copyright © 2005 Miyamae et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Systemic-onset juvenile idiopathic arthritis (SJIA) is a disease of unknown etiology with an unpredictable response to treatment. We examined two groups of patients to determine whether there are serum protein profiles reflective of active disease and predictive of response to therapy. The first group (n = 8) responded to conventional therapy. The second group (n = 15) responded to an experimental antibody to the IL-6 receptor (MRA). Paired sera from each patient were analyzed before and after treatment, using surface-enhanced laser desorption/ionization time-of-flight mass spectrometry (SELDI-TOF MS). Despite the small number of patients, highly significant and consistent differences were observed before and after response to therapy in all patients. Of 282 spectral peaks identified, 23 had mean signal intensities significantly different (P < 0.001) before treatment and after response to treatment. The majority of these differences were observed regardless of whether patients responded to conventional therapy or to MRA. These peaks represent potential biomarkers of active disease. One such peak was identified as serum amyloid A, a known acute-phase reactant in SJIA, validating the SELDI-TOF MS platform as a useful technology in this context. Finally, profiles from serum samples obtained at the time of active disease were compared between the two patient groups. Nine peaks had mean signal intensities significantly different (P < 0.001) between active disease in patients who responded to conventional therapy and in patients who failed to respond, suggesting a possible profile predictive of response. Collectively, these data demonstrate the presence of serum proteomic profiles in SJIA that are reflective of active disease and suggest the feasibility of using the SELDI-TOF MS platform used as a tool for proteomic profiling and discovery of novel biomarkers in autoimmune diseases.
==== Body
Introduction
Systemic-onset juvenile idiopathic arthritis (SJIA) is a form of childhood arthritis of unknown etiology, characterized by systemic features in addition to arthritis, including spiking fever, erythematous rash, articular involvement, and other, visceral manifestations [1]. Its clinical course is associated with changes in the levels of several serum proteins, including IL-6 [2]. Over half of children with SJIA eventually recover almost completely [3]. The other half have severe, unremitting arthritis, poorly responsive to conventional therapy, leading to poor functional outcome and substantial morbidity [4]. In view of the heterogeneity of clinical disease manifestations and the unpredictability of treatment responses in SJIA, there would be great clinical benefit in the discovery of biomarkers reflective of disease activity and predictive of response to therapy.
Proteomics, or protein pattern analysis, is the characterization and quantitation of proteins in tissues and body fluids [5]. Proteomic methods can be used to compare protein expression patterns between disease states. Although two-dimensional gel electrophoresis has been the primary technique in conventional proteomic analysis, it is relatively insensitive to proteins of low abundance and below 10 kDa in mass, is labor intensive, and has low throughput. A more recent technology known as surface-enhanced laser desorption/ionization time-of-flight mass spectrometry (SELDI-TOF MS), a derivative of conventional matrix-associated laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS), involves the application of a biologic sample, such as serum, to a protein-binding chip [6]. The chip is irradiated with a laser, resulting in ionization of the adherent molecules. The ions travel through a vacuum tube and their mass-to-charge ratios are calculated from their time of flight through the vacuum chamber. The technology is high throughput, rapid, and sensitive and provides a profile of low-molecular-weight peptides and proteins within a complex mixture such as serum.
SELDI-TOF MS does not directly identify specific proteins. It has been used to differentiate disease states from nondisease states by analysis of protein profiles in sera. Examples include the differentiation of neoplastic from non-neoplastic breast masses [7], prognostic and diagnostic classification of breast cancer [8], neoplastic versus non-neoplastic disease of the ovary [9], and prostate cancer from both men with benign hyperplasia and healthy men [10]. SELDI-TOF MS has also been used for the discovery of disease-related biomarkers in sera. Examples include detection of serum amyloid α in patients with renal cancer [11] and the quantitation of prostate-specific membrane antigen in prostate cancer [12].
The present study was designed to determine whether there are serum proteomic profiles in SJIA that are reflective of active disease and predictive of response to therapy, as well as to determine whether SELDI-TOF MS could be used as a tool for proteomic profiling and for discovery of novel biomarkers of SJIA.
Materials and methods
Patients and study subjects
Banked sera from 23 patients (14 boys, 9 girls) with SJIA according to the criteria established by the International League of Association for Rheumatology [13] were obtained from the Department of Pediatrics, Yokohama City University School of Medicine, Yokohama, Japan. All the patients were Asian and their mean age at the start of the study was 7.25 ± 0.92 years. Eight of them had obtained a clinical response to conventional therapy. Clinical response was defined as the absence of fever rash, hepatosplenomegaly, and arthritis for at least 3 months, accompanied by normalization of serum C-reactive protein. Briefly, conventional therapy consisted of three doses of intravenous methylprednisolone (30 mg/kg per day) or oral prednisolone (1 to 2 mg/kg), followed by nonsteroidal anti-inflammatory drugs (NSAIDs) and a tapering dose of oral prednisolone. In addition, methotrexate (2.5 to 5 mg/m2 per week orally) was used in three patients and cyclosporin (5 mg/kg per day orally) in two. The mean period from acute status to complete clinical response was 27.7 ± 14.6 months. Fifteen patients who had inadequate response to the above therapy, as well as to the addition of azathioprine (five patients), mizoribine (five patients), sulfasalazine (two patients), or plasma exchange (three patients), had been administered humanized anti-IL-6 receptor antibody (MRA; Chugai Pharmaceuticals, a subsidiary of Roche Pharmaceuticals). All 15 patients had a clinical response to MRA. The mean period from acute status to clinical response was 11.2 ± 5.1 months. Pretreatment sera were collected before starting conventional treatment or giving the initial dose of MRA. Post-treatment sera were collected 2 to 3 months after patients achieved a clinical response. Ethical approval for this study was granted by Yokohama University. The study was approved by both Chugai Pharmaceuticals and Roche Pharmaceuticals.
SELDI-TOF MS
Serum samples were thawed on ice, denatured, and processed in duplicate on IMAC-3 (immobilized metal affinity capture) copper ProteinChip® Arrays (Ciphergen Biosystems, Fremont, CA, USA). ProteinChips were loaded, processed, and prepared for mass spectrometry using a Biomek2000 liquid handling robot (Beckman-Coulter, Fullerton, CA, USA) and optimized for reproducibility using validated protocols. ProteinChips were read in a PBSIIc mass spectrometer (Ciphergen) with mass deflection at 1 kDa and time-lag focusing. The resulting mass spectra were examined between m/z values of 2 and 100 kDa for quantitative comparison of identifiable peak features. The parameters used for spectral preprocessing and peak selection were: external calibration (seven peptide calibrants, 1 to 7 kDa, Ciphergen), baseline subtraction by 8 × expected peak width and smoothing, filtering by average using 0.2 expected peak width, noise defined over 1500 Da, normalization by total ion current (TIC) over 1500 Da, peaks detected over 2000 kDa by centroid mass. Weak spectra were excluded from analysis if the normalization factor exceeded 2 standard deviations above the mean normalization factor.
Statistical analysis
Peak clustering among sample groups was performed with the Biomarker Wizard (Ciphergen) tool, with a peak detection threshold of 5 for signal-to-noise ratio, and mass tolerance of 0.3%, for any peak appearing in at least 5% of experimental spectra being compared. The Biomarker Wizard compares the mean intensity of peak clusters, by sample group, using the nonparametric Mann–Whitney U test (two–way comparisons) and generates P values that reflect probabilities that mean peak intensities at a given m/z value differ by random chance. The intensity values of the automatically clustered peaks (averaged between technical replicates of each sample) were used in classification tree analysis (CART) using Ciphergen's Biomarker Patterns Software. This supervised learning process uses cross-validation to optimize the minimization of classification error.
Immunoprecipitation of SAA
A pooled sample from four sera taken before conventional treatment was incubated with either Protein A–Sepharose beads alone or Protein A–Sepharose beads bound with 100 μg of anti-SAA antibody (Anogen, Yes Biotech Laboratories, Mississauga, ON, Canada). After immunoprecipitation, the depleted serum was subjected to SELDI-TOF MS.
LC-ESI-MS/MS-TOF analysis
Protein identification by MS was carried out as previously described [14]. Briefly, serum samples were subjected to immunoprecipitation with anti- SAA or with an IgG isotype control. Immunoprecipitates were washed extensively in phosphate-buffered serum and centrifuged, and the pellet was sonicated for 10 min in 8 M urea/400 mM NH4HCO3. The supernatant was diluted in water to a final concentration of 2 M urea/100 mM NH4HCO3 and digested overnight at 37°C with 1 μg trypsin. The tryptic digest was subjected to nano-LC-ESI-MS/MS analysis that was performed on a Q-TOF-2™ (Waters, Milford, MA, USA), coupled on line to a CapLC system equipped with three separate syringe pump modules, an auto injector, a 10-port valve and a 250-μm (inner diameter) × 1-mm pre-column. Separations were performed on a 7-cm × 75-μm (inner diameter) capillary column. Both columns were packed with Microsorb C18 (Varian, Mississauga, ON, Canada) reverse-phase material. Peptides were eluted at a flow rate of 0.25 μl/min with the following linear gradient of solvent B (80% aqueous acetonitrile with 10% isopropanol and 0.2% formic acid) in solvent A: from 0 to 60% B in 40 min, to 90% B in 7 min, and to 10% B in 8 min. Spectra were acquired in auto MS/MS mode conducted using survey scans to choose up to three precursor ions. Collision energies were selected automatically as a function of m/z value and charge state. The Q-TOF mass spectrometer was calibrated by infusing a solution of either NaI containing a small amount of cesium ion dissolved in 50% aqueous isopropanol (0.2 μg/μl) or Glu-fibrinopeptide B (1 pmol/μl) dissolved in 30% aqueous acetonitrile containing 0.2% formic acid. Protein identification was performed using the MASCOT search program (Matrix Science Limited, ) and the NCBI (National Center for Biotechnology Information) (Bethesda, MD, USA) protein database.
Results
Protein profiling by SELDI-TOF MS reveals distinct patterns differentiating active from well-controlled SJIA
To determine whether SELDI-TOF MS could be a valuable tool for analyzing serum protein profiles in autoimmunity, we chose SJIA as a test model, since the disease has systemic features, in addition to arthritis, likely to be reflected in the serum. Paired sera were available from patients who had been followed up for a mean of 24.4 ± 6.9 months. The availability of paired sera from each patient allowed for longitudinal comparison and substantially reduced sample variability between the two groups. Sera from eight patients with SJIA who responded to conventional therapy (therapy and definition of clinical response are described in detail in the Materials and methods section) were analyzed before and after therapy by SELDI-TOF MS using Ciphergen IMAC-3 (immobilized metal affinity capture) copper chips (Fig. 1). All samples were run in duplicate. Table 1 shows the most significant differences between mass spectra of sera before and after conventional therapy. Only variables with nonparametric P values of <0.001 are given. Of 282 spectral peaks identified, 23 had mean signal intensities that were significantly different (P < 0.001) before and after response to treatment. These peaks represent potential biomarkers of active disease. We next performed a similar analysis on paired sera from 15 patients who had failed conventional therapy but responded to an experimental antibody to the IL-6 receptor (MRA) [15]. These sera were obtained after failure of conventional therapy. Pre- and post-MRA sera revealed similar profiles to those observed in the pre- and post-conventional therapy group. Thus, substantial consistency was observed in protein profiles, regardless of whether patients with active disease responded to conventional therapy or to MRA. Eight of the differentially expressed peaks represent prominent, visually distinct spectral features. These peaks are represented in Table 1 in bold, along with the number of paired patient sera in which each peak was differentially expressed by visual inspection of the spectra. Representative examples of these peaks are shown in Fig. 2.
To determine the usefulness of the profiles in classifying active versus controlled SJIA, the data were subjected to CART (Biomarker Patterns Software, Ciphergen) analysis. This sample classification method is designed, through multivariate analysis, to construct classification trees recognizing a complex pattern of multiple peak intensities. The method is ideally suited for sample sets large enough to permit cross-validation internal to the 'training' data, but also the segregation of additional unused data as a validation or 'testing' set. On these relatively small sample sets, CART was used in training mode primarily as a data exploration tool. Whether using the training set as the MRA group, or as the conventional treatment group, the CART analysis returned simple classification trees consisting of one primary splitter, either 11.4 kDa or 11.6 kDa (m/z). The primary splitter at 11493 kDa correctly identified 13 of 14 pretreatment and 14 of 15 post-MRA treatment samples when conventional treatment was used as the training set. When MRA treatment was used as the training set, all of the pre- and post-conventional treatment samples were correctly identified as either pretreatment or post-treatment. The distinction between these samples by CART registered at the most extreme level of significance the program is capable of indicating. Even when forced to ignore the mass spectrum peaks at 11493 or 11650 Da, the CART program was able to effectively discriminate, using secondary peaks derived from them (at half these m/z values; attributed to doubly protonated species). This robust classification surpasses the performance of any other sample set being profiled and analyzed by this and several other statistical methods at this institution (data not shown).
Identification of serum amyloid A from SELDI-TOF MS mass spectra
A prominent group of peaks within the range 11.4 to 11.7 kDa m/z strongly distinguished the pre- and post-treatment samples (Fig. 3). The post-treatment groups showed an apparently single m/z peak at 11.75 kDa, which was also routinely observed in pooled reference sera from healthy adults (data not shown). A previous study in nasopharyngeal cancer, using the same SELDI-TOF technique and IMAC3 copper chip, identified two biomarkers, of 11.6 and 11.8 kDa, as serum amyloid A (SAA) [16]. Since SAA is a known biomarker of active SJIA [17], we compared the intensity of the 11.6-kDa peak with SAA levels in the sera, as determined by latex agglutination. As shown in Fig. 4, a strong correlation was observed (R2 = 0.74), suggesting that the 11.6-kDa peak might represent SAA. To further investigate the biochemical identity of this peak, serum containing high levels of the 11.6-kDa peak was subjected to immunoprecipitation using anti-SSA antibody bound to Protein A–Sepharose beads. As shown in Fig. 5, after immunoprecipitation and SELDI analysis, the 11.4- and 11.6-kDa peaks were markedly diminished. To confirm the identity of the immunoprecipitated protein, anti-SAA precipitates were digested with trypsin and subjected to η-scale liquid chromatography electrospray ionization tandem mass spectrometry time-of-flight (LC-ESI-MC/MS-TOF) analysis. Over 90 peptide ions were examined and only 2 proteins were identified, including immunoglobulin and SAA. The SAA peptide ions represented approximately 51% of the SAA sequence.
Protein profiling by SELDI-TOF MS reveals patterns differentiating the responding from the nonresponding SJIA group
The above data, using paired sera, demonstrate the ability of SELDI-TOF MS to identify biomarkers of active disease, as exemplified by the identification of SAA. A long-term goal is to predict clinical outcome, based on protein profiles present in the serum early in the disease course. To begin to approach this challenge, we compared the pretreatment serum profiles of the 8 patients who responded to conventional therapy with those from the 15 patients who responded poorly to conventional therapy. Similar to the preceding analysis, the latter samples were obtained after failure of conventional therapy and before MRA treatment, when the patients still had active disease. In this initial exploratory study, the number of available samples was too small for definitive conclusions; however, several interesting trends were apparent. Several highly significant differences were observed in the mass spectra of these sera, as shown in Table 2 and Fig. 6. These peaks may represent a profile predictive of response to conventional therapy. Alternatively, they could represent the effects of conventional therapy or differences between early versus long-standing disease. A number of peaks overlap with regions observed in Table 1, including the region of SAA (11.6 kDa) as well as 4504 kDa and 28 kDa.
We were fortunate to have pre-conventional treatment sera available from 3 of the 15 patients who responded poorly to conventional therapy and went on to receive MRA. These three pretreatment sera were compared with the pretreatment sera of the eight patients who responded to conventional therapy. Three peaks of interest were observed. As shown in Fig. 7, all the nonresponders had lower values for the 4825-Da feature and higher values for the 3276-Da and 3293-Da peaks than did the responders, with the exception of a single outlier sample. However the sample size is small and this observation needs further validation in a larger clinical cases series; this putative signature of nonresponse may be susceptible to statistical overfitting, even at this level of analysis.
Discussion
Current diagnostic techniques for rheumatic diseases are based on clinical presentation and nonspecific serum markers. Because the phenotype of a rheumatic disease such as SJIA is largely dependent on proteins, the present study was designed to determine whether serum protein expression profiling with SELDI-TOF MS could be used to search for new molecular diagnostic biomarkers and potential therapeutic targets. This approach has theoretical advantages over other modalities used to identify differentially expressed proteins. SELDI-TOF MS analysis is capable of detecting small amounts of protein, hence the potential to detect proteins of relatively low abundance with affinity for the ProteinChip surface. The technique is high throughput, allowing detection of hundreds of species in a single sample, and is capable of analyzing large number of samples. The data presented here show that it is possible to generate mass spectrometry protein expression profiles from serum that can differentiate active versus controlled SJIA.
Some of the difficulties inherent in gene or protein expression profiling in human disease include accounting for the genetic and environmental variability between patients and the potential for detecting chance associations when measuring large numbers of proteins or genes. The use of paired serum samples from individual patients in the present study removes most of these variables and makes it likely that the changes in protein profiles after successful treatment reflect the disease state rather than confounding variables. We found a surprising degree of consistency in the relative abundance of a number of serum proteins in ill versus well patients. The clear distinction in the levels of these various ionic species between these sample groups permits a robust classification based on simple thresholding on any one of a number of possible variables.
One disadvantage of the SELDI-TOF MS technology is that protein sequences, and thus specific identifications, are not obtained, requiring further biochemical/mass spectrometry analysis to identify differentially-expressed proteins. A recent study using two-dimensional gels and MALDI-TOF MS analysis of plasma and synovial fluids from patients with rheumatoid arthritis or osteoarthritis also revealed the presence of SAA in samples from rheumatoid arthritis but not osteoarthritis [18]. Although SELDI-TOF MS is not directly quantitative, it can detect changes in the relative abundance of proteins in a manner that compares favorably to quantitative methods such as latex agglutination or enzyme-linked immunosorbent assay. Identification of SAA by SELDI-TOF MS helps validate our experimental approach, since SAA is a known marker of active SJIA.
Although we were able to identify SAA by further analysis, there were many other peaks observed in the serum profiles that have yet to be explored or identified. The 66.6-kDa and 33.4-kDa peaks most likely represent serum albumin and its doubly protonated form, as they are the correct mass and they increase after response to therapy, reflective of the known rise in serum albumin levels in these patients (data not shown). Identification of the other peaks is currently being investigated and may yield novel information on the pathophysiology of SJIA. Furthermore, the proteins observed represent only a fraction of those present in the serum. We observed only a subset of relatively high-abundance proteins, limited by their concentration in the serum, their affinity with the copper matrix of the IMAC ProteinChip, and the relative desorption/ionization efficiencies of each protein. In addition, the fact that the mass spectra generated in this study were from unfractionated sera is likely to obscure many protein species that might otherwise be detectable in the absence of high-abundance serum proteins. Refinement of the methodologies for processing serum samples, including initial depletion of high-abundance proteins, is likely to substantially increase the information that can be derived from the resulting profiles.
There are likely many more subtypes of the group of diseases known collectively as idiopathic arthritis than have as yet been defined by clinical criteria. The ability to differentiate uncontrolled from controlled SJIA by serum protein profiling raises the possibility of more specific diagnostic and prognostic criteria for evaluating such patients. Furthermore, the dramatic mass spectral differences observed between the sample groups led us to compare sera obtained before any treatment, from patients who ultimately differed in their response to conventional therapy. While the current work can only comment on an anecdotal basis from a limited number of these samples, some early differences were observed that suggest that a prognostic profile might exist. Beyond the obvious clinical usefulness of such a profile, it also could provide a discovery tool for further characterization of the pathophysiology of SJIA.
Although SELDI-TOF MS was recently used to compare synovial fluids from patients with rheumatoid arthritis and osteoarthritis [19], the present study is, to our knowledge, the first to define a serum proteomic profile of a rheumatic disease using SELDI-TOF MS. The SELDI-TOF MS technique described here provides a rapid, high throughput, and mass accurate method for detecting relative quantities of multiple disease-related proteins simultaneously. Using this platform, we identified a protein (SAA) known to be elevated in active SJIA. This proteomic profiling approach has the potential to expand the current repertoire of molecular targets and to provide diagnostic and prognostic information useful for improving the care of and ultimate outcome for SJIA patients.
Conclusion
This study demonstrates the presence of serum proteomic profiles in SJIA that are reflective of active disease and suggests the feasibility of using the SELDI-TOF MS platform used as a tool for proteomic profiling in autoimmune diseases. Furthermore, the study validates the ability of the SELDI-TOF MS platform to identify a known biomarker of SJIA (SAA), suggesting that it may also be useful as a screening approach towards the discovery of novel biomarkers. To that end, identifying the 22 unknown m/z protein species in the serum profiles of our patients is now the focus of further investigation.
Abbreviations
IL = interleukin; IMAC-3 = immobilized metal affinity capture; LC-ESI-MS/MS-TOF = liquid chromatography electrospray ionization tandem mass spectrometry time-of-flight; MALDI-TOF MS = matrix-associated laser desorption/ionization time-of-flight mass spectrometry; MRA = humanized anti-IL-6 receptor monoclonal antibody; SAA = serum amyloid A; SELDI-TOF MS = surface-enhanced laser desorption/ionization time-of-flight mass spectrometry; SJIA = systemic juvenile idiopathic arthritis.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
TM, DM, BL, and RH participated in all experimental design, data collection, and analysis and helped draft the manuscript. MM, TI, SY, and NN provided patient sera and clinical data. MW carried out the sample preparation. KK and AV carried out the LC ESI-MS/MS-TOF analysis and helped draft the manuscript. All authors read and approved the final manuscript.
Figures and Tables
Figure 1 SELDI-TOF MS technique. Serum samples are spotted onto IMAC-3 copper chips® (Ciphergen Biosystems). The chip is irradiated with a laser, resulting in ionization of the adherent molecules. The ions travel through a vacuum tube and their mass-to-charge ratios are calculated from their time of flight through the vacuum chamber. IMAC-3, immobilized metal affinity capture; SELDI-TOF MS, surface-enhanced laser desorption/ionization time-of-flight mass spectrometry.
Figure 2 SELDI-TOF MS profiles for patients with SJIA treated conventionally. Six serum protein peaks can be clearly seen to have changed after conventional therapy. The profiles of a representative patient are shown here. Visually distinct peaks that were clearly different between pre- and post-treatment paired samples upon visual inspection of the profiles (in bold type in Table 1) are outlined in grey. Pre- and post-treatment spectra are shown on the same intensity scale in each frame. SELDI-TOF MS, surface-enhanced laser desorption/ionization time-of-flight mass spectrometry SJIA, systemic juvenile idiopathic arthritis.
Figure 3 SELDI-TOF MS profiles for patients with SJIA before and after conventional or MRA treatment. Sera taken before conventional and MRA treatment show similar patterns that are distinct from the post-treatment profiles. Mean spectra of all patients are shown in the 11- to 12-kDa m/z range. Means are compiled from 8 samples before and after conventional therapy and 15 samples before and after MRA therapy, each sample run in duplicate. Spectra were preprocessed as described in the Materials and methods section. MRA, humanized anti-IL-6 receptor monoclonal antibody; SELDI-TOF MS, surface-enhanced laser desorption/ionization time-of-flight mass spectrometry SJIA, systemic juvenile idiopathic arthritis.
Figure 4 Peak intensities of the 11.6-kDa m/z SELDI peak in serum after MRA treatment in SJIA. The peak intensities correlated with the SAA titers measured by latex agglutination. MRA, humanized anti-IL-6 receptor monoclonal antibody; SAA, serum amyloid A; SELDI, surface-enhanced laser desorption SJIA, systemic juvenile idiopathic arthritis.
Figure 5 Immunoprecipitation of SAA in SJIA, resulting in loss of the 11.4- to 11.6-kDA peak cluster. A pooled sample from four sera before conventional treatment (top panel) was incubated with either Protein A–Sepharose beads alone (middle panel) or Protein A–Sepharose beads bound with anti-SAA antibody (lower panel). After immunoprecipitation, the depleted serum was subjected to SELDI-TOF MS. The 11.4-to 11.6-kDa m/z peak cluster is shown in grey. MRA, humanized anti-IL-6 receptor monoclonal antibody; SAA, serum amyloid A; SELDI-TOF MS, surface-enhanced laser desorption/ionization time-of-flight mass spectrometry; SJIA, systemic-onset juvenile idiopathic arthritis.
Figure 6 Differences between mass spectra of sera before and after MRA treatment for SJIA. Sera were taken before and after treatment with MRA from patients for whom conventional therapy had failed. The high significance of these differences suggests a profile predictive of response to conventional therapy. The values represent mean intensities. The P values of these univariate comparisons are given in Table 2. Error bars represent standard deviations. MRA, humanized anti-IL-6 receptor monoclonal antibody; SJIA, systemic-onset juvenile idiopathic arthritis.
Figure 7 Serum proteins in SJIA patients according to whether they responded to conventional therapy. Three most significant differences distinguishing between pretreatment samples from conventional therapy responders (n = 8) and those from nonresponders (n = 3), suggesting a profile predictive of response to conventional therapy. The averaged peak intensity is shown for the eight pretreatment 'responder' patient samples (left panel) compared with the corresponding intensities of those same three peaks from the three pretreatment 'nonresponder' patient samples (right panel). SJIA, systemic-onset juvenile idiopathic arthritis.
Table 1 SELDI-TOF MS protein peaks differentially expressed in paired sera from SJIA before and after therapy
Before/after conventional therapy Before/after MRA
Mass (m/z) P Patients with visually distinct peaks/total no. of patients P Patients with visually distinct peaks/total no. of patients
4504 0.0001 0.005
4758 0.0005 0.0006
5739 0.0001 0.00000008
6441 0.00009 0.0001
6947 0.0005 0.003
9510 0.0007 6/8 0.00001 12/15
9725 0.00006 0.00001
11405 0.00001 8/8 0.000000003 14/15
11520 0.000007 8/8 0.000000001 14/15
11641 0.000005 8/8 0.0000000001 14/15
11718 0.0004 0.00002
11880 0.0002 0.0000006
12703 0.0006 7/8 0.04 6/15
20816 0.0002 7/8 0.01 9/15
22389 0.0002 0.00005
23627 0.0002 0.09
28734 0.0002 0.00006
33457 0.00001 6/8 0.00001 11/15
66637 0.0001 7/8 0.000009 11/15
74886 0.0006 0.00003
75622 0.0009 0.0001
76319 0.0008 0.00002
79216 0.0007 0.000003
Differences between mass spectra of sera before and after conventional therapy. Sera from eight patients with SJIA who responded to conventional therapy were analyzed before and after therapy by SELDI-TOF MS using Ciphergen IMAC-3 copper chips. All samples were run in duplicate. Only variables with nonparametric P values of <0.001 are given. Of 282 spectral peaks identified, the 23 listed here had mean signal intensities significantly different (P < 0.001) before and after response to treatment. Proteins are listed according to mass/charge ratio. Visually distinct peaks (in bold type) refers to peaks that were clearly different between paired samples from before and after treatment upon visual inspection of the profiles, as shown in Fig. 1. The numbers of patients in whom these peaks were visually distinct are shown in columns 3 and 5. These peaks represent potential biomarkers of active disease. IMAC-3, immobilized metal affinity capture; MRA, humanized anti-IL-6 receptor monoclonal antibody; SELDI-TOF MS, surface-enhanced laser desorption/ionization time-of-flight mass spectrometry; SJIA, systemic juvenile idiopathic arthritis.
Table 2 SELDI-TOF MS unpaired serum protein peaks differentially expressed in SJIA before and after conventional therapy
Mass (m/z) P
4504 0.0004
11650 0.0003
11691 0.0002
14047 0.00001
28107 0.000002
28958 0.0000005
39871 0.0004
46175 0.0008
60806 0.0002
Differences between mass spectra of pretreatment sera of 8 patients who went on to respond to conventional therapy compared with post-treatment sera of 15 patients who responded poorly to conventional therapy. Only variables with nonparametric P values of <0.001 are listed, from 272 peak clusters surveyed. These peaks may represent a profile predictive of response or nonresponse to conventional therapy. SELDI-TOF MS, surface-enhanced laser desorption/ionization time-of-flight mass spectrometry; SJIA, systemic juvenile idiopathic arthritis.
==== Refs
Still GF On a form of chronic joint disease in children Med Chir Trans 1897 80 47
de Benedetti F Massa M Robbioni P Ravelli A Burgio GR Martini A Correlation of serum interleukin-6 levels with joint involvement and thrombocytosis in systemic juvenile rheumatoid arthritis Arthritis Rheum 1991 34 1158 1163 1930333
Spiegel LR Schneider R Lang BA Birdi N Silverman ED Laxer RM Stephens D Feldman BM Early predictors of poor functional outcome in systemic-onset juvenile rheumatoid arthritis: a multicenter cohort study Arthritis Rheum 2000 43 2402 2409 11083261 10.1002/1529-0131(200011)43:11<2402::AID-ANR5>3.0.CO;2-C
Oen K Malleson PN Cabral DA Rosenberg AM Petty RE Reed M Schroeder ML Cheang M Early predictors of longterm outcome in patients with juvenile rheumatoid arthritis: subset-specific correlations J Rheumatol 2003 30 585 593 12610821
Petricoin EF Zoon KC Kohn EC Barrett JC Liotta LA Clinical proteomics: translating benchside promise into bedside reality Nat Rev Drug Discov 2002 1 683 695 12209149 10.1038/nrd891
Wright GL Jr SELDI proteinchip MS: a platform for biomarker discovery and cancer diagnosis Expert Rev Mol Diagn 2002 2 549 563 12465452 10.1586/14737159.2.6.549
Li J Zhang Z Rosenzweig J Wang YY Chan DW Proteomics and bioinformatics approaches for identification of serum biomarkers to detect breast cancer Clin Chem 2002 48 1296 1304 12142387
Laronga C Becker S Watson P Gregory B Cazares L Lynch H Perry RR Wright GL JrDrake RR Semmes OJ SELDI-TOF serum profiling for prognostic and diagnostic classification of breast cancers Dis Markers 2003 19 229 238 15258337
Petricoin EF Ardekani AM Hitt BA Levine PJ Fusaro VA Steinberg SM Mills GB Simone C Fishman DA Kohn EC Use of proteomic patterns in serum to identify ovarian cancer Lancet 2002 359 572 577 11867112 10.1016/S0140-6736(02)07746-2
Adam BL Qu Y Davis JW Ward MD Clements MA Cazares LH Semmes OJ Schellhammer PF Yasui Y Feng Z Serum protein fingerprinting coupled with a pattern-matching algorithm distinguishes prostate cancer from benign prostate hyperplasia and healthy men Cancer Res 2002 62 3609 3614 12097261
Tolson J Bogumil R Brunst E Beck H Elsner R Humeny A Kratzin H Deeg M Kuczyk M Mueller GA Serum protein profiling by SELDI mass spectrometry: detection of multiple variants of serum amyloid alpha in renal cancer patients Lab Invest 2004 84 845 856 15107802 10.1038/labinvest.3700097
Xiao Z Adam BL Cazares LH Clements MA Davis JW Schellhammer PF Dalmasso EA Wright GL Jr Quantitation of serum prostate-specific membrane antigen by a novel protein biochip immunoassay discriminates benign from malignant prostate disease Cancer Res 2001 61 6029 6033 11507047
Petty RE Southwood TR Baum J Bettay E Glass DN Manners P Maldonado-Cocco J Suarez-Almazor M Orozco-Alcala J Prieur AM Revision of the proposed classification criteria for juvenile idiopathic arthritis: Durban, 1997 J Rheumatol 1998 25 1991 1994 9779856
Vallejo AN Bryl E Klarskov K Naylor S Weyand CM Goronzy JJ Molecular basis for the loss of CD28 expression in senescent T cells J Biol Chem 2002 277 46940 46949 12324461 10.1074/jbc.M207352200
Choy E Interleukin 6 receptor as a target for the treatment of rheumatoid arthritis Ann Rheum Dis 2003 ii68 69 14532153
Cho WC Yip TT Yip C Yip V Thulasiraman V Ngan RK Lau WH Au JS Law SC Cheng WW Identification of serum amyloid a protein as a potentially useful biomarker to monitor relapse of nasopharyngeal cancer by serum proteomic profiling Clin Cancer Res 2004 10 43 52 14734450
Scheinberg MA Hubscher O Morteo OG Benson MD Serum amyloid protein levels in South American children with rheumatoid arthritis: a co-operative study Ann Rheum Dis 1980 39 228 230 7416812
Sinz A Bantscheff M Mikkat S Ringel B Drynda S Kekow J Thiesen HJ Glocker MO Mass spectrometric proteome analyses of synovial fluids and plasmas from patients suffering from rheumatoid arthritis and comparison to reactive arthritis or osteoarthritis Electrophoresis 2002 23 3445 3456 12373775 10.1002/1522-2683(200210)23:19<3445::AID-ELPS3445>3.0.CO;2-J
Uchida T Fukawa A Uchida M Fujita K Saito K Application of a novel protein biochip technology for detection and identification of rheumatoid arthritis biomarkers in synovial fluid J Proteome Res 2002 1 495 499 12645617 10.1021/pr025531w
| 15987476 | PMC1175022 | CC BY | 2021-01-04 16:02:37 | no | Arthritis Res Ther. 2005 Apr 4; 7(4):R746-R755 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1723 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17281598747810.1186/ar1728Research ArticleA monoclonal antibody against kininogen reduces inflammation in the HLA-B27 transgenic rat Keith James C [email protected] Irma M [email protected] Irma [email protected] Robin A [email protected] Yelena 1Albert Leo M 1Colman Robert W [email protected] Department of Cardiovascular and Metabolic Diseases Research, Wyeth Research, Cambridge, Massachusetts, USA2 The Sol Sherry Thrombosis Research Center, Temple University School of Medicine, Philadelphia, Pennsylania, USA2005 4 4 2005 7 4 R769 R776 31 1 2005 3 3 2005 Copyright © 2005 Keith et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The human leukocyte antigen B27 (HLA-B27) transgenic rat is a model of human inflammatory bowel disease, rheumatoid arthritis and psoriasis. Studies of chronic inflammation in other rat models have demonstrated activation of the kallikrein–kinin system as well as modulation by a plasma kallikrein inhibitor initiated before the onset of clinicopathologic changes or a deficiency in high-molecular-mass kininogen. Here we study the effects of monoclonal antibody C11C1, an antibody against high-molecular-mass kininogen that inhibits the binding of high-molecular-mass kininogen to leukocytes and endothelial cells in the HLA-B27 rat, which was administered after the onset of the inflammatory changes. Thrice-weekly intraperitoneal injections of monoclonal antibody C11C1 or isotype IgG1 were given to male 23-week-old rats for 16 days. Stool character as a measure of intestinal inflammation, and the rear limbs for clinical signs of arthritis (tarsal joint swelling and erythema) were scored daily. The animals were killed and the histology sections were assigned a numerical score for colonic inflammation, synovitis, and cartilage damage. Administration of monoclonal C11C1 rapidly decreased the clinical scores of pre-existing inflammatory bowel disease (P < 0.005) and arthritis (P < 0.001). Histological analyses confirmed significant reductions in colonic lesions (P = 0.004) and synovitis (P = 0.009). Decreased concentrations of plasma prekallikrein and high-molecular-mass kininogen were found, providing evidence of activation of the kallikrein–kinin system. The levels of these biomarkers were reversed by monoclonal antibody C11C1, which may have therapeutic potential in human inflammatory bowel disease and arthritis.
==== Body
Introduction
Human leukocyte antigen B27 (HLA-B27) transgenic Fisher rats are normal at birth but develop chronic inflammation of multiple organ systems as they age. Transgenic rats of this strain, overexpressing the human HLA-B27 and β2-microglobulin proteins, develop lesions of the gastrointestinal system, the joints, the skin, and the gonads, which seem similar to the spondyloarthropathies in humans that have been associated with the HLA-B27 and β2-microglobulin genes [1,2]. The gastrointestinal inflammation is mostly limited to the mucosa and submucosa, exhibiting histological features similar to those present in inflammatory bowel disease (IBD) [1-4]. Chronic intestinal inflammation is the first to occur, with clinical signs of diarrhea apparent after 12 weeks of age. About 4 weeks later, joint inflammation is seen, and these rats can also be used for a model of inflammatory arthritis [3].
The plasma kallikrein–kinin system (KKS), which is initiated by factor XIIa [5] or prolylcarboxypeptidase [6] after binding of high-molecular-mass kininogen (HK) and plasma prekallikrein (PK) to the surface of endothelial cells and leukocytes [7], generates the enzyme kallikrein. Kallikrein in turn cleaves HK to yield the inflammatory mediators bradykinin (BK) and cleaved high-molecular-mass kininogen (HKa) [8]. Kallikrein is chemotactic, aggregates neutrophils [9], stimulates superoxide formation, and releases elastase from neutrophils [10], all of which induce tissue injury. BK stimulates vascular permeability and angiogenesis after binding to endothelial cells [11] and also mediates pain through the release of prostanoids [12]. HKa stimulates cytokine release from rat [13] and human monocytes[14]. Thus, activation of the KKS is an inflammatory stimulus that might be operative in human disease, as represented in Fig. 1.
We have shown that KKS activation mediates the acute and chronic phases of T cell-mediated arthritis induced by peptidoglycan–polysaccharide complexes from Group A streptococci (PG-APS) in Lewis rats [15] and is selectively activated in granulomatous enterocolitis in these susceptible rats, but not in resistant Buffalo rats [16]. We have discovered a genetic difference in kininogen structure between resistant Buffalo and Fischer F344 inbred rats and the susceptible Lewis rat that results in accelerated cleavage of HK in the latter. This mutation consists of a single nucleotide polymorphism coding for the amino acid alteration, S511N, in the HK gene of Lewis (N511) (mutant) versus Buffalo and Fischer (S511) (wild-type) rats that results in an altered glycosylation state [17] and an increased rate of HK cleavage by plasma kallikrein with release of BK. We have shown that BK has a critical role in the PG-APS-mediated arthritis [18]. We have also implicated BK receptors as having a role in a different model of IBD, indomethacin-induced colitis [19]. Most recently, we have shown that a monoclonal antibody (mAb), C11C1, acting to prevent HK interaction with cells involved in inflammatory disorders, inhibited the development of acute and chronic arthritis in the PG-APS model [20].
To demonstrate that this effect was not specific for a single model and to allow us to assess the possibility of treating established chronic inflammation, we examined an HLA-B27 transgenic rat model of chronic inflammation of the intestine and peripheral joints. Administration of mAb C11C1 ameliorated colitis and tarsal joint inflammation.
Materials and methods
HLA-B27 transgenic male rats were purchased from Taconic Laboratories (Germantown, NY) and housed one per cage in accordance with Wyeth Research facility standard operating procedures. They received a standard regimen of food and water. Animals were thoroughly acclimated to the laboratory before the beginning of the study. The study was approved by the Wyeth Research (Cambridge) Institutional Animal Care and Use Committee.
At 23 weeks of age, 10 male rats presenting the clinical signs of colitis (diarrhea) and arthritis (erythematous and swollen hind paws) were randomized into either an isotype control mAb IgG (n = 5) or mAb C11C1 (n = 5) treatment group. Each rat was weighed daily and received an intraperitoneal injection of isotype IgG1 (6 mg/kg) or mAb C11C1 (1.9 mg/kg) three times per week for 16 days. Stool character observations for each animal on each day of study were assigned numerical scores of 3 for diarrhea, 2 for soft stool and 1 for normal stool. The clinical signs of arthritis in the tarsal joints were monitored daily in all of the animals. This assessment was performed visually with a scale for swelling (0 to 3) and for erythema (0 to 3) of the hindpaws (normal paw = 0, mild = 1, moderate = 2, severe = 3). The maximum possible score for arthritis per animal per paw per day was 6 (total per animal = 12 for both hindpaws).
Histological analyses
At the end of the experiment, the animals were killed with 100% carbon dioxide, and the distal 10 cm of colon of each rat was removed and opened. Four standardized samples of colon were immersed in 10% neutral buffered formalin [21]. Samples from each rat were prepared for histological evaluation. The formalin-fixed tissues were processed in a Tissue Tek vacuum infiltration processor, Model 4617 (Miles, Inc., West Haven, CT) for paraffin embedding. The samples were sectioned at 5 μm thickness and then stained with hematoxylin and eosin (H&E) for histological evaluation. Histological lesions were assigned scores in accordance with a previously defined scoring scheme [21-24]. In brief, the severity in the colonic sections was evaluated for ulcer size (none = 0, small = 1, large = 2), degree of inflammation (none = 0, mild = 1, moderate = 2, severe = 3), depth of lesion (none = 0, submucosa = 1, muscularis propria = 2, involving serosa = 3), and fibrosis (none = 0, mild = 1, severe = 2). The total histological scores for the colon specimens ranged from 0 to 10.
During necropsy, segments of the rear limbs (with the tarsal joints) were removed, fixed in 10% buffered formalin, and examined as described previously [22]. After decalcification, histological sections were obtained and stained with H&E or Safranin O/Fast Green stain. Synovial tissue from tarsal joints was evaluated on the basis of synovial hyperplasia (synovial cell proliferation: mild = 1, moderate = 2, villus formation = 3), fibroplasia (subsynovial fibrosis: minimal = 1, one-third to one-half of areolar tissue replacement = 2, whole thickness areolar tissue replacement = 3), inflammatory cell infiltrates (occasional = 0, small numbers/around blood vessels = 1, small focal collections = 2, large foci = 3), and pannus formation (organizing inflammatory exudates within the joint space: nondetectable = 0, detectable = 2). The total histological score for synovial inflammation ranged from 0 to 11 [25]. Articular cartilage was evaluated with Mankin's histological grading system [26]: cartilage organization changes (normal = 0, surface irregularity = 1, pannus and surface irregularity = 2, clefts to transitional zone = 3, clefts to radial zone = 4, clefts to calcified zone = 5, complete disorganization = 6), chondrocyte proliferation (none = 0, hypercellularity = 1, cloning = 2, hypocellularity = 3), proteoglycan contents (Safranin O/Fast Green staining, normal = 0, slight reduction = 1, modest reduction = 2, severe reduction = 3, no dye noted = 4), and tidemark integrity (intact = 0, crossed by blood vessels = 1). The total Mankin score ranged from 0 to 14. Histological H&E-stained sections taken from kidney, liver, and spleen from the mAb C11C1-treated group were evaluated for signs of systemic inflammation and/or toxicity.
Blood collection
Blood samples were obtained by cardiac puncture with a 19-gauge, 3/4-inch needle on a 10 ml polypropylene syringe (BD Medical Systems, Franklin Lakes, NJ). The sample was obtained from the left atrium as the heart beat. The sample of 3 to 5 ml was obtained by slow vacuum (to prevent hemolysis) within a minute (to prevent clotting in the syringe). The blood was then transferred into pre-marked, 1 ml Eppendorf polypropylene tubes (Fisher Scientific, Pittsburgh, PA) containing 100 μl of anticoagulant (citrate-phosphate-dextrose solution with adenine, Sigma C-4431; Sigma Chemical Co.) to a final volume of 1 ml and gently mixed. Plasma was isolated by double centrifugation of the citrated blood in polypropylene tubes (Fisher Scientific) at 23°C. Aliquots were stored at -70°C until assayed.
Assays of KKS activation ex vivo
PK function levels were performed by a microtiter, amidolytic assay using a chromogenic substrate, S-2302 (Pro-Phe-Arg-p-nitroanilide; Chromogenix, Moindal, Sweden), as described previously [27]. HK coagulant activity was evaluated by our modification of an APTT test assay [28,29], using total kininogen-deficient plasma purchased from George King (Overland Park, KS) [19]. In addition, factor XI and factor XII coagulant activity assays were performed with a similar method using the appropriate deficient plasma obtained from George King.
Statistical analyses
All the evaluations were made by examiners blinded to the treatment groups. All of the parameters were subjected to Students' t test between groups. Data were expressed as means ± SEM, and differences were deemed significant if P < 0.05.
Results
Twenty-four hours after the onset of therapy in the mAb C11C1-treated rats, the clinical signs of intestinal inflammation (diarrhea) had disappeared, and the stool character remained normal or nearly normal for the duration of the experiment (Fig. 2a). Histological analysis demonstrated significant reductions (from 60 to 75%) in lesion scores in the animals treated with mAb C11C1 in comparison with animals injected with isotype IgG1 (Fig. 2b,c).
Daily visual inspection of the tarsal joints in the mAb C11C1-treated animals revealed marked reductions in the degree of swelling and erythema of the joints compared with isotype-treated animals. As can be seen in Fig. 3a, within 24 hours of the onset of therapy, the mean joint histological scores in the mAb C11C1-treated rats decreased by about 50% compared with the mAb control group. By the end of 1 week of treatment, the clinical signs of arthritis had almost disappeared. Evaluation of the histological features of the arthritis in the tarsal joints at the termination of the experiment on day 16 showed a marked reduction in the parameters of synovitis in the rats treated with mAb C11C1 compared with those receiving isotype IgG1 (P < 0.05) (Fig. 3b,c). In a similar manner to the changes seen in the colon, 40 to 60% decreases in the various components of the synovitis score occurred. However, the effects on the articular cartilage were more modest. Nevertheless, the cartilage organization, chondrocyte proliferation and total Mankin score were significantly decreased (Fig. 3d). Tidemark integrity was preserved in all groups (data not shown). Histological analysis of kidney, liver and spleen sections showed normal architecture without any signs of inflammation or toxicity in both treated groups (results not shown).
KKS activation assays
To assess KKS system activation in this animal model of inflammation, we compared the experimental groups' results with a standard pool of normal Fischer 344 rat plasma (Fig. 4). We measured the plasma functional levels of four contact proteins. At the termination of the experiment (day 16), HK levels were reduced in both groups compared with the standard pool level. The values in the mAb C11C1-treated animals were closer to normal than those in the isotype-treated animals. HK levels were significantly lower in the isotype IgG-treated group (74.7 ± 1.0) than in the group receiving mAb C11C1 (83.9 ± 1.1) (P < 0.001). PK levels were significantly decreased in the IgG isotype group (52.5 ± 1.3%) versus the mAb C11C1-treated group (60.1 ± 1.3%; P < 0.005). Factor XI was similarly lower in both experimental groups but factor XII was not lower (in any group). Neither difference in factor XI or factor XII levels between the two experimental groups was significant. The results of these assays were similar to those observed in our previous studies [20], in which a decrease in HK and PK was the most consistent evidence for KKS activation.
Discussion
Therapy with C11C1, a mAb that interferes with the cellular binding of HK, evoked marked anti-inflammatory activity in both the colon and the tarsal joints of HLA-B27 transgenic rats. The onset of anti-inflammatory activity by mAb C11C1 was rapid and sustained throughout the study, with the first effect seen in the intestine. The joint changes began to resolve with improvement in stool character, but it took almost 10 days for the joint swelling and erythema to reach minimal levels (as reflected in joint score values). The histological effects in the colon seemed to be more complete than those seen in the tarsal joints because only a modest effect was seen on the articular cartilage lesions, as reflected in the Mankin score. However, if one compares the colonic score results with the synovitis score results, the effect was very similar in both character and magnitude. The isotype IgG1 group KKS assays showed a decrease in HK and PK levels consistent with this system activation, whereas the mAb C11C1-treated group showed significantly increased levels of both proteins. These observations are explained by the fact that mAb C11C1 inhibits the activation of HK, thus blocking KKS activation and decreasing the signs of inflammation [20].
The HLA-B27 transgenic rat model has been used for several years to evaluate the activity and mechanisms of actions of anti-inflammatory molecules [22,23,30-34]. This model is very reproducible and consistent, as long as the environmental conditions remain stable. The chronic inflammation seen in these transgenic rats seems to be the result of HLA-B27 transgene expression-induced alterations in antigen processing and subsequent immune responses to the microbial environment in the lumen of the animal's gastrointestinal tract [35,36]. These aberrant responses lead to CD4+ T cell activation and proinflammatory cytokine production. Broad-spectrum antibiotic therapy can produce significant remissions of the inflammatory lesions, but relapse occurs when antibiotic therapy stops [35]. If antibiotic therapy is followed by inoculation of the gut with probiotic agents such as Lactobacillus rhamnosus, relapse is prevented [36]. Lactobacilli have also been shown to be effective in treating patients with chronic pouchitis after ileal pouch–anal anastomosis for the treatment of ulcerative colitis [37].
In addition to antibiotics and probiotic agents, other standard anti-inflammatory agents used in the long-term treatment of IBD patients are also active in the HLA-B27 transgenic rat. Both dexamethasone and prednisolone produce dose-dependent reductions in the inflammation in these animals [38,39]. As in patients with IBD, sulfasalazine at low doses is without effect in the HLA-B27 transgenic rat [40], but high doses do ameliorate the disease [41].
Three approaches have been used in our laboratory to show that the KKS has a major role in inflammatory arthritis and enterocolitis with the use of the PG-APS models. First, we used a specific oral reversible tight-binding active-site inhibitor of plasma kallikrein, D-Pro-Phe-boro-Arg. This specific kallikrein inhibitor attenuated acute inflammatory changes (edema, and neutrophil infiltration) and prevented arthritis and chronic systemic complications (splenomegaly, hepatomegaly, leukocytosis and the acute-phase reaction) in the PG-APS model [42]. The same plasma kallikrein inhibition modulated acute intestinal changes [28] as well as chronic granulomatous intestinal inflammation [29] similar to human Crohn's disease. Second, we showed that antagonists of BK receptor type 2 ameloriate acute arthritis [43] whereas an antagonist of BK receptor type 1 aggravated the joint inflammation [44]. We have recently shown that BK receptor antagonists can upregulate or downregulate specific cell-adhesion molecules [44]. Third, kininogen deficiency was first described in Brown Norway rats [45]. We introduced this mutation into a Lewis genetic background with five generations of backcrosses and showed that the deficiency of kininogen ameliorated acute and chronic enterocolitis [46]. Because we have previously successfully used the mAb C11C1 to inhibit tumor growth in a syngeneic murine model (Sainz IM, Isordia-Salas I, Pixley RA, Colman RW, unpublished work) and in a human colon carcinoma grown in a nude (immunodeficient) mouse model [47], we used this fourth approach in the present study. This antibody has recently been successfully employed in the PG-APS model in which mAb C11C1 inhibited inflammatory changes in joints, systemic inflammation, and activation of the kallikrein–kinin system [20]. Here we have demonstrated its efficiency in treating HLA-B27-associated inflammatory disease.
Each of the previous approaches to inhibiting the KKS to control inflammation was successful but had certain limitations. The plasma kallikrein active-site inhibitor displayed hepatic toxicity. The BK receptor antagonist had only a modest effect. Kininogen deficiency is rare in humans and is not really an applicable therapeutic modality. However, we were encouraged by the success of mAb C11C1 in the PG-APS model in the prevention of systemic and joint inflammation [20] and the lack of obvious side effects. The fact that antibodies against other inflammatory agonists have been used in the treatment of human IBD, arthritis and cancer make its use attractive. Until this study, mAb C11C1 had been used in a preventive mode. The HLA-B27 transgenic rat model permitted the rapid treatment of an established disease model. On the basis of these results, we suggest that mAb C11C1 might be a candidate for a therapeutic agent in human inflammatory disease.
Conclusion
We have assessed a transgenic rat model in which the human gene encoding HLA-B27 has been overexpressed. These rats developed T cell-mediated, spontaneous arthritis resembling reactive or inflammatory arthritis. We were able to successfully treat an established disease with an antibody against kininogen without inducing side effects or toxicity in either the rat or the mouse model of the disease.
Abbreviations
BK = bradykinin; H&E = hematoxylin and eosin; HK = high-molecular-mass kininogen; HKa = cleaved high-molecular-mass kininogen; HLA-B27 = human leukocyte antigen B27; IBD = inflammatory bowel disease; KKS = kallikrein–kinin system; mAb = monoclonal antibody; PG-APS = peptidoglycan–polysaccharide polymers from group A streptococci; PK = prekallikrein.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
JCK planned and supervised the entire animal protocol. He also participated in the statistical analysis and writing of the clinical results section and in the editing of the manuscript. IMS assessed the potential toxic effects of the treatment on kidney, lungs and liver. She also prepared the final version of all figures and collaborated in the statistical analysis, editing, and typing of the manuscript. IIS performed the KKS assays and, together with RAP, purified the antibody. RAP also participated in the statistical analysis, editing of the manuscript, and preparation of the KKS figure. YL performed the animal protocol and collected the data. LMA participated in the planning and execution of the animal project. RWC planned and initiated the entire product, wrote the introduction and discussion portions of the manuscript, and was responsible for final editing. All authors read and approved the final manuscript.
Acknowledgements
We thank Virginia Sheaffer for careful manuscript preparation, and Dr Ricardo Espinola for his technical support in this study. Grant support was received from the National Institutes of Health (grants R01 CA83121 and R01 AR051713) and the Broad Medical Research Program (IBD-0080R).
Figures and Tables
Figure 1 Kallikrein–kinin system (KKS). The KKS is initiated by factor XIIa (FXIIa) or prolylcarboxypeptidase on the endothelial cell and leukocyte (polymorphonuclear cell (PMN)) surface, generating the enzyme kallikrein, which in turn cleaves high-molecular-mass kininogen (HK) to yield bradykinin (BK) and cleaved high-molecular-mass kininogen (HKa). Kallikrein is chemotactic, aggregates neutrophils, and stimulates the release of elastase and superoxide (potent inducers of tissue injury). BK stimulates vasodilation, mediates pain through the release of prostaglandins, and stimulates vascular permeability through the generation of nitrous oxide (NO). PK, prekallikrein.
Figure 2 Effect of mAb C11C1 on HLA-B27 transgenic rats colonic inflammation. (a) Effects of monoclonal antibody (mAb) C11C1 on diarrhea in human leukocyte antigen B27 (HLA-B27) rats. Stool score was determined five times a week (normal stool = 1, soft stool = 2, watery stool = 3). mAb C11C1 (1.9 mg/kg) was administered three times a week for 16 days. The control group received murine isotype IgG1 (6 mg/kg) three times a week for 16 days. All stool scores are significantly different between the two groups for each corresponding day (P < 0.005) except for day 11 (P = 0.03). Data are shown as means ± SEM. Filled circles, IgG1-treated group; open circles, mAb C11C1-treated group. (b) Effects of mAb C11C1 on colonic mucosa in HLA-B27 rats. Photomicrographs of representative sections of colon from C11C1-treated (left) and IgG-treated (right) HLA-B27 transgenic rats. Note the extensive inflammatory cell infiltrates within the mucosa (a) and submucosa (b) with loss of villus formation on the mucosal surface indicated by the arrow (a) in the IgG group (right) compared with the C11C1 group (left). The branched arrow (left) points to the villus formation normally present in the colon (mAb C11C1-treated group). Hematoxylin and eosin stain; original magnification × 100. (c) Effects of mAb C11C1 on colonic inflammatory changes in HLA-B27 rats. mAb C11C1 decreased inflammatory changes in the colonic sections as evaluated by ulceration (P = 0.02), inflammation (P < 0.001), depth of lesion (P = 0.004), and degree of fibrosis replacement (P = 0.01) compared with IgG1 administration. Treatment with mAb C11C1 (open bars) significantly decreased the extent and intensity of the total colonic inflammatory score (P = 0.004). Data are shown as means ± SEM. *P < 0.05; ***P < 0.005.
Figure 3 Effect of mAb C11C1 on HLA-B27 transgenic rat inflammatory arthritis. (a) Effects of monoclonal antibody (mAb) C11C1 on clinical signs of arthritis in human leukocyte antigen B27 (HLA-B27) rats. mAb C11C1 was administered at the same dose and frequency as in Fig. 2a. Mean joint score was determined daily, except at weekends. All joint scores are significantly different between the two groups for each corresponding day (P < 0.001) except for days 1 (P > 0.03), 2 (P = 0.01) and 3 (P = 0.006). Data are shown as means ± SEM. Filled circles, IgG1-treated group; open circles, mAb C11C1-treated group. (b) Effects of mAb C11C1 on joint histology in HLA-B27 rats. Photomicrographs of representative sections of tarsal joints from C11C1-treated (left) and IgG-treated (right) HLA-B27 transgenic rats. Note the clear joint space (a) and normal appearance of bone (b) in the mAb C11C1-treated group (left) compared with the inflamed villus formation (arrows) occupying the synovial space (a) in the IgG-treated group (right). Hematoxylin and eosin stain; original magnification × 100. (c) Effects of mAb C11C1 on synovial inflammatory changes in HLA-B27 rats. Treatment with mAb C11C1 (open bars) decreased synovial proliferation (hyperplasia) (P = 0.01), subsynovial fibrosis (fibroplasia) (P = 0.001), and degree of inflammation (P < 0.001), but not pannus formation. The total score of the control IgG1 of 9.6 ± 1.0 was reduced by mAb C11C1 to an inflammatory score of 5.0 ± 1.0 (P = 0.009). Data are shown as means ± SEM. *P < 0.05; **P < 0.01; ***P < 0.005. (d) Effects of mAb C11C1 on cartilage and bone inflammatory changes in HLA-B27 rats. mAb C11C1 (open bars) significantly improved (decreased the Mankin score of) the cartilage organization (P = 0.01) and the altered chondrocyte proliferation (P = 0.008). The proteoglycan cartilage contents (Safranin O/Fast Green staining) were similar in both experimental groups (P > 0.05) and the tidemark integrity was preserved (data not shown). The total Mankin score was significantly decreased in the mAb C11C1-treated group (P = 0.02). Data are shown as means ± SEM. *P < 0.05; **P < 0.01; ***P < 0.005.
Figure 4 Kallikrein–kinin system (KKS) assays. Plasma KKS protein concentrations in the human leukocyte antigen B27 transgenic rats treated with control monoclonal antibody IgG (filled bars) or monoclonal antibody C11C1 (open bars) at day 16 of the experimental protocol. Values were compared with a pool of normal Fischer 344 rat plasma. Both high-molecular-mass kininogen (HK) and prekallikrein (PK) were significantly decreased in the IgG1-treated group and were closer to normal in the C11C1-treated group. Both experimental groups showed decreased factor XI (FXI) with no significant differences between them. There were no significant changes between any groups in factor XII (FXII). ***P < 0.005.
==== Refs
Taurog JD Richardson JA Croft JT Simmons WA Zhou M Fernandez-Sueiro JL Balish E Hammer RE The germfree state prevents development of gut and joint inflammatory disease in HLA-B27 transgenic rats J Exp Med 1994 180 2359 2364 7964509 10.1084/jem.180.6.2359
Hammer RE Maika SD Richardson JA Tang JP Taurog JD Spontaneous inflammatory disease in7 transgenic rats expressing HLA-B27 and human beta 2m: an animal model of HLA-B27-associated human disorders Cell 1990 63 1099 1112 2257626 10.1016/0092-8674(90)90512-D
Sartor RB Rath HC Lichtman SN van Tol EA Animal models of intestinal and joint inflammation Baillieres Clin Rheumatol 1996 10 55 76 8674149
Breban M HLA-B27 transgenic rats model Ann Med Interne (Paris) 1998 149 139 141 11490535
Joseph K Shibayama Y Ghebrehiwet B Kaplan AP Factor XII-dependent contact activation on endothelial cells and binding proteins gC1qR and cytokeratin 1 Thromb Haemost 2001 85 119 124 11204562
Shariat-Madar Z Mahdi F Schmaier AH Identification and characterization of prolylcarboxypeptidase as an endothelial cell prekallikrein activator J Biol Chem 2002 277 17962 17969 11830581 10.1074/jbc.M106101200
Chavakis T Santoso S Clemetson KJ Sachs UJ Isordia-Salas I Pixley RA Nawroth PP Colman RW Preissner KT High molecular weight kininogen regulates platelet-leukocyte interactions by bridging Mac-1 and glycoprotein Ib J Biol Chem 2003 278 45375 45381 12952972 10.1074/jbc.M304344200
Colman RW Plasma and tissue kallikrein in arthritis and inflammatory bowel disease Immunopharmacology 1999 43 103 108 10596840 10.1016/S0162-3109(99)00068-5
Schapira M Despland E Scott CF Boxer LA Colman RW Purified human plasma kallikrein aggregates human blood neutrophils J Clin Invest 1982 69 1199 1202 6917855
Wachtfogel YT Kucich U James HL Scott CF Schapira M Zimmerman M Cohen AB Colman RW Human plasma kallikrein releases neutrophil elastase during blood coagulation J Clin Invest 1983 72 1672 1677 6556194
Colman RW Pixley RA Sainz I Song JS Isordia-Salas I Muhamed SN Powell JA JrMousa SA Inhibition of angiogenesis by antibody blocking the action of proangiogenic high-molecular-weight kininogen J Thromb Haemost 2003 1 164 170 12871554 10.1046/j.1538-7836.2003.00025.x
Regoli D Barabe J Pharmacology of bradykinin and related kinins Pharmacol Rev 1980 32 1 46 7015371
Liu B Li F Isordia-Salas I Bradford HN Colman RW Pixley RA Sartor RB Cleavage products of kininogen variably induce proinflammatory cytokines by Fischer rat splenocytes [Abstract] Gastroenterology 2004 126 A573
Khan MM Bradford HN Isordia-Salas I Espinola R Colman RW High molecular weight kininogen fragments stimulate the secretion of interleukin 1b through urokinase-type plasminogen activator receptor (uPAR) and CD11b/CD18 (Mac-1) in human blood mononuclear cells [abstract] Blood 2004 104 36b
DeLa Cadena RA Laskin KJ Pixley RA Sartor RB Schwab JH Back N Bedi GS Fisher RS Colman RW Role of kallikrein–kinin system in pathogenesis of bacterial cell wall-induced inflammation Am J Physiol 1991 260 G213 G219 1996642
Sartor RB DeLa Cadena RA Green KD Stadnicki A Davis SW Schwab JH Adam AA Raymond P Colman RW Selective kallikrein–kinin system activation in inbred rats differentially susceptible to granulomatous enterocolitis Gastroenterology 1996 110 1467 1481 8613052
Isordia-Salas I Pixley RA Parekh H Kunapuli SP Li F Stadnicki A Lin Y Sartor RB Colman RW The mutation Ser511Asn leads to N-glycosylation and increases the cleavage of high molecular weight kininogen in rats genetically susceptible to inflammation Blood 2003 102 2835 2842 12842992 10.1182/blood-2003-02-0661
Blais C JrCouture R Drapeau G Colman RW Adam AA Involvement of endogenous kinins in the pathogenesis of peptidoglycan-induced arthritis in the Lewis rat Arthritis Rheum 1997 40 1327 1333 9214434
Stadnicki A Sartor RB Janardham R Stadnicka I Adam A Blais C Colman RW Kallikrein–kininogen system activation and bradykinin (B2) receptors in indomethacin-induced enterocolitis in genetically susceptible Lewis rats Gut 1998 43 365 374 9863482
Espinola RG Uknis A Sainz IM Isordia-Salas I Pixley RA DeLa Cadena R Long W Agelan A Gaughan J Adam A A monoclonal antibody to high molecular weight knininogen is therapeutic in a rodent model of reactive arthritis Am J Pathol 2004 165 969 976 15331420
Keith JC JrAlbert L Sonis ST Pfeiffer CJ Schaub RG IL-11, a pleiotropic cytokine: exciting new effects of IL-11 on gastrointestinal mucosal biology Stem Cells 1994 12 79 89 7696971
Harris HA Albert LM Leathurby Y Malamas MS Mewshaw RE Miller CP Kharode YP Marzolf J Komm BS Winneker RC Evaluation of an estrogen receptor-beta agonist in animal models of human disease Endocrinology 2003 144 4241 4249 14500559 10.1210/en.2003-0550
Peterson RL Wang L Albert L Keith JC JrDorner AJ Molecular effects of recombinant human interleukin-11 in the HLA-B27 rat model of inflammatory bowel disease Lab Invest 1998 78 1503 1512 9881950
Boughton-Smith NK Wallace JL Morris GP Whittle BJ The effect of anti-inflammatory drugs on eicosanoid formation in a chronic model of inflammatory bowel disease in the rat Br J Pharmacol 1988 94 65 72 2840994
Poole AR Coombs RR Rheumatoid-like joint lesions in rabbits injected intravenously with bovine serum Int Arch Allergy Appl Immunol 1977 54 97 113 863527
Mankin HJ Dorfman H Lippiello L Zarins A Biochemical and metabolic abnormalities in articular cartilage from osteo-arthritic human hips. II. Correlation of morphology with biochemical and metabolic data J Bone Joint Surg Am 1971 53 523 537 5580011
Scott CF Silver LD Purdon AD Colman RW Cleavage of human high molecular weight kininogen by factor XIa in vitro. Effect on structure and function J Biol Chem 1985 260 10856 10863 3875612
Stadnicki A DeLa Cadena RA Sartor RB Bender D Kettner CA Rath HC Adam A Colman RW Selective plasma kallikrein inhibitor attenuates acute intestinal inflammation in Lewis rat Dig Dis Sci 1996 41 912 920 8625762
Stadnicki A Sartor RB Janardham R Majluf-Cruz A Kettner C Adam AA Colman RW Specific inhibition of plasma kallikrein modulates chronic granulomatous intestinal and systemic inflammation in genetically susceptible rats FASEB J 1998 12 325 333 9506476
Harnish DC Albert LM Leathurby Y Eckert AM Ciarletta A Kasaian M Keith JC Jr Beneficial effects of estrogen treatment in the HLA-B27 transgenic rat model of inflammatory bowel disease Am J Physiol Gastrointest Liver Physiol 2004 286 G118 G125 12958017 10.1152/ajpgi.00024.2003
Peterson RL Wang L Albert L Marchese E Erickson J Wong A Mounts WM Hayes L Bouchard P Keith J Pharmacogenomic analysis of rhIL-11 treatment in the HLA-B27 rat model of inflammatory bowel disease Pharmacogenomics J 2002 2 383 399 12629504 10.1038/sj.tpj.6500137
Greenwood-Van Meerveld B Venkova K Keith JC Jr Recombinant human interleukin-11 restores smooth muscle function in the jejunum and colon of human leukocyte antigen-B27 rats with intestinal inflammation J Pharmacol Exp Ther 2001 299 58 66 11561063
Albert LM Patel U Ferranti TJ Ahmed A Schaub RG Keith JCJ Soluble recombinant human p-selectin glycoprotein ligand-1 (sPSGL-1) decreases intestinal lesions in HLA-B27 transgenic rats and in acetic acid induced colitis in Sprague-Dawley rats [Abstract] Am Gastroenterology Soc 1998 114 A918
Chadwick CC Chippari S Matelan E Borges-Marcucci L Eckert AM Keith JC JrAlbert LM Leathurby Y Harris HA Bhat RA Identification of pathway-selective estrogen receptor ligands that inhibit NF-κB transcriptional activity PNAS 2005 102 2543 2548 15699342 10.1073/pnas.0405841102
Dieleman LA Hoentjen F Qian BF Sprengers D Tjwa E Torres MF Torrice CD Sartor RB Tonkonogy SL Reduced ratio of protective versus proinflammatory cytokine responses to commensal bacteria in HLA-B27 transgenic rats Clin Exp Immunol 2004 136 30 39 15030511 10.1111/j.1365-2249.2004.02410.x
Dieleman LA Goerres MS Arends A Sprengers D Torrice C Hoentjen F Grenther WB Sartor RB Lactobacillus GG prevents recurrence of colitis in HLA-B27 transgenic rats after antibiotic treatment Gut 2003 52 370 376 12584218 10.1136/gut.52.3.370
Gionchetti P Rizzello F Venturi A Brigidi P Matteuzzi D Bazzocchi G Poggioli G Miglioli M Campieri M Oral bacteriotherapy as maintenance treatment in patients with chronic pouchitis: a double-blind, placebo-controlled trial Gastroenterology 2000 119 305 309 10930365
Kim YS Son M Ko JI Cho H Yoo M Kim WB Song IS Kim CY Effect of DA-6034, a derivative of flavonoid, on experimental animal models of inflammatory bowel disease Arch Pharm Res 1999 22 354 360 10489873
Blanchard HS Dernis-Labous E Lamarque D Nhieu JT Szepes Z Flejou JF Wollman E Whittle BJ Breban M Inducible nitric oxide synthase attenuates chronic colitis in human histocompatibility antigen HLA-B27/human beta2 microglobulin transgenic rats Eur Cytokine Netw 2001 12 111 118 11282554
Aiko S Fuseler J Grisham MB Effects of nitric oxide synthase inhibition or sulfasalazine on the spontaneous colitis observed in HLA-B27 transgenic rats J Pharmacol Exp Ther 1998 284 722 727 9454820
Kerr SW Wolyniec WW Filipovic Z Nodop SG Braza F Winquist RJ Noonan TC Repeated measurement of intestinal permeability as an assessment of colitis severity in HLA-B27 transgenic rats J Pharmacol Exp Ther 1999 291 903 910 10525115
DeLa Cadena RA Stadnicki A Uknis AB Sartor RB Kettner CA Adam A Colman RW Inhibition of plasma kallikrein prevents peptidoglycan-induced arthritis in the Lewis rat FASEB J 1995 9 446 452 7896018
Uknis AB DeLa Cadena RA Janardham R Sartor RB Whalley ET Colman RW Bradykinin receptor antagonists type 2 attenuate the inflammatory changes in peptidoglycan-induced acute arthritis in the Lewis rat Inflamm Res 2001 50 149 155 11339503 10.1007/s000110050739
Sainz IM Uknis AB Isordia-Salas I Dela Cadena RA Pixley RA Colman RW Interactions between bradykinin (BK) and cell adhesion molecule (CAM) expression in peptidoglycan-polysaccharide (PG-PS)-induced arthritis FASEB J 2004 18 887 889 15001555
Hayashi I Hoshiko S Makabe O Oh-ishi S A point mutation of alanine 163 to threonine is responsible for the defective secretion of high molecular weight kininogen by the liver of brown Norway Katholiek rats J Biol Chem 1993 268 17219 17224 8349607
Isordia-Salas I Pixley RA Li F Sainz I Sartor RB Adam A Colman RW Kininogen deficiency modulates chronic intestinal inflammation in genetically susceptible rats Am J Physiol Gastrointest Liver Physiol 2002 283 G180 G186 12065305
Song JS Sainz I Cosenza SC Isordia-Salas I Bior AD Bradford HN Guo YL Pixley RA Reddy EP Colman RW Inhibition of tumor angiogenesis in vivo by monoclonal antibody targeted to domain 5 of high molecular weight kininogen Blood 2004 104 2065 2072 15161672 10.1182/blood-2004-02-0449
| 15987478 | PMC1175023 | CC BY | 2021-01-04 16:02:38 | no | Arthritis Res Ther. 2005 Apr 4; 7(4):R769-R776 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1728 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17301598747710.1186/ar1730Research ArticleChondrocytes, synoviocytes and dermal fibroblasts all express PH-20, a hyaluronidase active at neutral pH El Hajjaji Hafida 1Cole Ada Asbury [email protected] Daniel-Henri [email protected] Christian de Duve Institute of Cellular Pathology, Department of Biochemistry, Connective Tissue Group, Université Catholique de Louvain in Brussels, Brussels, Belgium2 Department of Biochemistry, Rush Medical College, Rush-Presbyterian-St. Luke's Medical Center, Chicago, IL, USA3 Department of Rheumatology, Saint Luke's University Hospital, Catholic University of Louvain in Brussels, Brussels, Belgium2005 4 4 2005 7 4 R756 R768 18 3 2004 8 4 2004 21 2 2005 7 3 2005 Copyright © 2005 Hajjaji et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Hyaluronan (HA), an important component of connective tissues, is highly metabolically active, but the mechanisms involved in its catabolism are still largely unknown. We hypothesized that a protein similar to sperm PH-20, the only mammalian hyaluronidase known to be active at neutral pH, could be expressed in connective tissue cells. An mRNA transcript similar to that of PH-20 was found in chondrocytes, synoviocytes, and dermal fibroblasts, and its levels were enhanced upon stimulation with IL-1. In cell layers extracted with Triton X-100 – but not with octylglucoside – and in culture media, a polyclonal antipeptide anti-PH-20 antibody identified protein bands with a molecular weight similar to that of sperm PH-20 (60 to 65 kDa) and exhibiting a hyaluronidase activity at neutral pH. Further, upon stimulation with IL-1, the amounts of the neutral-active hyaluronidase increased in both cell layers and culture media. These findings contribute potential important new insights into the biology of connective tissues. It is likely that PH-20 facilitates cell-receptor-mediated uptake of HA, while overexpression or uncontrolled expression of the enzyme can cause great havoc to connective tissues: not only does HA fragmentation compromise the structural integrity of tissues, but also the HA fragments generated are highly angiogenic and are potent inducers of proinflammatory cytokines. On the other hand, the enzyme activity may account for the progressive depletion of HA seen in osteoarthritis cartilage, a depletion that is believed to play an important role in the apparent irreversibility of this disease process.
==== Body
Introduction
Hyaluronan (HA), a linear, megadalton glycosaminoglycan, has important biological and structural functions [1]. First, by interacting with transmembrane proteins, such as CD44 and other members of the heterogeneous group of proteins termed hyaladherins, HA initiates signaling pathways and contributes to the formation of the pericellular matrix that prevents direct contact between cells and protects them against attack from viruses, bacteria, and immune cells. Second, in the extracellular matrix further removed from the cell and in the basement membrane, the hydrophilic HA network not only gives turgor pressure and resilience, but also functions as a scaffold about which other macromolecules associate and orient themselves [2-4]. Within the abundant extracellular matrix of articular cartilage, the long, filamentous HA molecules form the backbone upon which the viscoelastic aggrecan molecules align to form aggregates, a supramolecular organization that immobilizes aggrecans at very high concentrations within the collagen network, thereby providing remarkable biomechanical properties to the articular tissue [5]. Third, in its unaggregated form, HA is the major macromolecular species in synovial fluid, being thereby responsible for the viscoelastic properties of what is otherwise a simple plasma diffusate [6].
On the other hand, because HA degradation products may interact with various cells and initiate a program of gene expression leading to cell proliferation, migration, or activation [3,4,7], these products exhibit biological functions that are quite distinct from those of the native, high-molecular-weight polymer. Thus, by stimulating the proliferation and migration of vascular endothelial cells via multiple signaling pathways, HA fragments induce angiogenesis, whereas high-molecular-weight HA inhibits angiogenesis [8,9]. Studies in vitro have also indicated that HA fragments similar in size to that of fragmented HA molecules found in vivo in inflammatory sites induce the expression of inflammatory genes in dendritic cells, macrophages, eosinophils, and certain epithelial cells [7]; this effect is in contrast to that of high-molecular-weight HA molecules, which inhibit the production of IL-1, prostaglandin E2, and matrix metalloproteinases [10-12]. Further, HA depolymerization, such as occurs in osteoarthritis and other arthritides, compromises the biomechanical properties of diarthrodial joints and thereby contributes to joint destruction. Indeed, the synovial fluid of diseased joints contains smaller HA molecules, which dramatically reduce the lubricating properties of the joint fluid [13]. On the other hand, upon stimulation with IL-1, the HA molecules in articular cartilage explants are fragmented and lost into the conditioned medium [14], a finding which implies that in the presence of this cytokine the viscoelastic aggrecan molecules are no longer firmly entrapped within the collagen network of the articular tissue.
Since the fragmentation of HA compromises the integrity of tissues, it is obviously important to investigate the mechanisms involved in the production of its fragments. Although oxygen-derived free radicals are known to fragment HA randomly [15], one cannot exclude the possible role of hyaluronidases, because mammalian hyaluronidases, in contrast to bacterial ones, yield a heterogeneous mixture of oligosaccharides and HA fragments of various sizes.
Thus far, the only human hyaluronidase known to be active at neutral pH is the sperm adhesion molecule 1, also termed PH-20, which enables the sperm to penetrate the HA-rich cumulus oophorus [16,17]. Therefore, and although PH-20 mRNA has been detected only in testis, mouse kidneys, some cancer cell lines, and fetal and placenta cDNA libraries [18-21], we hypothesized that connective tissue cells may express a hyaluronidase similar to the one present on sperm. Herein we provide evidence that, at both the mRNA and protein levels, fibroblasts, chondrocytes, and synoviocytes all express a neutral, active hyaluronidase similar to sperm PH-20.
Materials and methods
Culture media and reagents
The following reagents were used: DMEM, FCS, penicillin, streptomycin, the PCR TOPO vector, and the SuperScript Reverse Transcriptase (Invitrogen, Merelbeke, Belgium); the Complete Protease Inhibition Cocktail, the Tripure isolation kit, collagenase P, n-octylthioglucoside, recombinant human IL-1β (IL-1; specific activity 5 × 107 units/mg) (Roche Applied Science, Brussels, Belgium); radiolabelled precursors, Hybond-N+ membranes, CNBr-activated Sepharose, EAHSepharose, Megaprime DNA labeling system, and restriction enzymes (Amersham, Roosendaal, the Netherlands); standards and reagents for protein electrophoresis (Bio-Rad laboratories, Nazareth, Belgium); Pierce BCA protein reagent, Pierce Immuno-Pure Gentle Ag/Ab Elution buffer, and Clontech Advantage PCR cDNA mix kit (Perbio Science, Erembodegem, Belgium); DNA molecular-weight markers and custom-made primers (Eurogentec, Seraing, Belgium); the pGEM-T Easy Vector kit (Promega Benelux, Leiden, the Netherlands); Qiagen kits for plasmid purification (Westburg, Venlo, the Netherlands); and BSA, high-purity fraction V, from Calbiochem (VWR International, Leuven, Belgium). Unless specified, all other chemicals used were of analytical grade and obtained from Sigma-Aldrich-Fluka (Bornem, Belgium).
Tissue acquisition and cell culture
Human chondrosarcoma cells (SW 1353) were purchased from the American Tissue Culture Collection (LGC Promochem, Teddington, UK). Tissue sampling was conducted in accordance with the rules and regulations of the Saint Luc University Hospital ethics committee (UCL, Brussels). Foreskin was obtained from newborn babies undergoing surgery for phymosis. Human articular cartilage and synovium were harvested from patients undergoing either leg amputation or arthroplasty of the knee joint. Dermal fibroblasts and synoviocytes were isolated by sequential enzymatic digestion with 0.15% collagenase P and 0.03% trypsin, and the cells were used routinely at passage numbers 4 to 7. Chondrocytes were isolated [22] and used at passage number 2 or 3.
Cells were cultured in DMEM supplemented with 10% heat-inactivated FCS, penicillin (50 units/ml), and streptomycin (50 μg/ml). At 2 days before cell sampling and/or cell stimulation by cytokines, FCS was replaced by ITS (insulin, transferrin, and sodium selenite) and 0.5% BSA. In contrast to other tested preparations, the BSA (fraction V, high purity) that we routinely used from Calbiochem did not contain any hyaluronidase activity as assessed by HA substrate gel electrophoresis.
At the end of the culture period, conditioned media were removed and enriched with Triton X-100 (1% final concentration) and a complete protease inhibitor cocktail before further processing. The Tripure solution was then directly added to the culture dish kept on ice to extract the total cellular RNA. In parallel experiments, 50 mM sodium phosphate, pH 7.4, containing the cocktail of protease inhibitors as well as either 1% Triton X-100 or 20 mM octylthioglucoside was added to culture dishes kept on ice to extract proteins from cell layers. Adherent cells were scraped off with a rubber scraper and further incubated with the extracting solution on ice for 15 min with intermittent vortexings. At the end of the incubation period, the solution was centrifuged and the supernatant was either used immediately or stored at -20°C in glycerol until use. The pellet represented a very small proportion of the original cells.
Reverse transcriptase polymerase chain reaction
For reverse transcription reactions, about 250 ng of total RNA was converted into first-strand cDNA using SuperScript Reverse Transcriptase in accordance with the manufacturer's instructions, in a final volume of 20 μl. The target cDNA was then amplified by using the Advantage cDNA PCR kit, a reaction that included 2 μl of the cDNA synthesis reaction. After 35 cycles at 94°C for 30 s, 64°C for 1 min, and 72°C for 5 min, amplification products were detected by electrophoresis in 1% agarose gel containing 0.5 μg/ml ethidium bromide, into which 10 μl of sample per lane was injected. DNA molecular-weight markers were included on each gel for sizing.
Comparison between the sequence of testis PH-20 cDNA (GenBank accession number S67798) and the sequence of chromosome 7q31.3 (4676254 to 4710990) reveals that sperm PH-20 cDNA contains apparently four exons, the coding sequence starting in exon 2 and ending in exon 4. Therefore, to avoid genomic DNA contamination, two sense primers in exon 2 and two antisense primers in exon 4 were chosen. The first set of primers (primers A) consisted of 5'-CCA-TGTTGCTTGACTCTGAATTTCA-3' (oligo sense) and 5'-CCGAACTCGATTGCGCACATAGAGT-3' (oligo antisense), with an expected RT-PCR product of 759 bp. The second set of primers (primers B) consisted of 5'-GCCTGGAATGCCCCAAGTGA-3' (oligo sense) and 5'-TCCTTGCTCCTGGCAAAGCAC-3' (oligo antisense), with an expected RT-PCR product of 1,000 bp. A third set of primers (primers C) was chosen both in exon 4: 5'-TGCTTTGCCAGGAGCAAGGA-3' (sense primer) and 5'-CCTGCGCAATTACAAACT-CGCTACA-3' (oligo antisense), with an expected RT-PCR product of 403 bp. Primers used for the housekeeping gene β-actin were 5'-TGATGGTGGGCATGGGTCAG-3' (oligo sense) and 5'-TCTTCTCGCGGTTGGCCTTG-3' (oligo antisense), with an expected RT-PCR product of 226 bp.
Cloning and sequencing of PCR products
Aliquots of RT-PCR products were cloned in the pGEM-T Easy Vector. Recombinant plasmids were isolated and sequenced by using the Big Dye Terminator Cycle sequencing kit from Applied Biosystems (Lennik, Belgium) and an automated sequencer (ABI Prism model 310).
Northern blot analysis
As the 1,000-bp RT-PCR product amplified by PH-20 primers B does not contain any EcoRI restriction sites, this product was cloned into the PCR TOPO vector that exhibits EcoRI sites flanking the PCR product insertion site. The approximately 1,000-bp EcoRI restriction fragment was radiolabelled with [α-32P]dCTP by random priming (Megaprime system) and used as a PH-20 probe.
Intact RNA samples (25 μg) were subjected to electrophoresis in 1% agarose/formaldehyde gels and transferred by capillarity to Hybond-N+ membranes prior to fixation by UV cross-linking. After a prehybridization step for 4 hours in 6 × SSC (1 × SSC is 0.15 M NaCl, 0.015 M sodium citrate, pH 7.0), 5 × Denhardt's solution, 0.5% SDS, and 20 μg/ml denatured Salmon sperm DNA (Invitrogen, Merelbeke, Belgium), the membranes were hybridized overnight at 64°C in a similar solution containing either the PH-20 or the actin probe. At the end of the hybridization procedure, membranes were washed at 64°C (2 × SCC, 3 × 10 min; 2 × SSC/0.1% SDS, 1 × 10 min; 2 × SSC, 1 × 10 min; 0.2 × SSC, 1 × 10 min), dried, and then exposed to x-ray film for 12 hours (β-actin) or 48 hours (PH-20).
In separate experiments, 20 μg of total RNA from cells either nonstimulated or stimulated with IL-1 at a concentration of 5 or 10 ng/ml were subjected to northern blotting as stated above and the dried membranes were exposed to a PhosphorImager screen (Molecular Dynamics, Sunnyvale, CA, USA) for the detection and quantification of the radioactive signals. As the 1.4-kb transcript represented approximately 10% of the total radioactivity, quantification was restricted to the 2.4-kb transcript. Results obtained for stimulated cells were normalized with values obtained for the corresponding unstimulated control cells, to which a value of 1 arbitrary unit was assigned. In each set of experiments, three Petri dishes (79 cm2 each) containing cells at near confluence were used for each condition. Data represent the means ± standard deviations obtained for five different sets of experiments.
Antibody production and PH-20 purification
In accordance with the sequence of human PH-20 (Swissprot: locus HYAP_HUMAN, accession P38567), the peptide NH2-CAR NWK PKD VYK NRS I-CONH2 (amino acids 155 to 169) was chosen. The sequence of its first 10 amino acids is also present in bovine PH-20. Further comparison between the sequence of the immunizing peptide and that of other known human proteins ('blast', National Institure of Health) also reveals that both hyaluronidases 2 [23] and 4 [20] exhibit two stretches of three to four identical amino acids separated by different amino acids, whereas hyaluronidase 1 [24] contains three stretches of two identical amino acids. After synthesis, the 15-mer oligopeptide was coupled to keyhole limpet hemocyanin and injected into rabbits for immunization. After the peptide had been coupled to EAHSepharose (efficiency 98%), the peptide–gel complex was incubated with the serum and then washed extensively with Tris-buffered saline. The specific antibodies were eluted with 1.5 M guanidine hydrochloride and stored in PBS containing 0.01% sodium azide and 1% BSA.
Purified antibodies were covalently coupled to CNBr-activated Sepharose in accordance with the manufacturer's instructions (3 mg IgG per gram dry gel). Cell-layer extracts and conditioned media containing a cocktail of protease inhibitors were applied to columns packed with the immunoaffinity gel, which was then washed with 20 column volumes of PBS containing 0.1% Triton X-100 before eluting the bond proteins with either 50 mM glycine–HCl, pH 3.0, or the Pierce Immuno-Pure 'gentle' Ag/Ab elution buffer. Fractions eluted with the glycine–HCl buffer were neutralized immediately by adding 0.1 volume of 1 M Tris–HCl, pH 8.0. The BCA assay was used to determine protein concentrations.
The immunoaffinity gel was also used to estimate the changes in the amounts of PH-20 proteins present in cell layers and conditioned media of the three cell lines upon stimulation with IL-1. For each cell line and in each experiment, three Petri dishes (79 cm2 each) containing cells at near confluence were used for each condition. The data reported are means ± standard deviations obtained for three to five sets of experiments.
Sodium dodecyl sulfate–polyacrylamide gel electrophoresis
After being concentrated by ultrafiltration, proteins eluted from the immunoaffinity gel were characterized by Tricine (N-tris [(hydroxymethyl)methyl]glycine) SDS–PAGE in accordance with the procedure of Schagger and von Jagow [25]. Gels were fixed in 50% methanol and 10% acetic acid before being stained with 0.025% Serva Blue G in 10% acetic acid.
Assays for hyaluronidase activity
In a first set of studies, hyaluronidase activity was detected by HA substrate gel electrophoresis [26]. Briefly, after electrophoresis at 4°C (20 mA/gel), gels were washed twice for 20 min in 2.5% Triton X-100 before overnight incubation in 0.15 M sodium chloride, 0.5 mM calcium chloride, 7 mM 1,4-saccharolactone buffered with either 0.1 M acetate, pH 4.5, or 0.1 M Mes (2-(N-Morpholino)ethane sulfonic acid), pH 6.5. When required, apigenin dissolved in dimethylsulfoxide was added to the incubation solution to give a final concentration of 5 μg/ml. Control gels were incubated with a similar volume of dimethylsulfoxide. After incubation, gels were washed twice for 30 min in water and further incubated at 37°C with proteinase K to remove proteins that may interfere with gel staining with alcian blue. The presence of bands without staining on the blue background of undegraded HA indicates hyaluronidase activity.
Hyaluronidase activity was also assessed by using high-molecular-weight radiolabelled HA (5 × 106 dpm (disintegrations per minute) per milligram). To generate this substrate, preconfluent synoviocytes were incubated in the presence of 2-acetamido-2-deoxy-D-glucurono-1,5-lactone (50 mM) and tritium-labelled glucosamine (50 μCi/ml); after digestion with proteinase K and precipitation in ethanol, the glycosaminoglycans were passed through a DEAESepharose column equilibrated in 50 mM pyridine, pH 5.5 (buffer A), and the high-molecular-weight tritium-labelled HA molecules were eluted at a NaCl concentration of 0.4 M. Assays were conducted in duplicate as follows: specimens and known relative turbidity-reducing units (rTRUs) of testicular hyaluronidase were incubated with purified tritium-labelled HA molecules (5 to 8 μg, 4 × 104 dpm) in either 0.1 M Hepes (4-(2-hydroxyethyl)-1-piperazine-ethanesulfonic acid) (pH 7 to 8), 0.1 M Mops (3-(N-morpholino)propanesulfonic acid) (pH 6.5 to 7.5), 0.1 M Mes (pH 5.5 to 6.5), or 0.1 M acetate (pH 4.5 to 5.5) containing a final concentration of 0.15 M NaCl, 7 mM 1,4-saccharolactone, and 1% Triton X-100. After incubations, specimens were heated at 95°C for 5 min, diluted with 10 volumes of buffer A, and applied on 1 ml DEAESepharose gel equilibrated in the same buffer. Gels were then washed with 10 volumes of buffer A before being eluted step by step with increasing amounts of NaCl dissolved in the same buffer: 10 volumes of 0.2 M NaCl for step 1; 10 volumes of 0.3 M NaCl for step 2; 10 volumes of 0.4 M NaCl for step 3, and 3 volumes of 0.8 M NaCl for step 4. More than 95% of the radioactivity applied onto the gel was recovered in steps 1 to 3. HA disaccharides produced by chondroitinase ABC elute at 0.2 M NaCl, the HA tetrasaccharides and hexasaccharides produced by testicular hyaluronidase elute at 0.2 and 0.3 M NaCl, and the intact HA molecules elute at 0.4 M NaCl. For amounts of testicular hyaluronidase ranging between 0.005 and 0.05 rTRUs, there is a highly statistically significant linear correlation (r = 0.9) between, on the one hand, the enzyme activity and, on the other hand, the % of radioactivity recovered in steps 2 and 3 (range 10 to 45%). The intra-assay and interassay variations are less than 5 and 10%, respectively.
Results
mRNA for PH-20 is present in chondrocytes and other connective tissue cell lines
RT-PCR was used to investigate whether PH-20 mRNA was present in human chondrocytes, fibroblasts, or synoviocytes and also in a chondrosarcoma cell line (Fig. 1). Primers A amplified an expected RT-PCR product of 759 bp in both testis (lane 1) and chondrocytes (lane 2), but the amplified product was more abundant in testis than in chondrocytes. RT-PCR products of expected size were also amplified by using either PH-20 primers B (1,000 bp: lanes 5, 8, and 10), PH-20 primers C (403 bp; lanes 4 and 7) or actin primers (226 bp; lanes 6, 9, and 11). In all the cell lines tested, the band obtained for the housekeeping gene β-actin was stronger than that observed for PH-20.
After purification, these products were ligated into the pGEM-T vector and sequenced on both strands. The sequence obtained for each cell line was identical to the sequence of human testis PH-20.
Northern blotting was used to further characterize the PH-20 mRNA (Fig. 2a). As observed in testis (lane 1), the probe detected a strong 2.4-kb mRNA band in synoviocytes (lane 3), chondrocytes (lane 4), fibroblasts (lane 7), and a chondrosarcoma cell line (lane 8), whereas no signal was detected in the RNA extracted from liver (lane 2), a tissue reported as not expressing PH-20 [27]. The PH-20 probe also detected a fainter, 1.4-kb band whose abundance was somewhat related to the abundance of the major 2.4-kb band. Thus far, it is not clear whether the smaller transcript indicates alternative splicing of PH-20 mRNA or reflects the existence of another potential hyaluronidase. This notwithstanding, it is worth noting that the signal corresponding to the other known hyaluronidases (1, 2, 3, and 4) is greater than 2 kb [20] and that PH-20 RT-PCR conducted either with sense primer located in exon 1 and antisense primer located in exon 3 or with sense primer located in exon 2 and antisense primer in exon 4 always gave a product of expected size (not shown).
Because HA depolymerization is consistently observed in inflammatory sites [7], and as the HA molecules present in articular cartilage explants are fragmented and lost into the conditioned medium upon stimulation with IL-1 [14], we examined the effect of this proinflammatory cytokine on PH-20 mRNA levels in chondrocytes: the band intensity was enhanced at IL-1 concentrations of both 5 ng/ml (lane 5) and 10 ng/ml (lane 6). The intensity of the 2.4-kb transcript was semiquantified by using the PhosphoImager (Fig. 2c): the intensities increased (mean ± standard deviation) by a factor of 1.7 ± 0.2 (n = 5) at 5 ng/ml IL-1 and 2.4 ± 0.3 (n = 5; P = 0.0004) at 10 ng/ml IL-1. Concentration of 10 ng/ml.
Evidence for the presence of the PH-20 protein in cell-associated extracts and conditioned media
Because, thus far, chondrocyte and fibroblast cell layers extracted with octylglucoside have been reported to contain a hyaluronidase activity that cannot be detected above pH 5 [22,28], a preparation of fibroblast cell layers extracted with octylglucoside was subjected to HA substrate gel electrophoresis (Fig. 3). As observed by Stair-Nawy and colleagues [28], the fibroblast preparation (lanes 1) exhibited a clear-cut band of activity with an apparent molecular weight of approximately 55 kDa at pH 4, but no activity was detected at pH 6.5. Likewise, a preparation of liver lysosomes used as positive control (lanes 2) was very active at pH 4, with a major band at approximately 50 kDa and a minor band at approximately 120 kDa, but was totally inactive at pH 6.5. On the other hand, a commercial preparation of bovine testicular hyaluronidase (lane 3) exhibited several bands of activity at near neutral pH. Therefore, we hypothesized that octylglucoside was apparently unable to extract from cell layers a neutral-active hyaluronidase, or at least not in amounts sufficient to be detected by HA substrate gel electrophoresis.
Because obviously fibroblasts expressed PH-20 at the mRNA level, we used the anti-PH-20 antibodies to test whether these cells expressed the neutral-active hyaluronidase at the protein level. After having extracted fibroblast cell layers with either octylglucoside or Triton X-100, a detergent used to solubilize sperm PH-20 [17], specimens from both extracts (approximately 3 mg proteins) were applied to columns packed with anti-PH-20-antibodies–Sepharose. Fractions eluted with the glycine–HCl buffer were concentrated by ultracentrifugation and analyzed using SDS-PAGE (Fig. 4a). A strong band (lane 1) with the apparent molecular weight of 65 kDa reported for sperm PH-20 [29] was present in specimens extracted with Triton X-100, whereas a faint band with a similar molecular weight could be barely detected in specimens extracted with octylglucoside (lane 2). These results confirmed that a substantial population of proteins recognized by anti-PH-20 antibodies and having a molecular weight similar to that of the sperm hyaluronidase are extractable by Triton X-100 but not by octylglucoside [17,29].
In parallel experiments, chondrocytes were cultured in the absence and in the presence of IL-1 (5 ng/ml). Cell layers extracted with Triton X-100 as well as their corresponding culture media were allowed to interact with the immunoaffinity gel, and proteins eluted with the glycine–HCl buffer were subjected to SDS-PAGE (Fig. 4b). Specimens from unstimulated and stimulated conditioned media (lanes 1 and 2, respectively) as well as extracts from unstimulated and stimulated cell layers (lanes 3 and 4, respectively) all exhibited several bands ranging from approximately 60 to 65 kDa. This range of molecular weights has been observed for human and macaque sperm PH-20 and is thought to be related to the glycoprotein structure of PH-20 [29,30]. Further, although the commercial preparation of bovine testis hyaluronidase (lane 5) contained a major band of approximately 33 kDa and a well-defined band at approximately 69 kDa, and a diffuse pattern of bands ranging from 60 to 65 kDa could also be observed. SDS-PAGE of specimens from cell-layer extracts and conditioned media of synoviocytes gave similar results (not shown).
The hyaluronidase activity of the proteins recognized by PH-20 antibodies was examined next. Proteins eluted from the immunoaffinity gel either with the rather harsh glycine–HCl buffer or with the milder Immuno-Pure® Gentle elution buffer were subjected to HA substrate gel electrophoresis at pH 6.5 (Fig. 5). Surprisingly, specimens eluted with the gentle buffer (Fig. 5a, lane 1) and those eluted with the glycine–HCl buffer (Fig. 5a, lane 3) both exhibited a single band of activity with an apparent molecular weight of above 200 kDa. Because specimens were not heat-denatured and because primate sperm PH-20 has a HA binding domain that is distinct from the hyaluronidase domain [17,31], we hypothesized that the binding of high-molecular-weight HA molecules to the PH-20 molecules may hamper the protein's movement to its expected location. Therefore, because HA oligosaccharides are known to dissociate the non-link-stabilized aggrecan molecules from their high-molecular-weight HA backbone [32], we accordingly prepared HA oligosaccharides varying in length from 10 to about 30 monosaccharides, which were then allowed to interact with specimens (about 1 mg of HA oligosaccharides/mg proteins) before being loaded into the gel. When this procedure was used (Fig. 5b), the band of activity switched from the apparent molecular weight of >200 kDa to the expected apparent molecular weight of 60 to 65 kDa in specimens eluted with the gentle buffer (lanes 1 and 2) and also in specimens eluted with the glycine–HCl buffer (lanes 3 and 4). We were unable to detect any hyaluronidase activity (Fig. 5c) in the HA oligosaccharide preparation (lane 1) or the gentle buffer (lane 2). On the other hand, the hyaluronidase activity was stronger in specimens eluted with the gentle buffer (Fig. 5a, lane 1, and 5b, lane 2) than in specimens eluted with the glycine buffer (Fig. 5a, lane 3, and 5b, lane 4), an observation suggesting that the glycine buffer is likely to be harmful to the hyaluronidase activity of the eluted proteins. Further, both the hyaluronidase activity eluted from the immunoaffinity gel and that of the preparation of bovine testicular hyaluronidase disappeared when the HA substrate gels were incubated in the presence of apigenin (Fig. 6), a well-known inhibitor of hyaluronidase [33]. This observation and the presence of saccharolactone in the incubation buffer strongly suggest that HA degradation reflects the hyaluronidase activity alone rather than the combined action of hyaluronidase and exoglycosidases.
Further characterization of the hyaluronidase activity present in cell-layer extracts and conditioned media
The pH profile of the hyaluronidase immunoprecipitated from cell-layer-associated extracts and conditioned media was further examined using the radiolabelled HA substrate assay, a procedure that can detect with reliability an enzymatic activity ranging between 0.005 and 0.05 rTRUs (Fig. 7, upper panel) and that is thus quite a bit more sensitive than HA substrate gel electrophoresis. Although elution with a very acidic pH seems to reduce the hyaluronidase activity of the eluted proteins as assessed by HA substrate gel electrophoresis, preparations of cell layer extracts and conditioned media as well as a commercial preparation of testicular hyaluronidase were eluted from the immunoaffinity gel with the glycine–HCl buffer, because the composition of the gentle elution buffer from Pierce was not known. Eluted proteins were then diluted to be in the linear range of the assay before being incubated for 16 hours at 37°C in appropriate buffers whose pHs were systematically adjusted for a temperature of 37°C. Because the stock of tritium-labelled HA substrate was in 0.4 M NaCl, the assay had to be conducted in 0.10 M NaCl, although this salt concentration has been reported to decrease the apparent pH optima of testicular hyaluronidase preparations [34]. This notwithstanding, between pH 5 and 8 (Fig. 7, lower panel), the profile of activity of specimens from both conditioned media and cell layer extracts was similar to that of the preparation of testicular hyaluronidase, with a maximum between pH 6 and 7, thereby demonstrating that the hyaluronidase activity detected at near neutral pH does not represent the 'tail' of the pH profile of an enzyme essentially active at acid pH.
Effect of IL-1 on the relative amounts of PH-20 in cell layers extracts and conditioned media
Based on the above results, the immunoaffinity gel was used to evaluate to what extent IL-1 could modulate the amount of PH-20 present in the cell layers and conditioned media of chondrocytes, synoviocytes, and fibroblasts. The various cell lines were incubated for 16 hours in the absence and in the presence of IL-1 at a concentration of 5 ng/ml. For each cell line and for each condition, the amounts of PH-20 molecules immunopurified from Triton-X-100 cell layer extracts and corresponding culture media were summed and expressed as the relative percentage of the total amount of proteins present in Triton-X-100-extracted cell layers (Fig. 8, upper panels). Striking differences were observed between the three cell lines. In the absence of IL-1, fibroblasts expressed the lowest amounts of PH-20 and synoviocytes produced the highest amounts. However, stimulation with IL-1 (5 ng/ml) increased the total amount of expressed PH-20 molecules by a factor of 1.9 in fibroblasts and 1.5 in both chondrocytes and synoviocytes.
The relative amounts of PH-20 present in culture media also differed markedly from one cell line to another (Fig. 8, lower panel). For unstimulated cells, the relative amounts of PH-20 liberated into conditioned media were lowest in fibroblasts and highest in synoviocytes. However, when cells were stimulated with IL-1, the relative increase in the amount of PH-20 secreted into the culture medium were higher in fibroblasts than in either chondrocytes or synoviocytes.
Discussion
The data reported herein strongly suggest, for the first time, that human chondrocytes, synoviocytes, and dermal fibroblasts all express, at both the mRNA and protein levels, a neutral-active hyaluronidase similar to PH-20. Further, and importantly, IL-1 enhanced PH-20 levels not only in cell layer extracts but also in culture media of the three cell lines. These observations are likely to be of great interest to both clinicians and scientists, because they shed new light on the biology of connective tissue. Indeed, this neutral-active hyaluronidase is likely to play a key role in the normal turnover of HA and in the receptor-mediated pathways of HA endocytosis. The enzyme also generates HA fragments that are highly angiogenic and are potent inducers of inflammatory cytokines. On the other hand, overexpression or uncontrolled expression of PH-20 can cause great havoc in the extracellular matrix of connective tissues, since the filamentous molecule of HA serves as a backbone upon which other macromolecules associate. Thus, like bacteria, fungi, viruses, leeches, bees, lizards, and snakes, which all use neutral-active hyaluronidase(s) to open up tissue spaces and facilitate penetration [35], spermatozoa use PH-20 to penetrate the cumulus oophorus [17], and invasive and metastatic breast cancers express high levels of PH-20 [21].
Several factors could explain why the enzyme has not been detected before in connective tissue cell lines. Northern blot analysis and the more sensitive RT-PCR technique detected the PH-20 transcript in testis, normal breast tissue, metastatic cancer cell lines, fetal and placental cDNA libraries, and murine kidneys, as well as in trace amounts in the prostate, but those studies never investigated the RNA from articular cartilage, synovium, or skin [17-21]. Two previous studies did not detect the PH-20 transcript in chondrocytes. Because of their 'gene-homology' RT-PCR approach, Flannery and colleagues [22] used reverse primers that do not match correctly the reported PH-20 cDNA sequence, whereas, although they were far from being ideal (according to the Oligo and Primer-3 programs), the primers used by Nicoll and colleagues [36] were apparently able to detect the PH-20 mRNA in testis cDNA but not in chondrocytes and fibroblasts. On the other hand, we provide strong evidence that, in contrast to Triton X-100, a detergent used to solubilize sperm PH-20 [17], octylglucoside, which was used in previous studies to solubilize the hyaluronidase activity present in chondrocytes and fibroblasts [22,28], is unable to extract PH-20 in substantial amounts from connective cell layers. Further, our data also show that detection of the neutral-active hyaluronidase at the expected molecular weight by HA substrate gel electrophoresis can be missed because the migration of the enzyme into the gel may be hampered by the interaction between, on the one hand, HA molecules of relatively high molecular weight and, on the other hand, the HA binding domain of PH-20, which is separate from the hyaluronidase domain of PH-20 [17].
HA is not an inert constituent of the extracellular matrix, but, rather, is highly metabolically active, with a half-life ranging from less than a day in the skin to about 3 weeks in articular cartilage [37,38]. As there is evidence that after a first degradation within the tissues of origin, HA is drained by the lymphatic system before being further degraded in lymph nodes, liver, and kidney [39], PH-20 is likely to contribute to the pool of HA molecules that are easily released from the tissue into the lymphatic system. There is also strong evidence that, via CD44 and/or other cell surface receptors, many cell types can bind and internalize HA, a process that, because of steric inhibition, is dependent upon the size of both HA and HA-bound macromolecules [40,41]. Since PH-20 degrades not only HA but also the chondroitin sulfate chains of the various proteoglycan molecules present in the extracellular matrix, the enzyme is likely to facilitate HA internalization, a process that does not necessarily lead to the degradation of the glycosaminoglycan to small oligosaccharides within lysosomes. Indeed, HA networks have been observed in both the cytoplasm and nucleus of several cell lines [42]. These observations and the growing list of intracellular hyaladherins, such as RHAMM, suggest that the intracellular networks of HA may regulate intracytoplasmic and intranuclear signaling events that are thought to contribute to various inflammatory processes [43].
PH-20 generates a mixture of HA oligosaccharides and fragments that may interact with various cells and produce distinct and important biological effects quite different from those induced by the native, high-molecular-weight polymer. Thus, by stimulating the proliferation and migration of vascular endothelial cells via multiple signaling pathways, HA fragments, but not high-molecular-weight HA molecules, are angiogenic [8,9]. Obviously, PH-20 produces angiogenic HA fragments, since cancer cell lines expressing this hyaluronidase induce angiogenesis in the cornea of mice whereas cancer cell lines lacking PH-20 mRNA do not [18]. Further, since PH-20 activity is enhanced by IL-1, the hyaluronidase may also contribute to the production of HA fragments that accumulate under inflammatory conditions and act as signaling molecules. Indeed, while high-molecular-weight HA molecules suppress the proliferation of synovial cells as well as the production of IL-1, prostaglandin E2, and matrix metalloproteinase-3 by arthritic synovium [10-12], fragmented HA molecules not only induce irreversible phenotypic and functional maturation of dendritic cells [44] but also stimulate the production of cytokines, chemokines, and nitric oxide by macrophages, an activity involving nuclear factor κB and several other transcription factors [7].
The finding that human dermis contains a neutral-active hyaluronidase suggests that depolymerization of HA can occur locally within the dermis. Accumulation of dermal HA with its associated water of hydration, as seen in urticaria and bullous skin lesions, can arise from both increased synthesis and local catabolic failures. On the other hand, PH-20-induced depolymerization of HA may contribute to the scar formation that occurs in adult wounds: fetal wounds have an extracellular matrix rich in HA of high molecular weight and heal without fibrosis, whereas the addition of hyaluronidase to the wound fluid enhances wound fibrosis [45,46]. By binding transforming growth factor β, the most critical factor involved in inflammatory fibrosis, HA helps to concentrate and to protect from proteolytic degradation this growth factor [47], which could be then released by PH-20.
Uncontrolled and/or up-regulated PH-20 activity may be very deleterious to cartilage matrix, both directly and indirectly, since even a couple of cleavages along the filamentous backbone of aggrecan aggregates dramatically reduce the viscoelastic properties of articular cartilage as well as the size of these aggregates, which become no longer effectively immobilized within the collagen network of the articular tissue. Further, the oligosaccharides produced by PH-20 have been shown to induce a dose-dependent chondrocytic chondrolysis as well as up-regulation of aggrecan synthesis and HA synthase 2 mRNA [48]. In both experimental and human osteoarthritis, the progressive reduction in the HA content of articular cartilage is believed to contribute to the apparent irreversibility of the disease process [49-51]. Because the loss of HA from cartilage explants occurs in spite of an up-regulation in HA biosynthesis [52,53], a likely explanation is that HA strands are being degraded at an accelerated rate by a hyaluronidase active at near neutral pH. On the other hand, PH-20 may be also responsible for the release of aggrecan ternary complexes made of aggrecans, link protein, and HA from cartilage matrix upon stimulation with retinoic acid, a process that persists when cartilage explants are bathed with AG3340 at concentrations that completely inhibit the collagenolytic activity present in explants as well as the enzymatic activity of both aggrecanase-1 and aggrecanase-2 [54].
The observation that IL-1 up-regulates the production of PH-20 by both chondrocytes and synoviocytes is worth noting, because enhanced expression of the enzyme may contribute to the degradation of cartilage matrix in arthritides such as rheumatoid arthritis; this contention is strengthened by the observation that proinflammatory cytokines increase the loss of HA fragments from cartilage explants [14]. On the other hand, there is evidence that proinflammatory cytokines up-regulate the levels of chondrocyte lysosomal hyaluronidases [22] as well as the expression of the CD44/HA receptor by chondrocytes [41], thereby suggesting that, within articular cartilage, HA catabolism also involves endocytosis and intracellular degradation. Although this intracellular pathway does not explain the loss of HA molecules from cartilage matrix, the intracellular and the extracellular PH-20 pathways can be complementary, since, as stated above, the degradation of high-molecular-weight HA molecules facilitates their endocytosis [40,41].
Furthermore, because the pH of synovial fluid is usually above 6 [55], PH-20 can contribute to the depolymerization of HA molecules present in this body fluid. The local production of HA fragments is likely to enhance joint degradation by producing HA fragments that, concomitantly, trigger the inflammatory reaction, and reduce dramatically the rheological properties of the joint fluid. It has been suggested that the nonsteroidal anti-inflammatory drugs, such as indomethacin, may exert a portion of their anti-inflammatory properties by inhibiting hyaluronidase and, hence, the generation of small HA fragments [56].
As PH-20 has a HA-binding domain that is distinct from its hyaluronidase domain, the molecule may also act as a HA receptor at the cell surface. Indeed, binding of HA to this distinct domain of sperm PH-20 results in thyrosine phosphorylation and an increase in intracellular calcium [17]). Glycosylphosphatidylinositol-anchored proteins involved in signaling are often associated with nonreceptor protein kinases that are bound to the cytoplasmic leaflet of the plasma membrane and are believed to regulate signal transduction [57].
Conclusion
Our study provides strong evidence that connective tissue cells express PH-20 and that the production of this neutral-active hyaluronidase is up-regulated by IL-1. Since, besides having unique physicochemical properties, HA can modify cell behavior and plays a key role in the organization of the extracellular matrix of connective tissues, this finding may contribute new insights into the pathophysiology of several disorders including skin and joint diseases. Although the overall hyaluronidase activity detected at neutral pH was relatively low, remodeling of the extracellular matrix during wound healing, inflammatory processes, and cancer growth and metastasis are slow processes. In this context, even very slow enzymatic activities at physiological pH, operating on a time scale of hours to days rather than minutes, may suffice to create great havoc.
Abbreviations
bp = base pairs; BSA = bovine serum albumin; DMEM = Dulbecco's modified Eagle's medium; dpm = distintegrations per minute; FCS = fetal calf serum ; HA = hyaluronan; IL-1 = interleukin-1; kb = kilobases; PBS = phosphate-buffered saline; PH-20 = sperm hyaluronidase or sperm adhesion molecule 1; RT-PCR = reverse transcriptase polymerase chain reaction; rTRUs = relative turbidity-reducing units; SSC = 0.15 M Na Cl, 0.015 M sodium citrate, pH 7.0.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
HEL played a leading role in the coordination of the study, contributed to the research protocols, participated in the interpretation of results, and prepared the manuscript. AAC conducted the in situ hybridization and histological studies. DHM designed the research protocols, supervised the studies, and drafted the manuscript. All authors read and approved the final manuscript.
Acknowledgements
The authors thank Professor Emile Van Schaftingen for constructive discussions, Annette Marcelis for technical assistance and Professor Marie-Paule Mingeot for providing the preparation of liver lysozomes.
Figures and Tables
Figure 1 Gene-specific RT-PCR survey of mRNA for the hyaluronidase PH-20 in human connective tissue cells. Three sets of primers amplified the expected product: 759 base pairs (bp) with primers A (lane 1, testis; lane B, chondrocytes; lane 3, negative control), 1,000 bp with primers B (lanes 5, 8, and 10), and 403 bp with primers C (lanes 4 and 7). Primers for actin also amplified an expected product of 226 bp (lanes 6, 9, and 11).
Figure 2 Northern blot analysis of mRNA from various connective tissue cell lines. (a) A radiolabelled PH-20 cDNA probe was hybridized to a northern blot containing 25 μg/lane of total RNA from (1) testis (positive control); (2) liver (negative control); (3) synoviocytes; (4) chondrocytes; (5) chondrocytes stimulated with 5 ng/ml of IL-1; (6) chondrocytes stimulated with 10 ng/ml of IL-1; (7) fibroblasts; (8) a chondrosarcoma cell line. (b) For loading control, the same blot was stripped and hybridized with a β-actin probe. (c) Column box–whisker plot showing the relative increase of the PH-20 2.4-kilobase transcript in chondrocytes stimulated with IL-1 at two concentrations (n = 5; *P = 0.0004 by paired t-test).
Figure 3 Hyaluronan (HA) substrate gel electrophoresis of fibroblast cell layers extracted with octylglucoside. After electrophoresis, gels were incubated at 37°C for 16 hours at pH 4 or pH 6.5 before being treated with proteinase K and stained with alcian blue. Lane 1, fibroblast cell layers extracted with octylglucoside; lane 2, a preparation of liver lysosomes; lane 3, a commercial preparation of bovine testicular hyaluronidase. Standards were Precision Plus Protein standards from Bio-Rad, with major bands at 75 and 50 kDa).
Figure 4 Tricine SDS-PAGE (T-SDS-PAGE) of proteins purified with the anti-PH-20-antibodies–Sepharose gel. (a) Triton X-100-extracted (lane 1) and octylglucoside-extracted (lane 2) fibroblast cell layers were allowed to interact with the immunoaffinity gel, and fractions eluted with the glycine–HCl buffer were analyzed using T-SDS-PAGE. Lane 3, commercial preparation of bovine testis hyaluronidase; lane 4, Precision Plus Protein standards from Bio-Rad. A strong band was detected in the Triton X-100 extracts, whereas a faint band was barely detected in the octylglucoside extracts. (b) Triton X-100-extracted chondrocyte cell layers and their conditioned media were applied to immunoaffinity columns. Samples eluted with the glycine–HCl buffer were subjected to T-SDS PAGE. Unstimulated chondrocyte cell layers (lane 3; 1.4 μg) and their conditioned medium (lane 1; 0.8 μg). Cell layers of chondrocytes stimulated with 5 ng/ml of IL-1 (lane 4; 1.2 μg) and their conditioned medium (lane 2; 1 μg). Lane 5, commercial preparation of bovine testis PH-20; lane 6, Precision Plus Protein standards from Bio-Rad. Several bands ranging from approximately 60 to approximately 65 kDa can be identified in each specimen.
Figure 5 Hyaluronan (HA) substrate gel electrophoresis of chondrocyte cell layers extracted with Triton X-100. (a) Specimens of fibroblast cell layers extracted with Triton X-100 (3 mg as total protein) were applied to an anti-PH-20-antibody Sepharose gel and eluted with either the ImmunoPure Gentle Ag/Ab buffer (lane 1) or the glycine–HCl buffer (lane 3) were subjected to HA substrate gel electrophoresis. A band of activity with an unexpected molecular weight (>200 kDa) was observed in both specimens. (b) When HA oligosaccharides were added to specimens eluted with either the 'gentle' buffer or the glycine–HCl buffer, the band of activity appeared at the expected molecular weight: 'gentle' buffer-eluted specimens without (lane 1) and with (lane 2) HA oligosaccharides; glycine–HCl buffer-eluted specimens without (lane 3) and with (lane 4) HA oligosaccharides; lane 5, commercial preparation of testicular hyaluronidase. Similar results were obtained with specimens from fibroblast and synoviocyte cell layers. (c) No hyaluronidase activity was detected in the HA oligosaccharide preparation (lane 1) or in the 'gentle' buffer preparation (lane 2); lane 3, commercial preparation of testicular hyaluronidase.
Figure 6 Apigenin inhibits the hyaluronidase activity of immunopurified proteins. The hyaluronidase activity present in a commercial preparation of testicular hyaluronidase (Api(-), left lane) as well the hyaluronidase activity purified from cell layers with the immunoaffinity gel (Api(-), right lane), both disappeared when HA substrate gels were incubated in the presence of apigenin (Api(+)).
Figure 7 Detection of hyaluronidase activity by a radiolabelled hyaluronan (HA) substrate assay. (a)Standard curve between the relative percentage of degraded HA molecules and the relative turbidity-reducing units (rTRUs) of hyaluronidase. (b) pH profile of the activity of proteins immunoprecipitated from chondrocyte-layer extracts (open squares), chondrocyte-conditioned media (open circles) and a commercial preparation of bovine testicular hyaluronidase (closed circles). Samples were analyzed in triplicate and activity was expressed as a percentage of maximal activity (= relative activity). A similar pH profile was obtained with specimens from synoviocyte-layer extracts and conditioned media.
Figure 8 Effect of IL-1 on total PH-20 production and hyaluronidase secretion by connective tissue cell lines. Chondrocytes, synoviocytes, and fibroblast cell layers either were not stimulated or were stimulated with IL-1 at a concentration of 5 ng/ml. At the end of the stimulation period, the PH-20 proteins present in conditioned media and Triton X-100-extracted cell layers were purified by the immunoaffinity gel and quantified using the commercial BCA assay. Upper panel: total amount of PH-20 (cell layer + corresponding medium) expressed as the relative percentage of the total amount of proteins present in the Triton-X-100 cell layer extract. Lower panel: relative percentage of the total PH-20 liberated into the conditioned medium. P = P value as assessed by paired t-test.
==== Refs
Laurent TC Fraser JRE Hyaluronan FASEB J 1992 6 2397 2404 1563592
Toole BP Hay E Proteoglycans and hyaluronan in morphogenesis and differentiation Cell Biology of Extracellular Matrix 1991 2 New York: Plenum Press 305 341
Sherman L Sleeman J Herrlich P Ponta H Hyaluronate receptors; key players in growth, differentiation, migration and tumor progression Curr Opin Cell Biol 1994 6 726 733 7530464 10.1016/0955-0674(94)90100-7
Entwistle J Hall CL Turley EA HA receptors: regulators of signalling to the cytoskeleton J Cell Biochem 1996 61 569 577 8806080 10.1002/(SICI)1097-4644(19960616)61:4<569::AID-JCB10>3.0.CO;2-B
Hardingham T Seibel MJ, Robins SP, Bilezikian JP Proteoglycans and glycosaminoglycans Dynamics of Bone and Cartilage Metabolism 1999 San Diego: Academic Press 71 81
Balazs EA Denlinger JL Sodium hyaluronate and joint function J Eq Vet Sci 1985 5 217 228
Noble PW Hyaluronan and its catabolic products in tissue injury and repair Matrix Biol 2002 21 25 29 11827789 10.1016/S0945-053X(01)00184-6
Slevin M Krupinski J Kumar S Gaffney J Angiogenic oligosaccharides of hyaluronan induce protein tyrosine kinase activity in endothelial cells and activate a cytoplasmic signal transduction pathway resulting in proliferation Lab Invest 1998 78 987 1003 9714186
Lokeshwar VB Selzer MG Differences in hyaluronic acid-mediated functions and signaling in arterial, microvessel, and vein-derived human endothelial cells J Biol Chem 2000 275 27641 27649 10882722
Asari A Mizuno S Tanaka I Sunose A Kuriyama S Miyazaki K Namiki O Suppression of hyaluronan and prostaglandin E2 production in traumatic arthritic synovial cells by Na HA Connect Tissue 1997 29 1 5
Homandberg GA Hui F Wen C Kuettner KE Williams JM Hyaluronic acid suppresses fibronectin fragment mediated cartilage chondrolysis in vitro Ostteoarthritis Cartilage 1997 5 309 319
Takahashi K Goomer RS Harwood F Kubo T Hirasawa Y Amiel D The effects of hyaluronan on matrix metalloproteinases-3 (MMP-3), interleukin-1-beta (IL-1 beta) and tissue inhibitor of metalloproteinase-1 (TIMP-1) gene expression during the development of osteoarthritis Osteoarthritis Cartilage 1999 7 182 190 10222217 10.1053/joca.1998.0207
Dahl LB Dahl IM Engstrom-Laurent A Granath K Concentration and molecular weight of sodium hyaluronate in synovial fluid from patients with rheumatoid arthritis and other arthropathies Ann Rheum Dis 1985 44 817 822 4083937
Fosang AJ Tyler JA Hardingham TE Effect of interleukin-1 and insulin like growth factor-1 on the release of proteoglycan components and hyaluronan from pig articular cartilage in explant culture Matrix 1991 11 17 24 2027327
Deguine V Menasche M Ferrari P Fraisse L Pouliquen Y Robert L Free radical depolymerization of hyaluronan by Maillard reaction products: role in liquefaction of aging vitreous Int J Biol Macromol 1998 22 17 22 9513812 10.1016/S0141-8130(97)00084-6
Gmachl M Sagan S Ketter S Kreil G The human sperm protein PH-20 has hyaluronidase activity FEBS Lett 1993 336 545 548 8282124 10.1016/0014-5793(93)80873-S
Cherr GN Yudin AIY Overstreet JW The dual functions of GPI-anchored PH-20: hyaluronidase and intracellular signaling Matrix Biology 2001 20 515 525 11731269 10.1016/S0945-053X(01)00171-8
Liu D Pearlman E Diaconu E Guo K Mori H Haqqi T Markowitz S Willson J Sy M-S Expression of hyaluronidase by tumor cells induces angiogenesis Proc Natl Acad Sci USA 1996 93 7832 7837 8755562 10.1073/pnas.93.15.7832
Sun L Feusi E Sibalic A Beck-Schimmer B Wuthrich RP Expression profile of hyaluronidase mRNA transcripts in the kidney and in renal cells Kidney Blood Press Res 1998 21 413 418 9933825 10.1159/000025893
Csòka AB Scherer SW Stern R Expression analysis of six paralogous human hyaluronidase genes clustered on chromosomes 3p21 and 7q31 Genomics 1999 60 356 361 10493834 10.1006/geno.1999.5876
Beech DJ Madam AK Deng N Expression of PH-20 in normal and neoplastic breast tissue J Surg Res 2002 103 203 207 11922735 10.1006/jsre.2002.6351
Flannery CR Little CB Hughes CE Caterson B Expression and activity of articular cartilage hyaluronidases Biochem Biophys Res Commun 1998 251 824 829 9790994 10.1006/bbrc.1998.9561
Lepperdinger G Strobl B Kreil G HYAL2, a human gene expressed in many cells, encodes a lysosomal hyaluronidase with a novel type of specificity J Biol Chem 1998 273 22466 22470 9712871 10.1074/jbc.273.35.22466
Frost GI Csóka TB Wong T Stern R Purification, cloning, and expression of human plasma hyaluronidase Biochem Biophys Res Commun 1997 236 10 15 9223416 10.1006/bbrc.1997.6773
Schagger H von Jagow G Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis for the separation of proteins in the range from 1 to 100 kDa Anal Biochem 1987 166 368 379 2449095 10.1016/0003-2697(87)90587-2
Miura RO Yamagata S Miura Y Harada T Yamagata T Analysis of glycosaminoglycan-degrading enzymes by substrate gel electrophoresis (Zymography) Anal Biochem 1995 225 333 340 7762800 10.1006/abio.1995.1163
Jones MH Davey PM Aplin H Affara NA Expression analysis, genomic structure, and mapping to 7q31 of the human sperm adhesion molecule gene SPAM1 Genomics 1995 29 796 800 8575780 10.1006/geno.1995.9931
Stair-Nawy S Csoka AB Stern R Hyaluronidase expression in human skin fibroblasts Biochem Biophys Res Commun 1999 266 268 273 10581201 10.1006/bbrc.1999.1802
Sabeur K Cherr GN Yudin AI Primakoff P Li MW Overstreet JW The PH-20 protein in human spermatozoa J Androl 1997 18 151 158 9154509
Li MW Yudin AI Robertson KR Cherr GN Overstreet JW Importance of glycosylation and disulfide bonds in hyaluronidase activity of macaque sperm surface PH-20 J Androl 2002 23 211 219 11868814
Yudin AI Li MW Robertson KR Cherr GN Overstreet JW Characterization of the active site of monkey sperm hyaluronidase Reproduction 2001 121 735 743 11427161 10.1530/rep.0.1210735
Hascall VC Heinegard D Aggregation of cartilage proteoglycans. II. Oligosaccharide competitors of the proteoglycan-hyaluronic acid interaction J Biol Chem 1974 249 4242 4249 4277352
Kuppusamy UR Das NP Inhibitory effects of flavonoids on several venom hyaluronidases Experientia 1991 47 1196 1200 1765130
Gasesa P Savitsky MJ Dodgson KS Olavesen AH Effect of ionic strength on the activity of bovine testicular hyaluronidase Biochem Soc Trans 1979 7 1287 1289 43812
Csòka AB Frost GL Stern R Hyaluronidases and tissue invasion Invasion Metastasis 1997 17 297 311 9949289
Nicoll SB Barak O Csoka AB Bhatnagar RS Stern R Hyaluronidases and CD44 undergo differential modulation during chondrogenesis Biochem Biophys Res Commun 2002 292 819 825 11944887 10.1006/bbrc.2002.6697
Tammi R Saamanen AM Maibach HI Tammi M Degradation of newly synthesized high molecular mass hyaluronan in the epidermal and dermal compartments of human skin in organ culture J Invest Dermatol 1991 97 126 130 2056182 10.1111/1523-1747.ep12478553
Morales TI Hascall VC Correlated metabolism of proteoglycans and hyaluronic acid in bovine cartilage organ cultures J Biol Chem 1988 263 3632 3638 3346212
Fraser JRE Brown TJ Laurent TC Laurent TC Catabolism of hyaluronan The Chemistry, Biology and Medical Applications of Hyaluronan and its Derivatives 1998 London: Portland Press 85 92
Lee JY Spicer AP Hyaluronan: a multifunctional, megaDalton, stealth molecule Curr Opin Cell Biol 2000 12 581 586 10978893 10.1016/S0955-0674(00)00135-6
Knudson W Chow G Knudson CB CD-44 mediated uptake and degradation of hyaluronan Matrix Biology 2002 21 15 23 11827788 10.1016/S0945-053X(01)00186-X
Evanko SP Wight TN Intracellular localization of hyaluronan in proliferating cells J Histochem Cytochem 1999 47 1331 1342 10490462
Hascall VC Majors AK De La Motte CA Evanko SP Wang A Drazba JA Strong SA Wight TN Intracellular hyaluronan: a new frontier for inflammation? Biochim Biophys Acta 2004 1673 3 12 15238245
Termeer CC Hennies J Voith U Ahrens T Weiss JM Prehm P Simon JC Oligosaccharides of hyaluronan are potent activators of dendritic cells J Immunol 2000 165 1863 1870 10925265
West DC Shaw DM Lorenz P Adzick NS Longaker MT Fibrotic healing of adult and late gestation fetal wounds correlates with increased hyaluronidase activity and removal of Hyaluronan Int J Biochem Cell Biol 1997 29 201 210 9076955 10.1016/S1357-2725(96)00133-1
Mackool RJ Gittes GK Longaker MT Scarless healing. The fetal wound Clin Plast Surg 1998 25 357 365 9696898
Locci P Marinucci L Lilli C Martinese D Becchetti E Transforming growth factor beta 1-hyaluronic acid interaction Cell Tissue Res 1995 281 317 324 7648625
Knudson W Casey B Nishida Y Eger W Kuettner KE Knudson CB Hyaluronan oligosaccharides perturb cartilage matrix homeostasis and induce chondrocytic chondrolysis Arthritis Rheum 2000 43 1165 1174 10817571 10.1002/1529-0131(200005)43:5<1165::AID-ANR27>3.0.CO;2-H
Sweet MB Thonar EJ Immelman AR Solomon L Biochemical changes in progressive osteoarthrosis Ann Rheum Dis 1977 36 387 398 144463
Manicourt DH Pita JC Progressive depletion of hyaluronic acid in early experimental osteoarthritis in dogs Arthritis Rheum 1988 31 538 544 3358811
Muller FJ Setton LA Manicourt DH Mow VC Howell DS Pita JC Centrifugal and biochemical comparison of proteoglycan aggregates from articular cartilage in experimental joint disuse and joint instability J Orthop Res 1994 12 498 508 8064480 10.1002/jor.1100120406
Ryu J Treadwell BV Mankin HJ Biochemical and metabolic abnormalities in normal and osteoarthritic human articular cartilage Arthritis Rheum 1984 27 49 57 6691859
Manicourt DH Druetz-Van Egeren A Haazen L Nagant de Deuxchaisnes C Effects of tenoxicam and aspirin on the metabolism of proteoglycans and hyaluronan in normal and osteoarthritic human articular cartilage Br J Pharmacol 1994 113 1113 1120 7889262
Sugimoto K Iizawa T Harada H Yamada K Katsumata M Takahashi M Cartilage degradation independent of MMP/aggrecanases Osteoarthritis Cartilage 2004 12 1006 1014 15564068 10.1016/j.joca.2004.09.003
Mapp PJ Stevens CR Blake DR Maddison PJ, Isenberg DA, Woo P, Glass DN The physiology of the joint and its disturbance in inflammation Oxford Textbook of Rheumatology 1993 1 Oxford: Oxford University Press 256 268
Mio K Stern R Inhibitors of hyaluronidases Matrix Biology 2002 21 31 37 11827790 10.1016/S0945-053X(01)00185-8
Harder T Simons K Clusters of glycolipid and glycosylphosphatidylinositol-anchored proteins in lymphoid cells: accumulation of actin regulated by local tyrosine phosphorylation Eur J Immunol 1999 29 556 562 10064071 10.1002/(SICI)1521-4141(199902)29:02<556::AID-IMMU556>3.0.CO;2-2
| 15987477 | PMC1175024 | CC BY | 2021-01-04 16:02:38 | no | Arthritis Res Ther. 2005 Apr 4; 7(4):R756-R768 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1730 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17311598747310.1186/ar1731Research ArticleRegeneration of the immunoglobulin heavy-chain repertoire after transient B-cell depletion with an anti-CD20 antibody Rouzière Anne-Sophie [email protected] Christian [email protected] Arumugam [email protected]örner Thomas [email protected] Hans-Peter [email protected] Department of Medicine II, Rheumatology and Clinical Immunology, University of Wuerzburg, Germany2 Charité University Hospital, Berlin, Germany2005 1 4 2005 7 4 R714 R724 5 8 2004 10 9 2004 1 3 2005 7 3 2005 Copyright © 2005 Rouzière et al, licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is cited.
B-cell depletive therapies have beneficial effects in patients suffering from rheumatoid arthritis. Nevertheless, the role of B cells in the pathogenesis of the disease is not clear. In particular, it is not known how the regeneration of the B-cell repertoire takes place. Two patients with active rheumatoid arthritis were treated with rituximab, and the rearranged immunoglobulin heavy-chain genes (Ig-VH) were analysed to follow the B-cell regeneration. Patient A was treated with two courses of rituximab, and B-cell regeneration was followed over 27 months by analysing more than 680 Ig-VH sequences. Peripheral B-cell depletion lasted 7 months and 10 months, respectively, and each time was accompanied by a clinical improvement. Patient B received one treatment course. B-cell depletion lasted 5 months and was accompanied by a good clinical response. B cells regenerated well in both patients, and the repopulated B-cell repertoire was characterised by a polyclonal and diverse use of Ig-VH genes, as expected in adult individuals. During the early phase of B-cell regeneration we observed the expansion and recirculation of a highly mutated B-cell population. These cells expressed very different Ig-VH genes. They were class-switched and could be detected for a short period only. Patient A was followed long term, whereby some characteristic changes in the VH2 family as well as in specific mini-genes like VH3–23, VH 4–34 or VH 1–69 were observed. In addition, rituximab therapy resulted in the loss of clonal B cells for the whole period.
Our data show that therapeutic transient B-cell depletion by anti-CD20 antibodies results in the regeneration of a diverse and polyclonal heavy-chain repertoire. During the early phase of B-cell regeneration, highly mutated B cells recirculate for a short time period in both the patients analysed. The longitudinal observation of a single patient up to 27 months shows subtle intraindividual changes, which may indicate repertoire modulation.
==== Body
Introduction
Although the role of B cells in autoimmunity is not completely understood, their importance in the pathogenesis of autoimmune diseases has been further appreciated in the past few years. It is now well known that B cells are more than just the precursors of (auto)antibody-secreting cells [1-4]. They are also involved in the regulation of T-cell-mediated autoimmune diseases by different mechanisms. In this regard, B cells are very efficient antigen-presenting cells. Activated B cells express co-stimulatory molecules, such as CD154, and in this way contribute to the evolution of T effector cells. They can produce chemokines and cytokines, like lymphotoxin α/β, that are essential for the differentiation of follicular dendritic cells in secondary lymphoid organs and for the organisation of effective lymphoid architecture.
There are also indications that B-cell activity is enhanced in rheumatoid arthritis (RA) [2,5]. B cells are found in the synovium, where they form aggregates with T cells and develop tertiary lymphoid tissue structures [5]. The mutational activity of these B cells is markedly enhanced and abnormalities in positive selection and negative selection are found [2]. Furthermore, elimination of B cells by anti-CD20 antibodies from the synovial tissue provokes a disruption of T-cell activation and provokes the production of proinflammatory monokines [6], proving an important role of B cells in the pathogenesis of RA.
The B-cell repertoire is shaped by a complex set of gene rearrangements, somatic hypermutation and receptor-driven selection. These processes are highly regulated during development, ontogeny and the response to antigen [7]. B cells develop in the bone marrow and in the foetal liver, and they mature in the peripheral lymphoid organs. The immunoglobulins they produce contain two heavy polypeptide chains and two light polypeptide chains. The different gene segments are assembled together during recombination to produce a unique rearrangement. The diversity of the repertoire is increased by the addition of or the deletion of nucleotides at the junction between the different gene segments and by random pairing of the heavy chains and light chains. Additional diversity is created by somatic hypermutation, which introduces point mutations to change amino acid codons. This final event takes place in germinal centres, when B cells encounter antigen. The composition of the antibody repertoire is regulated and constrained, and there is substantial evidence that the B-cell repertoire is changed in autoimmune diseases, such as systemic lupus erythematosus [1], Sjögren's syndrome [8,9], myasthenia gravis [10], diabetes mellitus [11,12] or RA [13,14].
B-cell depletive therapies have beneficial effects in patients suffering from RA [14-22]. Rituximab is a chimeric anti-CD20 monoclonal antibody that consists of human IgG1 and kappa constant regions and of mouse variable regions from a hybridoma directed at human CD20. Rituximab has mainly been used for the treatment of non-Hodgkin lymphomas [23]. It selectively depletes CD20+ B cells from the peripheral blood, the spleen and the bone marrow for several months [18]. Because early B-cell precursors do not express CD20, the bone marrow is able to repopulate B lymphocytes after therapy.
The aim of the present study was to determine whether a polyclonal and diverse B-cell repertoire is regenerated after temporary B-cell depletion by anti-CD20 monoclonal antibodies. To address this issue, we have analysed the immunoglobulin heavy-chain gene (Ig-VH) repertoire of two patients suffering from active RA before and after treatment with rituximab, up to a time period of 27 months.
Materials and methods
Patients
A 46-year-old male patient (patient A) was diagnosed with RA using the American College of Rheumatology criteria. The patient was unsuccessfully treated over a time period of 8 years with three disease-modifying anti-rheumatic drug regimens including methotrexate, and also failed therapy with tumour necrosis factor alpha blockers. After giving informed consent, the patient was treated in an open-label protocol, which was approved by the local ethics committee.
Rituximab (Mabthera®; Hoffmann La-Roche, Grenzach-Whylen, Germany) was administered intravenously at a dose of 375 mg/m2 once a week for a total of four infusions (days 1, 8, 15 and 22). The disease eventually relapsed 15 months after the beginning of the study. The patient therefore received another rituximab treatment (four once-weekly doses of 375 mg/m2) at the time point of 17 months. Except for 5 mg prednisolone equivalent daily, the patient received no other antiproliferative treatment during the whole study.
Patient B was a 63-year-old female patient with rheumatoid factor (RF)-positive RA. She had been unsuccessfully treated over a time period of 7 years with three different disease-modifying anti-rheumatic drug regimens including methotrexate. The patient has not yet been treated with a tumour necrosis factor alpha blocking therapy. After informed consent, she was treated with rituximab according to the same protocol as patient A. In addition, patient B continued to receive 20 mg methotrexate weekly. She did not receive any glucocorticoids during the study.
The analysis of lymphocyte subsets by immunofluorescence staining was performed by incubating peripheral blood mononuclear cells (PBMCs) with anti-CD19 and anti-CD3 antibodies (Phycoerythrin (PE) or Fluorescein isothiocyanate (FITC)-labelled as indicated; all antibodies from Becton-Dickinson, Heidelberg, Germany) using a FACSCalibur (Becton-Dickinson, San Jose, CA, USA). Frequencies of cell populations were calculated using CellQuest software.
Disease activity was regularly determined using the disease activity score (DAS28 index) and by monitoring C-reactive protein (CRP) levels.
Amplification of rearranged immunoglobulin V genes by PCR
Genomic DNA was isolated from PBMCs using the QIAamp® DNA Blood Mini Kit (Qiagen, Hilden, Germany). Rearranged VHDJH gene rearrangements were amplified for all VH families using a nested PCR approach [24]. Genomic DNA was amplified in separate reactions for the six VH families (VH1 oligonucleotide primers also co-amplify perfectly VH7 gene rearrangements). The final concentrations of the reagents were 200 μM each dNTP (Peqlab, Erlangen, Germany), 0.625 μM each primer, 2.5 mM MgCl2, 10 × PCR buffer II and 2.5 U AmpliTaq DNA polymerase (Applied Biosystems, Foster City, CA, USA).
During the external amplification round, 250 ng (5 μl) DNA were amplified in a 75 μl reaction containing primers specific for the leader peptide sequence and a mixture of external JH primers in a Gene Amp PCR System 2400 (Perkin Elmer, Applied Biosystems, Foster City, CA, USA). The internal amplification round was conducted with the 5' primer specific for framework region FR1 and a mixture of internal JH primers using 5 μl of the product of the first amplification reaction as template. The cycling parameters have been described previously [24]. Briefly, the annealing temperatures were 50°C for the external amplification round and 65°C for the internal round.
The polymerase error rate for the amplification of VH genes using nested PCR has been documented to be 1 × 10-4 mutations/bp [25].
RT-PCR
Total cellular RNA was extracted from 1 × 107 PBMCs after lysis of the cells in 1.5 ml TRIZOL reagent (Gibco, Karlsruhe, Germany) following the manufacturer's instructions.
The RT-PCR reaction was performed using Titan One Tube RT-PCR system (Roche, Mannheim, Germany). First-strand cDNA was synthesised at 42°C for 60 min in a 50 μl reaction mix containing 5 mM dithiothreitol, 400 ng oligo-dT15, 200 μM dNTP, 8 U RNAse inhibitor, 5 × RT-PCR buffer, 20 U high-fidelity enzyme mix RT-AMV and 1 μg RNA. VH mRNA transcripts were amplified by the nested PCR protocol described earlier using 5 μl cDNA and C-region primers (Cμ, TCA GGA CTG ATG GGA AGC CC; Cγ, CGA GCC GCT GGT CAG AGC G; Cα, ACC CTC AGC GGG AAG ACC TT) as the 3' primers in the first round.
Cloning of rearranged immunoglobulin V genes
All PCR products were separated by electrophoresis through 1.5% agarose and were visualised with ethidium bromide. Successful amplifications were identified as a band corresponding to a product of approximately 350 bp. The bands were excised and subsequently purified using the MinElute Gel extraction kit (Qiagen). The purified PCR products were polished using the PCR polishing kit (Stratagene, Amsterdam Zuidoost, The Netherlands), were ligated with Zero Blunt pCR-blunt vector and were transformed into One Shot® TOP10 cells (Invitrogen, Karlsruhe, Germany)
Sequencing and analysis of rearranged V genes
Plasmid DNA from clones containing gene inserts was prepared using the Wizard Plus SV Minipreps DNA Purification System kit (Promega, Mannheim, Germany). The DNA sequences were determined using BigDye Terminator Cycle Sequencing Ready Reaction kit (Perkin Elmer, Applied Biosystems) and the M13 forward and reverse universal primers in an automated genetic analyser ABI PRISM 310 (Applied Biosystems). Germline immunoglobulin V genes were identified by blast searching the VBase Sequence Directory [26].
Single cell sorting and PCR amplification
The procedure for single cell sorting and subsequent PCR amplifications has been described previously [24]. Briefly, single CD19+CD27- and CD19+CD27+ cells were sorted into wells of 96-well plates using a FACStar Plus flow cytometer with an automated single cell deposit unit (Beckton-Dickinson, USA). After primer extension pre-amplification, the rearranged VHDJH genes were amplified by nested PCRs using the same oligonucleotides as those already described. After gel purification (Qiagen), the PCR products were directly sequenced using the BigDye Terminator Cycle Sequencing Ready Reaction kit (Perkin Elmer, Applied Biosystems) and the 5' V primer used for the internal amplification.
Statistical analysis
The statistical analyses were performed using GraphPad software . Sequences were analysed by the Fisher's exact test to compare the differences in the distribution of particular gene segments. The average lengths of CDR3 were studied by the unpaired t test. The chi-square test was used to compare the mutational frequencies.
Results
Clinical data
Figure 1 shows the clinical response of RA patient A during the two treatment periods with rituximab. B cells accounted for 10.1% of peripheral lymphocytes at the beginning of the study. A rapid B-cell depletion occurred after each therapy, and lasted up to 7 months and 10 months, respectively. The time points of 7 months and 27 months were the first times where B cells were detectable in peripheral blood, either by flow cytometry or by PCR (data not shown). The B-cell frequency was 3% of total lymphocytes at 7 months and was 2.4% of total lymphocytes at 27 months (Fig. 1). Patient A had an active disease before therapy, as indicated by a high CRP level and a high DAS28 clinical activity index. The disease activity declined continuously after rituximab therapy. The patient had a good clinical response starting from 3 months and lasting over 1 year. We observed a deterioration of the clinical parameters about 13 months after the first therapy. The patient therefore received a second treatment with rituximab at 17 months. Again, patient A presented a clinical improvement following B-cell depletion. The arrows in Fig. 1 indicate the four time points when the B-cell repertoire was studied. Table 1 presents the RF values. The RF activity declined quite rapidly after rituximab therapy and followed the inflammatory activity, with increasing values in relapse and falling values after the second treatment.
Patient B started with B cells at 11% of the peripheral blood lymphocytes. B cells were not detectable during the 5 months following the rituximab treatment. B cells represented 2.4% of peripheral lymphocytes at the time point of 5 months, and represented 3.1% at 6 months. The patient experienced a good clinical response. The DAS28 declined from 6.0 before therapy to 3.3 at 5 months, and the CRP values declined from 1.9 mg/dl to 0.65 mg/dl, respectively.
Distribution of VH genes in patient A (Fig. 2)
Immunoglobulin VH gene rearrangements from peripheral blood B cells of patient A were analysed using nested PCR, followed by subcloning and sequencing at the time points indicated in Fig. 1. A total of 687 clones were analysed: 179 clones before treatment, 199 clones during the first phase of B-cell regeneration (7 months after the first therapy), 149 clones after 17 months, and 160 clones after 27 months (at the time of the second B-cell regeneration). Only the productive rearrangements were taken into consideration.
VH1 family
VH1–69 and VH1–18 were the most frequently used VH1 family members before therapy, comprising 68% of the sequences analysed in this family. Significant changes in the VH1 distribution could be observed 7 months after therapy. VH1–02 was increased and became the predominant gene (36% versus 4%, P = 0.0038), whereas VH1–69 was decreased (12% versus 36%, P = 0.0594). The VH1 gene distribution 17 months after therapy was largely comparable with that before treatment (i.e. VH1–18 and VH1–69 were the most often used genes). Notably, VH1–03 was increased and was then the third most frequent gene (18%). The distribution of the VH1 genes after 27 months was quite stable with no significant differences to the previous time point, except that VH1–03 was no longer found.
VH2 family
The gene VH2–05 was slightly predominant in the VH2 family before treatment. The distribution was completely shifted toward the usage of VH2–05 7 months after therapy (96% versus 59%, P = 0.0041), which 10 months later shifted back to the rearrangement frequency found before treatment. At the time point of 27 months, during the second regeneration phase, the frequencies of the VH2 genes were similar to those observed during the regeneration phase following the first treatment (7 months after therapy).
VH3 family
The VH3 family is the largest family, comprising 22 members. Ten family members were found before treatment, two of them accounting for 45% of all VH3 rearrangements (VH3–23, 29%; VH3–30/3–30.5, 16%). The newly regenerated B cells used a greater variety of genes 7 months after therapy (16 different VH3 gene segments were observed). The overall distribution was similar to the first time point, except for VH3–07, which was then over-represented (15% versus 2%, P = 0.059). The gene VH3–23 was significantly increased 17 months after therapy when compared with the previous time point (37% versus 17%, P = 0.0481), and was the most often represented gene, followed by VH3–09 (16% of all VH3 rearrangements). No significant alterations in VH3 gene distribution could be observed in the final time point, despite the tendency for VH3–23 to decrease to its level found at the time point of 7 months.
VH4 family
One single gene in the VH4 family (VH4–34) was overexpressed before therapy, accounting for more than 40% of all VH4 rearrangements. The therapy induced some significant changes in the distribution of the VH4 genes. The frequency of VH4–34 decreased (16% versus 41%, P = 0.0198), whereas VH4–04 increased (21% versus 3%, P = 0.0302). VH4–59 was also increased and became the most frequently rearranged gene (39% of the VH4 sequences after therapy). As described for the VH3 family, there was also a greater variety of genes used in the VH4 rearrangements after therapy. Seven different gene segments were used before treatment, whereas nine different gene segments were found after treatment. The overall rearrangement frequency of the different genes at the time point of 17 months was comparable with that observed before therapy, except for certain genes like VH4–34 that remained at the level seen directly after therapy. Significant changes could be described between the points of 7 months and 17 months after therapy. An increased frequency of the genes VH4–39 and VH4–30.1/4–31 was observed (18% versus 3%, P = 0.0443 and 24% versus 5%, P = 0.0372, respectively), as well as a decreased use of the genes VH4–59 and VH4–04 (12% versus 39%, P = 0.0146 and 6% versus 21%, P = 0.0931, respectively). This distribution was then quite stable up to 27 months.
VH5 family
Within the two-member VH5 family, VH5–51 was the predominant gene at all four time points analysed. Its frequency significantly increased after the first therapy (86% versus 65%, P = 0.036), however, with a further tendency to fall to the pretreatment level. In this family, two B-cell clones were found in the repertoire before therapy (Fig. 3). The first clone comprised six clonally related sequences, with the number of shared mutations varying from zero to six per sequence. This first clone's 33-nucleotide CDR3 involved VH5–51 rearranged to D6-6 and JH5. The second clone consisted of seven clonally related sequences (from which three were nonproductive since one mutation in position 90 generated a stop codon). The CDR3 was 30 nucleotides long. It was composed by VH5-a rearranged to D4–14 and JH4, and had between zero and four mutations. Both B-cell clones disappeared after the first anti-CD20 therapy; their rearrangements were no longer observed and no other B-cell clones were detected during follow-up.
Use of D segments, distribution of JH gene segments and CDR3 length in patient A
All D gene families could be detected in the sequences analysed. Before treatment, the four-member D1 family was under-represented compared with its representation in the genome (3% versus 16% expected). On the contrary, the D6 family that comprises only three members was over-represented (31% versus 15%). A significant increase of the D1 family members was observed after therapy, bringing the frequency to its expected level. Its usage decreased continuously until 27 months. In parallel, D3 became the most frequently used family starting from 17 months after therapy.
The analysis of the JH segments indicated that the overall distribution of these components did not vary with the treatment. JH4 was represented most frequently (accounting for 50% or more of all rearrangements), followed by JH6 (between 20% and 30% of all rearrangements). The other genes were less frequently used. A significant reduction of the JH6 use (20% versus 30%, P = 0.0223), as well as a significant increase of JH2 (7% versus 1%, P = 0.0069), was observed after the first B-cell depletion. The rearrangement frequency of these two gene segments returned to the pretreatment level at 17 months. No significant changes were observed in the other families.
The CDR3 length was calculated by determining the number of nucleotides from residues 95–102. The average length of CDR3 before therapy was 38.0 nucleotides (± 11.0), ranging from 9 to 75 nucleotides. After the first anti-CD20 treatment (7 months after therapy) the average length was 36.9 nucleotides (± 9.4), ranging from 15 to 69 nucleotides; 17 months after therapy the average length was 42.6 nucleotides (± 12.1), ranging from 15 to 78 nucleotides; and at the time point of 27 months the average length was 42.3 nucleotides (± 10.7), ranging from 18 to 75 nucleotides.
Mutational frequencies in VH rearrangements in both patients (Fig. 4)
At the beginning of the study, the overall mutational frequency in the VH genes of patient A was 1.4% (681 mutations/48,891 bp) (Fig. 4a). The mutational frequencies varied from nine mutations/2,448 bp (0.4%) for the VH6 family to 278 mutations/12,390 bp (2.2%) for the VH3 family. The majority of the clones (113 out of 178) contained two mutations or less per rearrangement (data not shown). During the first B-cell regeneration phase, the mutational frequencies varied from 383 mutations/6,940 bp (5.5%) for the VH2 family to 638 mutations/6,542 bp (9.8%) for the VH1 family. The overall mutational frequency was highly and significantly increased at this time point (4,032 mutations/54,720 bp [7.4%] versus 681 mutations/48,891 bp [1.4%] before therapy, P < 0.0001). Almost 90% of the sequences analysed (175 out of 198) had more than 10 mutations per sequence. The frequency of mutations at the time point of 17 months was decreased again to the level found before treatment (519 mutations/39,307 bp [1.4%]) and 92 sequences out 145 contained two mutations or less per rearrangement. The overall mutational frequency stayed in the same low range at the later time points up to 27 months (725 mutations/44,037 bp [1.6%]). Only the distribution of the mutations per rearrangement was somewhat distinct, since 24 out of 159 sequences (15.1%) contained more than 10 mutations (versus 10.1% and 9.7%, before and 17 months after therapy, respectively).
The second patient (patient B) was analysed for mutational frequencies in the Ig-VH repertoire. The results for the VH4 family are presented in Fig. 4b. The mutational frequency before therapy was 1.5% (110 mutations/7,148 bp). Only two sequences out of 26 presented more than 10 mutations per sequence. At the early regeneration point (5 months after therapy), the frequency of mutations was significantly increased to 5.4% (394 mutations/7,292 bp, P < 0.0001) and the majority of the sequences (15 out of 26) contained more than 10 mutations. Four weeks later, the mutational frequency was decreased to 2.8% (169 mutations/6,151 bp). The number of sequences containing more than 10 mutations (five out of 23) was lower than in the previous time point but was still elevated compared with that before therapy.
Overall mutational frequencies of VH rearrangements from single CD19+ cells in patient A (Table 2)
In order to substantiate the unexpected high mutation rate observed 7 months after the first B-cell depletion, PBMCs of patient A were sorted into single CD19+ cells that were either CD27+ or CD27-. This different approach also revealed very high mutational frequencies in both B-cell populations. The overall mutational frequencies of CD19+CD27- single cells before therapy were as low as expected (0.6%). At the time point 7 months, during the first regeneration phase, the mutational frequency was significantly elevated in both CD27- B cells (5.3%) and CD27+ B cells (8.3%).
Discussion
The implication of B cells in the pathogenesis of RA is now well established, but their precise role is still unknown. Numerous studies have shown that B-cell depletion by anti-CD20 therapy can be beneficial for patients suffering from RA [15,16,19-22]. However, it is not known how the B-cell repertoire regenerates after anti-CD20-mediated transient B-cell depletion. In particular, whether a polyclonal and diverse repertoire is reconstituted has not been studied. To address this question, we decided to compare the B-cell repertoire of a RA patient before and after effective clinical B-cell depletive therapy by analysing his Ig-VH repertoire over a time period of 27 months.
A patient with active RA was selected for B-cell depletion using rituximab. He showed a good clinical response for over 1 year after antibody treatment. The disease eventually relapsed and the patient was retreated with rituximab. The second B-cell depletive therapy again induced a significant clinical response lasting about 10 months.
At the beginning of the study, the Ig-VH repertoire of the RA patient basically resembled the published distributions for healthy people [24,27]. Nevertheless, certain genes already described with bias in autoimmune diseases were used in a different proportion. In particular, the genes VH1–69 and VH4–34 were over-represented. The gene VH1–69, which represented 35.7% of the rearrangements for the VH1 family in the present study, has been found at the frequency of 11.1% in healthy persons [24]. The gene VH4–34 represented 41.2% of the VH4 genes in our RA patient. Its frequency in healthy people has been described as 14.3% [24] and 15.7% [27]. On the other hand, the gene VH3–07 was found in a smaller proportion (2.2% of the VH3 genes versus 7.5% [24] and 10.8% [27] in healthy controls). These genes have been shown to exhibit some evidence for (auto)antigen selection; for example, for RF activity [8]. In particular, the gene VH4–34 is very often used by anti-DNA antibodies [28,29] and is exclusively used by cold agglutinins [30]. In agreement with Huang and colleagues' data [13], the gene VH3–30 was found less frequently in our RA patient than in the controls. In addition to these genes, the proportion of some other variable genes like VH1–02, 1–18 or 4–04 also differed from the published data of normal controls and provided a distinct pattern for this patient.
The analysis of D segments and JH genes showed distributions comparable with those described in healthy individuals [24,27,31]. D6 represented the predominant D family and JH4 was the most frequently used JH segment, followed by JH6. The only difference we detected before therapy was the under-representation of the D1 family. The CDR3 length average (38.0 ± 11.0 nucleotides) was in agreement with published data [32].
These observations suggest that the overall representation of individual VH genes in peripheral B cells in the present RA patient resembled the repertoire expected in an adult, but also contained characteristic differences in certain genes already seen to often be biased in autoimmune diseases.
B cells regenerated well after B-cell depletion, and showed a diverse and polyclonal repertoire. Nevertheless, changes in the Ig-VH genes were detected, with the most profound effects observed 7 months after the beginning of therapy, during the early regeneration phase. The intraindividual long-term changes were more subtle.
Seven months after the first therapy was the earliest time point when peripheral B cells could be detected either by flow cytometry or by PCR. These early regenerated B cells presented a distribution of Ig-VH genes significantly different from that before therapy. Some VH genes (such as VH1–02 or VH3–07) were more often used, whereas other genes (e.g. VH4–34) were decreased. Also, significant changes were observed in the distribution of the D segments and JH genes. The most striking differences were the mutational frequencies found in the VH genes (Fig. 4a). At this time point, 3% of the peripheral lymphocytes were CD19+ B cells. All the amplified sequences were extensively mutated (mean mutation rate, 7.4%): 88% contained more than 10 mutations per sequence. This was highly significant when compared with the data observed before therapy and at the time point of 17 months, where only 10% of the rearrangements comprised more than 10 mutations per sequence. This increase of mutations correlated well with the diminution of JH6 usage and the tendency of lower CDR3 length, and argues the influence of antigen contact and T-cell help [8,31].
The high mutation rates were unexpected. This result is not likely to be related to selective amplification of specific sequences by our PCR protocol, since the high mutation rates were observed in all VH families amplified using different PCR conditions. Furthermore the detected repertoire was polyclonal, and even presented an extended number of VH genes. We nevertheless wanted to substantiate this result using a different approach. We therefore sorted single cells from this time point in CD27+ and CD27- B-cell subpopulations. The increase of mutational frequency was confirmed: again, the newly recirculating B cells showed increased mutation rates. This was detectable in both CD27+ and CD27- B cells (8.3% for CD27+ and 5.3% for CD27- versus 0.6% for the CD27- cells before therapy; Table 2). The fact that even the CD27- B cells were highly mutated was surprising, since CD27 is assumed to be a marker for memory B cells with mutated immunoglobulin receptors [33,34]. Mutated CD27- B cells have, however, been described in a study by Hansen and colleagues in patients with Sjögren's syndrome [35]. Reparon-Schuijt and colleagues also described, in the synovium of RA patients, a population of B lymphocytes that were functionally and phenotypically distinct from classic memory cells [36]. These cells were CD20+, CD38- and CD27-, and they produced immunoglobulins under induction but had a defective proliferative responsiveness.
To determine the heavy chain class distribution, we performed RT-PCR on total RNA from this time point (7 months), using primers specific for IgM, IgG and IgA. The IgM population was slightly mutated as expected (1.7%), but the IgG (9.0%) and IgA (8.9%) populations were highly mutated, in the same range as the genes amplified from genomic DNA (7.4%). This is in line with the assumption that the regenerating B cells at this time point were class-switched B cells.
To address the question of whether the circulation of highly mutated B cells in the early regeneration phase may be related to the anti-CD20 mediated B-cell depletion, a second patient was studied. We analysed the mutational frequencies in 75 VH4 rearrangements before treatment and 5 months and 6 months later, when B cells were again detectable in the periphery (Fig. 4b). The elevated mutation rate of expressed Ig-VH genes during the early regeneration phase was confirmed. Before B-cell depletion, similar to patient A, about 8% of the Ig-VH sequences were highly mutated. B-cell depletion in the periphery lasted 5 months in patient B. At the time point of 5 months, 2.4% of peripheral lymphocytes were CD19+ B cells and 46% of the analysed Ig-VH sequences were highly mutated. As in patient A, the phenomenon seems to be transient since a decrease in the mutation rate was already observed 4 weeks later in patient B. Only 22% of the rearrangements were highly mutated at this time point.
Although our findings are restricted to a small number of patients, the observed changes in the B-cell repertoire are highly probably related to the regeneration of B cells. We did not observe other possible confounding factors. During this phase, the patients had no change in their medication and did not show any clinical signs of infection. Also, the CRP levels did not change during this time period.
It is not known from which B-cell pool peripheral B cells regenerate after an anti-CD20-mediated B-cell depletion. A recent paper using a mouse model for anti-CD20-mediated immunotherapy demonstrates a hierarchy of B-cell sensitivities using rituximab-mediated B-cell depletion [37]. Particularly, germinal centre B cells and marginal zone B cells were more resistant to depletion in vivo. The microenvironment, such as resident macrophages, B-cell survival factors or circulatory dynamics of B-cell subsets, influences their sensitivity to anti-CD20-mediated depletion. It therefore seems probable that B-cell regeneration arises from a distorted composition of the mature B-cell compartment. The recirculating B cells during the early regeneration phase seem not to be newly generated cells, but are more probably resident cells that were resistant to the antibody treatment. These cells form the first wave of regenerating cells. Later B-cell repopulation is then taken over by newly produced naive cells. Alternatively, B-cell regeneration changes the local environment in the central immunological organs in a way that plasma cells, which usually home in the tissues or in bone marrow, recirculate in the periphery for a short time period.
Except for the highly mutated population observed during the early regeneration phase, the regenerated B-cell repertoire was overall relatively stable. This is in agreement with the data of Dijk-Hard and Lundkvist that followed the distribution of VH gene families in five individuals over a time period of 9 years [38]. A high degree of stability in the VHgene family repertoire was described in that paper, except for one individual where a changing pattern was observed that correlated with the presence of RF in serum at one time point.
Our study was not designed to detect specific disease-relevant changes in the repertoire. However, in addition to the treatment-induced reduction in disease activity, a specific decline in RF activity was observed. The patient showed a high RF activity before treatment. This activity decreased rapidly after the first B-cell depletion and rose again in relapse at 17 months. The second rituximab treatment again resulted in a significant fall of RF activity (Table 1). This rapid decrease in RF is in line with a report in a larger series of patients treated with rituximab [39]. The mechanism of a more selective effect on autoantibody production is not clearly understood. Possibly, RF-producing plasma cells are more dependent on the new regeneration from the CD20+ B-cell pool. Since we do not know precisely the Ig-VH genes used by RF-secreting B cells, it is not possible to relate RF to the use of Ig-VH genes. Nevertheless, there were distinct changes in the Ig-VH repertoire that paralleled the decrease in clinical activity – certain genes were shown to fluctuate, for instance the genes of the VH2 family or the gene VH3–23. The predominant gene of the VH3 family, VH3–23, was found in a high proportion before therapy, decreased 7 months after therapy, increased again 17 months after therapy, accompanied by a clinical relapse, and decreased again 27 months after the first therapy. It is also interesting to note that the use of JH6 segment in the VH3 rearrangements as well as the shorter CDR3s correlated with the disease activity – when the disease was active, the JH6 segment was found in a higher proportion accompanied by a shorter CDR3.
Regarding the VH4–34 gene already described to be frequently used in autoimmune disorders, it was significantly decreased with therapy and its frequency remained low after therapy. Irrespective of the possible pathogenic role of these changes, these results give evidence for a long-term modulation of the VH gene repertoire induced by anti-CD20 antibody treatment. Clonal expansion is a characteristic feature of the patients with RA. B-cell clones have been found in peripheral blood [13,14] and in synovial tissue [5,14]. In the present study, we were able to detect two clones within the VH5 family before therapy. The rearrangements used by these two clones were no longer observed after therapy at all studied time points up to 27 months. The inducible loss of clonal B cells is also an indication for the modulation of the B-cell repertoire. Even if their specificity is not known, the relevance of clonal B cells in disease activity can be speculated.
Conclusion
The present study describes the Ig-VH repertoire development after transient anti-CD20-mediated B-cell depletion. The results show that therapeutic, even repeated, transient B-cell depletions by anti-CD20 antibodies result in the regeneration of a diverse and polyclonal heavy-chain repertoire. The early phase of B-cell regeneration is characterised by the recirculation of highly mutated B cells during a short time period in both the patients analysed. The longitudinal observation of a single patient up to 27 months indicates subtle intraindividual changes, which cautiously favour the hypothesis of a therapeutic B-cell repertoire modulation.
Abbreviations
bp = base pair; CDR = complementary determining region; CRP = C-reactive protein; DAS28 = disease activity score; Ig-VH = immunoglobulin heavy-chain; PBMC = peripheral blood mononuclear cells; PCR = polymerase chain reaction; RA = rheumatoid arthritis; RF = rheumatoid factor.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
A-SR carried out the molecular genetic study, analysed the results and drafted the manuscript. CK participated in the design and coordination of the study, and performed the characterisation of cells. AP was involved in the molecular analysis of the second patient. TD carried out the single cell sorting and was involved in the molecular analysis. H-PT conceived the study, participated in its design and coordination. All authors read and approved the final manuscript.
Acknowledgements
The authors thank Kathrin Zehe and Karin Reiter for technical assistance, Dr Ioana Visan, Dr Martin Goller and Dr Martin Feuchtenberger for helpful discussions, and Prof. Vogt for his help with the statistical analyses. This work was supported by the IZKF of the University of Würzburg (BMBF 01KS 9603) and by the Graduate College 520 'Immunomodulation' in Würzburg.
Figures and Tables
Figure 1 Clinical response of rheumatoid arthritis (RA) patient A. The disease activity score (DAS28 index) and C-reactive protein (CRP) (mg/dl) levels of the RA patient. The patient was treated twice with rituximab (at 0 months and 17 months). Arrows indicate the percentage of B cells detected in peripheral blood for the four time points analysed (0, 7, 17 and 27 months)
Figure 2 Immunoglobulin heavy-chain gene (Ig-VH) distribution in peripheral B cells from patient A at the different time points (0, 7, 17 and 27 months): (a) VH1 genes, (b) VH2 genes, (c) VH3 genes, (d) VH4 genes and (e) VH5 genes. Results presented as the percentage of rearrangements expressing one particular gene within one VH family. * P < 0.05 using Fisher's exact test.
Figure 3 Genealogical trees of B-cell clones found before therapy in VH5 family in patient A. The best matching germline VH gene segments are shown in ellipses. The letters in the circles refer to individual sequences. Upper circle, parental clones with the gene segments they are using. Dotted circles, deduced intermediates. The numbers alongside the arrows represent the number of mutations between the different sequences. Brackets, mutated codons; underlined, replacement mutations; italicised, mutation to stop codon.
Figure 4 Mutational frequencies in VH rearrangements in (a) patient A and (b) patient B. * P < 0.0001 using the chi-square test.
Table 1 Rheumatoid factor values at different time points following the first anti-CD20 antibody therapy
Rheumatoid factor values (U/ml)
0 monthsa 2 months 5 months 7 months 13 months 17 months 27 months
Patient A 682 295 196 148 186 284 80
Patient B 196 124 130
aMonths after first therapy.
Table 2 Comparison of the overall mutational frequencies of VH rearrangements obtained from individual peripheral CD19+ (CD27- or CD27+) B cells of rheumatoid arthritis patient A before therapy and 7 months after first therapy
Before therapy 7 months after therapy
n Mutations (n) Total bp Mutational frequency (%) n Mutations (n) Total bp Mutational frequency (%)
CD19+CD27- 18 27 4,349 0.62 15 183 3,443 5.31*
CD19+CD27+ Not determined 22 477 5,745 8.30
* P < 0.0001 versus before therapy using chi-square test.
==== Refs
Lipsky PE Systemic lupus erythematosus: an autoimmune disease of B cell hyperactivity Nat Immunol 2001 2 764 766 11526379 10.1038/ni0901-764
Dorner T Burmester GR The role of B cells in rheumatoid arthritis: mechanisms and therapeutic targets Curr Opin Rheumatol 2003 15 246 252 12707577 10.1097/00002281-200305000-00011
Jonsson R Brokstad KA Lipsky PE Zouali M B-lymphocyte selection and autoimmunity Trends Immunol 2001 22 653 654 11738977 10.1016/S1471-4906(01)02052-X
Silverman G Carson D Roles of B cells in rheumatoid arthritis Arthritis Res Ther 2003 5 Suppl 4 S1 S6 15180890 10.1186/ar1010
Kim HJ Berek C B cells in rheumatoid arthritis Arthritis Res 2000 2 126 131 11094422 10.1186/ar77
Takemura S Klimiuk PA Braun A Goronzy JJ Weyand CM T cell activation in rheumatoid synovium is B cell dependent J Immunol 2001 167 4710 4718 11591802
Link JM Schroeder HW Jr Clues to the etiology of autoimmune diseases through analysis of immunoglobulin genes Arthritis Res 2002 4 80 83 11879542 10.1186/ar393
Dorner T Lipsky PE Abnormalities of B cell phenotype, immunoglobulin gene expression and the emergence of autoimmunity in Sjogren's syndrome Arthritis Res 2002 4 360 371 12453312 10.1186/ar603
Stott DI Hiepe F Hummel M Steinhauser G Berek C Antigen-driven clonal proliferation of B cells within the target tissue of an autoimmune disease. The salivary glands of patients with Sjogren's syndrome J Clin Invest 1998 102 938 946 9727062
Sims GP Shiono H Willcox N Stott DI Somatic hypermutation and selection of B cells in thymic germinal centers responding to acetylcholine receptor in myasthenia gravis J Immunol 2001 167 1935 1944 11489973
Hillorn V Soderstrom I Feld S Cilio C Forsgren S Hagg E Lundkvist I Holmberg D Aberrant V(H) gene utilization in patients with established insulin dependent diabetes mellitus J Autoimmun 1997 10 157 163 9185877 10.1006/jaut.1996.0117
Soderstrom I Dijk-Hard I Feld S Hillorn V Holmberg D Lundkvist I Altered VH6-D-JH repertoire in human insulin-dependent diabetes mellitus and autoimmune idiopathic thrombocytopenic purpura Eur J Immunol 1999 29 2853 2862 10508260 10.1002/(SICI)1521-4141(199909)29:09<2853::AID-IMMU2853>3.3.CO;2-P
Huang SC Jiang R Hufnagle WO Furst DE Wilske KR Milner EC VH usage and somatic hypermutation in peripheral blood B cells of patients with rheumatoid arthritis (RA) Clin Exp Immunol 1998 112 516 527 9649224 10.1046/j.1365-2249.1998.00580.x
Itoh K Patki V Furie RA Chartash EK Jain RI Lane L Asnis SE Chiorazzi N Clonal expansion is a characteristic feature of the B-cell repertoire of patients with rheumatoid arthritis Arthritis Res 2000 2 50 58 11219389 10.1186/ar68
Edwards JC Leandro MJ Cambridge G B-lymphocyte depletion therapy in rheumatoid arthritis and other autoimmune disorders Biochem Soc Trans 2002 30 824 828 12196207 10.1042/BST0300824
De Vita S Zaja F Sacco S De Candia A Fanin R Ferraccioli G Efficacy of selective B cell blockade in the treatment of rheumatoid arthritis: evidence for a pathogenetic role of B cells Arthritis Rheum 2002 46 2029 2033 12209504 10.1002/art.10467
Silverman GJ Weisman S Rituximab therapy and autoimmune disorders: prospects for anti-B cell therapy Arthritis Rheum 2003 48 1484 1492 12794814 10.1002/art.10947
Kneitz C Wilhelm M Tony HP Effective B cell depletion with rituximab in the treatment of autoimmune diseases Immunobiology 2002 206 519 527 12607727
Kneitz C Wilhelm M Tony HP Improvement of refractory rheumatoid arthritis after depletion of B cells Scand J Rheumatol 2004 33 1 8 15124935 10.1080/03009740310004379
Oligino T Dalrymple S Targeting B cells for the treatment of rheumatoid arthritis Arthritis Res Ther 2003 5 Suppl 4 S7 S11 15180891 10.1186/ar1006
Hainsworth J Safety of rituximab in the treatment of B cell malignancies: implications for rheumatoid arthritis Arthritis Res Ther 2003 5 Suppl 4 S12 S16 15180892 10.1186/ar1008
Gorman C Leandro M Isenberg D B cell depletion in autoimmune disease Arthritis Res Ther 2003 5 Suppl 4 S17 S21 15180893 10.1186/ar1007
Sacchi S Federico M Dastoli G Fiorani C Vinci G Clo V Casolari B Treatment of B-cell non-Hodgkin's lymphoma with anti CD 20 monoclonal antibody Rituximab Crit Rev Oncol Hematol 2001 37 13 25 11164715
Brezinschek HP Brezinschek RI Lipsky PE Analysis of the heavy chain repertoire of human peripheral B cells using single-cell polymerase chain reaction J Immunol 1995 155 190 202 7602095
Dorner T Brezinschek HP Brezinschek RI Foster SJ Domiati-Saad R Lipsky PE Analysis of the frequency and pattern of somatic mutations within nonproductively rearranged human variable heavy chain genes J Immunol 1997 158 2779 2789 9058813
Tomlinson IM Williams SC Corbett SJ Cox JBL Winter G V BASE Sequence Directory 1999 Cambridge: MRC Centre for Protein Engineering
Brezinschek HP Foster SJ Brezinschek RI Dorner T Domiati-Saad R Lipsky PE Analysis of the human VH gene repertoire. Differential effects of selection and somatic hypermutation on human peripheral CD5(+)/IgM+ and CD5(-)/IgM+ B cells J Clin Invest 1997 99 2488 2501 9153293
Stevenson FK Longhurst C Chapman CJ Ehrenstein M Spellerberg MB Hamblin TJ Ravirajan CT Latchman D Isenberg D Utilization of the VH4–21 gene segment by anti-DNA antibodies from patients with systemic lupus erythematosus J Autoimmun 1993 6 809 825 8155258 10.1006/jaut.1993.1066
van Vollenhoven RF Bieber MM Powell MJ Gupta PK Bhat NM Richards KL Albano SA Teng NN VH4–34 encoded antibodies in systemic lupus erythematosus: a specific diagnostic marker that correlates with clinical disease characteristics J Rheumatol 1999 26 1727 1733 10451069
Pascual V Victor K Spellerberg M Hamblin TJ Stevenson FK Capra JD VH restriction among human cold agglutinins. The VH4–21 gene segment is required to encode anti-I and anti-i specificities J Immunol 1992 149 2337 2344 1382098
Rosner K Winter DB Tarone RE Skovgaard GL Bohr VA Gearhart PJ Third complementarity-determining region of mutated VH immunoglobulin genes contains shorter V, D, J, P, and N components than non-mutated genes Immunology 2001 103 179 187 11412305 10.1046/j.1365-2567.2001.01220.x
Brezinschek HP Foster SJ Dorner T Brezinschek RI Lipsky PE Pairing of variable heavy and variable kappa chains in individual naive and memory B cells J Immunol 1998 160 4762 4767 9590222
Klein U Rajewsky K Kuppers R Human immunoglobulin (Ig)M+IgD+ peripheral blood B cells expressing the CD27 cell surface antigen carry somatically mutated variable region genes: CD27 as a general marker for somatically mutated (memory) B cells J Exp Med 1998 188 1679 1689 9802980 10.1084/jem.188.9.1679
Tangye SG Liu YJ Aversa G Phillips JH de Vries JE Identification of functional human splenic memory B cells by expression of CD148 and CD27 J Exp Med 1998 188 1691 1703 9802981 10.1084/jem.188.9.1691
Hansen A Odendahl M Reiter K Jacobi AM Feist E Scholze J Burmester GR Lipsky PE Dorner T Diminished peripheral blood memory B cells and accumulation of memory B cells in the salivary glands of patients with Sjogren's syndrome Arthritis Rheum 2002 46 2160 2171 12209521 10.1002/art.10445
Reparon-Schuijt CC van Esch WJ van Kooten C Ezendam NP Levarht EW Breedveld FC Verweij CL Presence of a population of CD20+, CD38- B lymphocytes with defective proliferative responsiveness in the synovial compartment of patients with rheumatoid arthritis Arthritis Rheum 2001 44 2029 2037 11592364 10.1002/1529-0131(200109)44:9<2029::AID-ART352>3.0.CO;2-2
Gong Q Ou Q Ye S Lee WP Cornelius J Diehl L Lin WY Hu Z Lu Y Chen Y Importance of cellular microenvironment and circulatory dynamics in B cell immunotherapy J Immunol 2005 174 817 826 15634903
Dijk-Hard I Lundkvist I Long-term kinetics of adult human antibody repertoires Immunology 2002 107 136 144 12225372 10.1046/j.1365-2567.2002.01466.x
Cambridge G Leandro MJ Edwards JCW Ehrenstein MR Salden M Bodman-Smith M Webster ADB Serologic changes following B lymphocyte depletion therapy for rheumatoid arthritis Arthritis Rheum 2003 48 2146 2154 12905467 10.1002/art.11181
| 15987473 | PMC1175025 | CC BY | 2021-01-04 16:02:37 | no | Arthritis Res Ther. 2005 Apr 1; 7(4):R714-R724 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1731 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17321598747510.1186/ar1732Research ArticleThe role of interleukin-1 in the pathogenesis of human Intervertebral disc degeneration Le Maitre Christine Lyn [email protected] Anthony J [email protected] Judith Alison [email protected] Division of Laboratory and Regenerative Medicine, School of Medicine, University of Manchester, Manchester, UK2005 1 4 2005 7 4 R732 R745 29 10 2004 3 12 2004 16 2 2005 Copyright © 2005 Le Maitre et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
In this study, we investigated the hypotheses that in human intervertebral disc (IVD) degeneration there is local production of the cytokine IL-1, and that this locally produced cytokine can induce the cellular and matrix changes of IVD degeneration. Immunohistochemistry was used to localize five members of the IL-1 family (IL-1α, IL-1β, IL-1Ra (IL-1 receptor antagonist), IL-1RI (IL-1 receptor, type I), and ICE (IL-1β-converting enzyme)) in non-degenerate and degenerate human IVDs. In addition, cells derived from non-degenerate and degenerate human IVDs were challenged with IL-1 agonists and the response was investigated using real-time PCR for a number of matrix-degrading enzymes, matrix proteins, and members of the IL-1 family.
This study has shown that native disc cells from non-degenerate and degenerate discs produced the IL-1 agonists, antagonist, the active receptor, and IL-1β-converting enzyme. In addition, immunopositivity for these proteins, with the exception of IL-1Ra, increased with severity of degeneration. We have also shown that IL-1 treatment of human IVD cells resulted in increased gene expression for the matrix-degrading enzymes (MMP 3 (matrix metalloproteinase 3), MMP 13 (matrix metalloproteinase 13), and ADAMTS-4 (a disintegrin and metalloproteinase with thrombospondin motifs)) and a decrease in the gene expression for matrix genes (aggrecan, collagen II, collagen I, and SOX6).
In conclusion we have shown that IL-1 is produced in the degenerate IVD. It is synthesized by native disc cells, and treatment of human disc cells with IL-1 induces an imbalance between catabolic and anabolic events, responses that represent the changes seen during disc degeneration. Therefore, inhibiting IL-1 could be an important therapeutic target for preventing and reversing disc degeneration.
==== Body
Introduction
Low back pain is a common, debilitating, and economically important disorder. Current evidence implicates loss of intervertebral disc (IVD) matrix consequent upon disc 'degeneration' as a major cause of low back pain [1]. Although many treatments aimed at relieving back pain are directed towards the degenerate IVDs (e.g. removal of protruding disc material, disc replacement, etc.), none of these are aimed at the processes of degeneration. Modern advances in therapeutics, particularly cell and tissue engineering, offer potential methods for inhibiting or reversing IVD degeneration that have not previously been possible, but they require a level of understanding of the pathobiology of degeneration of the IVDs that is not currently available [2].
Degeneration is characterized by increased degradation of the normal IVD matrix by locally produced matrix metalloproteinases (MMPs) and ADAMTS (a disintegrin and metalloproteinase with thrombospondin motifs) [3-6]. In addition, the nature of the matrix produced in the degenerate IVDs differs from that in normal IVDs, as a consequence of switches in the production of collagen within the inner annulus fibrosus (IAF), and nucleus pulposus (NP) from type II to type I [7] and in the synthesis of proteoglycan from aggrecan [8] to versican, biglycan, and decorin [9,10]. The resultant changes within the extracellular matrix have a number of consequences, resulting in loss of structural integrity, decreased hydration, and a reduced ability to withstand load.
Similar matrix changes have been reported in articular cartilage in osteoarthritis [11,12]. In this disease, the body of evidence points towards these being part of a more profound change in chondrocyte biosynthesis [13] driven by local production of IL-1 and tumour necrosis factor α [14-17]. Despite the similarities between IVD degeneration and the cartilage changes in osteoarthritis, there has been relatively little interest in exploring the possibility that the disease processes involved in IVD degeneration might be driven by similar alterations in local tissue cytokine biology, and particularly by IL-1 and tumour necrosis factor α. TNF α has been implicated in disc herniation and sciatic pain [18-21], but not in disc degeneration. There is, however, some circumstantial evidence implicating IL-1 in human IVD degeneration [22-26]. This evidence comes from studies on annulus fibrosus (AF) cells from rabbit IVDs [24,26,27] and NP cells from ovine [25] and rabbit IVDs [28], which suggest that IL-1 may have similar effects on the chondrocyte-like cells of IVDs to those seen in articular chondrocytes. IL-1 has been identified in herniated, displaced human discal tissue [23,29,30] but has not been investigated within the degenerate IVDs themselves. Two recent genetic studies suggest that IL-1 gene cluster polymorphisms contribute to the pathogenesis of lumbar IVD degeneration and low back pain [31,32]. Despite these data, there is no clear evidence that IL-1 is synthesized by native human disc cells (as opposed to cells within herniated disc tissue) or whether it can induce the altered synthesis of matrix molecules and degrading enzyme production by human IVD cells characteristic of IVD degeneration, particularly in the NP, where degenerative changes first appear.
This study investigates two hypotheses: that in human IVD degeneration, there is local production of the cytokine IL-1 by native disc cells, and that locally produced IL-1 can induce the cellular and matrix changes of IVD degeneration.
Materials and methods
Tissue samples
Human IVD tissue was obtained either at surgery or at post-mortem examination, with the informed consent of the patient or relatives. Local research ethics committee approval was given for this work by the following local research ethics committees: Salford and Trafford (Project number 01049), Bury and Rochdale (BRLREC 175(a) and (b)), Central Manchester (Ref No: C/01/008), and her Majesty's coroner (LMG/RJ/M6).
Tissue samples for Immunohistochemical analysis
Post-mortem tissue
Preliminary studies from our laboratory (data not shown) have shown that IVD cells remain viable for at least 48 hours after death. We also have evidence that NP cells from retrieved cadaveric IVDs are biosynthetically identical to age-matched cells from non-cadaveric tissue, an observation borne out by others [4,33,34]. In all, eight discs recovered from six patients within 18 hours of death were used in this study (Table 1). They consisted of full-thickness wedges of IVD of 120° of arc removed anteriorly. This allowed well-orientated blocks of tissue incorporating AF and NP to be cut for histological study. The family practitioner's notes were examined for evidence of a history of sciatica sufficient to warrant seeking medical opinion, and such patients were excluded from the study.
Surgical tissue
Patients were selected on the basis of IVD degeneration diagnosed by magnetic resonance imaging and progression to anterior resection either for spinal fusion or disc replacement surgery to relieve chronic low back pain. Some patients underwent fusion at more than one level, because of instability. Occasionally the specimens retrieved from multilevel fusion included discs with low (0–3) degeneration scores (i.e. morphologically normal) at one level (Table 1) (The scoring system is described below). Wedges of disc tissue were removed in a manner similar to that described for cadaveric tissue.
Treatment of tissue specimens
A block of tissue incorporating AF and NP in continuity was fixed and processed into paraffin wax. As some specimens contained bone, all the samples were decalcified in ethylenediaminetetraacetic acid (EDTA) (we have previously shown that EDTA decalcification does not affect detectable levels of product using in situ hybridization or immunohistochemical staining [35] when compared to snap-frozen tissue). Sections from the tissue blocks were taken for H&E staining to score the degree of morphological degeneration according to previously published criteria [8]. This scoring system provided a representation of the grade of degeneration within a disc: scores of 0 to 3 represent a histologically normal (non-degenerate) disc; 4 to 6, histological evidence of low-level degeneration; 7 to 9, an intermediate degree of degeneration; and 10 to 12, severe degeneration. From this scoring, 30 discs were selected to represent a range of scores from non-degenerate (1 to 3) up to the most severe level of degeneration (12).
Tissue samples for in vitro cell studies
Samples of degenerate IVD tissue (graded 6 to 10) were obtained from patients undergoing surgery for disc replacement for the treatment of chronic low back pain. Non-degenerate IVD tissue (graded 0 to 2) was also obtained from surgery for disc removal after trauma. Ten discs were used in triplicate for all treatments; all discs were lumbar in origin and the ages of the patients ranged from 18 to 44 years (mean 29.9).
Production and localization of IL-1 family proteins
Immunohistochemistry was used to localize the two IL-1 agonists (IL-1α and IL-1β) and their antagonist IL-1Ra together with the active receptor IL-1RI (IL-1 receptor, type I) and the IL-1β-converting enzyme (ICE; caspase-1) within the 30 disc samples described in Table 1. In addition, rheumatoid synovium was selected as a positive control tissue for members of the IL-1 family. The immunohistochemistry protocol followed was as previously published [6]. Briefly, 4-μm wax sections were dewaxed and rehydrated, and endogenous peroxidase was blocked using hydrogen peroxide. Sections were washed in dH2O and then treated with chymotypsin enzyme antigen retrieval system (0.01% w/v chymotrypsin (Sigma, Poole, Dorset, UK), 20 min at 37°C) for IL-1α, IL-1β, IL-1Ra, and ICE. No enzyme retrieval was necessary for IL-1RI. After washing, non-specific binding sites were blocked at room temperature for 45 min, either with 20% w/v rabbit serum (Sigma), for IL-1Ra and IL-1RI, or with 20% w/v donkey serum (Sigma), for IL-1α, IL-1β, and ICE. Sections were incubated overnight at 4°C with mouse monoclonal primary antibodies against human IL-1Ra (1:200 dilution, R&D Systems, Abingdon, UK), IL-1RI (1:50 dilution, R&D Systems), and goat polyclonal primary antibodies against human IL-1α (1:300 dilution, Santa Cruz Biotechnology, Santa Cruz, CA, USA), IL-1β (1:300 dilution, SantaCruz), and ICE (1:10 dilution, SantaCruz). Negative controls in which mouse or goat IgGs (Dako, Cambridgeshire, UK) replaced the primary antibody (at an equal protein concentration) were used.
Following washes, sections reacted with mouse monoclonal antibodies were incubated in a 1:400 dilution of biotinylated rabbit anti-mouse antiserum (Dako), and sections reacted with goat polyclonal primary antibodies were incubated in a 1:300 dilution of biotinylated donkey anti-goat antiserum (Santa Cruz Biotechnology), all for 30 min at room temperature. Disclosure of secondary antibody binding was by the streptavidin-biotin complex (Dako) technique with 3,3'-diaminobenzidine tetrahydrochloride solution (Sigma). Sections were counterstained with Mayer's haematoxylin (Raymond A Lamb, East Sussex, UK), dehydrated, and mounted in XAM (BDH, Liverpool, UK).
Image analysis
All slides were visualized using a Leica (Leica, Cambridge, UK) RMDB research microscope and images captured using a digital camera and Bioquant Nova image analysis system (Bioquant, Nashville, TN, USA). Each section was divided into three areas of disc for the purposes of analysis – the NP, the Inner annulus fibrosus (IAF), and, where present, the outer annulus fibrosus (OAF) – and analysed separately. Within each area, 200 cells were counted and the number of immunopositive cells (brown-stained cells) expressed as a proportion of this. Averages and standard deviations were calculated for disc sections grouped with the scores 0 to 3, 4 to 6, 7 to 9, and 10 to 12. Data was then presented on graphs as means ± 2 standard errors to represent the 95% confidence intervals [36].
Statistical analysis
Data was non-parametric, and hence the Mann-Whitney U tests were used to compare the numbers of immunopositive cells in degenerate discs (groups 4 to 6, 7 to 9, and 10 to 12) with those in non-degenerate discs (scores 0 to 3). These tests were performed for each area of the disc analysed (i.e. NP, IAF, and OAF). In addition, the Wilcoxon paired samples tests were used to compare proportions of immunopositive cells in the different areas of the discs (i.e. NP vs IAF, NP vs OAF, and IAF vs OAF). This analysis was performed using all disc sections, regardless of level of degeneration.
Investigation of the effect of IL-1 on human IVD cells in alginate culture
Issolation of Disc cells
Tissue samples were separated into NP and IAF tissue and transported to the laboratory in DMEM and Ham's F12 nutrient medium (DMEM + F12) (Gibco BRL, Paisley, UK) on ice. Tissue samples were finely minced and digested with 2 U/ml protease (Sigma) in DMEM + F12 media for 30 min at 37°C and washed twice in DMEM + F12. NP cells were isolated in 0.4 mg/ml collagenase type 1 (Gibco), and AF cells in 2 mg/ml collagenase type 1 (Gibco) for 4 hours at 37°C.
Alginate bead culture
It is well recognized that cells derived from the IVDs change their morphology and phenotype in monolayer culture, becoming similar to fibroblasts [37]. However, culturing the cells in systems such as alginate can restore the IVD cell phenotype [37]. We have therefore used cells in alginate gels to investigate the effects of IL-1 on cell behaviour. To achieve this, following isolation, cells were expanded in monolayer culture for 2 weeks, prior to trypsinization and resuspension in 1.2% medium-viscosity sodium alginate (Sigma) in 0.15 M NaCl at a density of 1 × 106 cells/ml and formation of alginate beads using 200 mM CaCl2. Following washes in 0.15 M NaCl, 2 ml of complete culture medium was then added to each well and cultures were maintained at 37°C in a humidified atmosphere containing 5% CO2. The culture medium was changed every other day.
Assessment of re-differentiated state in alginate
To ensure that the phenotype of cells treated with IL-1 were similar to the phenotype of cells within the IVDs in vivo, the cell phenotype was assessed in monolayer culture and at increasing times in alginate culture. The phenotype was then compared with that of uncultured, directly extracted cells. Phenotype was assessed using immunohistochemistry on cellular cytospins for directly extracted and monolayer cells, and wax-embedded alginate beads sectioned at 4 μm and mounted onto slides for analysis. Immunohistochemistry was performed for aggrecan, collagen type II, and collagen type I as described previously [38]. In addition, RNA was extracted from cells and reverse transcription performed using Avian Myeloblastosis Virus (AMV) reverse transcriptase (Roche, East Sussex, UK), and gene expression for the chondrogenic transcription factor SOX9 and the matrix constituents aggrecan, collagen II, and collagen type I were assessed (see below).
Image analysis
All slides were visualized using the Leica RMDB research microscope and images were captured using a digital camera and the Bioquant Nova image analysis system. Within each area, 200 cells were counted and the number of immunopositive cells was expressed as a proportion of this.
Statistical analysis
One-way ANOVA and Tukey post hoc tests were used to compare cellular gene expression of cells cultured in monolayer and alginate to uncultured, directly extracted cells. To perform this analysis, 2-ΔCt (where Ct represents the cycle at which the set threshold is reached) for each sample was calculated to generate relative gene expression for each sample, including all control values. These values were then used in ANOVA and post hoc tests.
Treatment of cells with IL-1, RNA extraction, and cDNA formation
After 4 weeks in this culture system (the time required to allow redifferentiation to the same phenotype as that of uncultured, directly extracted disc cells), cells were treated for 48 hours with either 10 ng/ml IL-1α or 10 ng/ml IL-1β, or were left untreated to serve as controls; all treatments were performed in triplicate. Following treatment, RNA was extracted using Trizol reagent (Gibco). Prior to Trizol extraction, alginate beads were washed in 0.15 M NaCl and dissolved in dissolving buffer (55 mM sodium citrate, 30 mM EDTA, 0.15 M NaCl; pH 6) at 37°C for 15 min and then were subsequently digested in 0.06% w/v collagenase type I (Gibco) for 30 min to allow digestion of matrix. Following RNA extraction, reverse transcription was performed as described previously.
Real-time PCR
Real-time PCR was used to investigate the effects of IL-1 on a range of targets, namely, the members of the IL-1 family (IL-1α, IL-1β, IL-1Ra, and IL-1RI), matrix-degrading enzymes (MMP-3, MMP-13, ADAMTS-4, and ADAMTS-5), matrix proteins (aggrecan and collagen types I and II), and two SOX genes (6 and 9). Primers and Probes for all of these targets were designed using the Primer Express computer program (Applied Biosystems, Warrington, UK), using the rules of primer design recommended by Applied Biosystems. The total gene specificity of the nucleotide sequences chosen for the primers and probes were confirmed by BLAST searches (GenBank database sequences). The nucleotide sequences of the oligonucleotide hybridization primers and probes are shown in Table 2. Primers and probes were purchased from Applied Biosystems, as were pre-designed primers and probe (PDAR) for human glyceraldehyde-3-phosphate dehydrogenase (GAPDH). For each set of primers and probes, the efficiency of the amplification was assessed using template titrations as recommended by Applied Biosystems.
PCR reactions were then performed and monitored using the ABI Prism 7700 Sequence Detection System (Applied Biosystems). The PCR master mix was based on the AmpliTaq Gold DNA polymerase (Applied Biosystems). cDNA samples (2.5 μl in a total of 25 μl per well) were analysed in duplicate; primers were used at a concentration of 900 nmol/l and probe at 250 nmol/l. After real-time amplification, the ABI 7700 expressed the data as an amplification plot, from which a baseline was set from cycle number 3 upto a few cycles before the first visible amplification. In addition to the baseline, the threshold was set at a level above background levels and within the exponential phase of the PCR amplification. The same threshold was used for a target between runs. The Ct values for each target gene (cycle at which the set threshold is reached) were then exported into an Excel file, where analysis was performed using the 2-ΔΔCt method, using GAPDH as the housekeeping gene, and normalized to untreated controls [39].
Statistical analysis
One-way ANOVA and Tukey post hoc tests were used to compare cells treated with IL-1 with those untreated samples. To perform this analysis, 2-ΔΔCt for each sample was calculated using an average of untreated control ΔCt values to generate the relative gene expression for each sample, including all control values. These values were then used in ANOVA and post hoc tests; each treatment group was compared with untreated controls.
Results
Immunohistochemical localisation
Immunoreactivity for the five molecules studied (IL-1α, IL-1β, IL-1Ra, IL-1RI, and ICE) was seen in degenerate and non-degenerate IVDs. The immunostaining was generally restricted to the cytoplasm of native disc cells in normal and degenerate discs (Fig. 1). Staining was particularly prominent in the cytoplasm of the chondrocyte-like cells of the NP and IAF. No significant difference was observed between the proportions of cells in the NP and IAF reacting for IL-1α, IL-1Ra, and ICE (P = 1.525, 0.870, and 0.639, respectively). IL-1β and IL-1RI immunopositive cells were more frequent in the NP than the IAF (IL-1β, P < 0.05; IL-1RI, P < 0.05)).
IgG controls were always negative and all positive controls showed strong immunoreactivity (Fig. 1). No immunopositivity was observed in the matrix of the IVDs or in blood vessels, with the exception of immunopositivity for ICE, which showed some staining in the matrix and blood vessels of the most degenerate discs (histological degenerative scores 10 to 12). Although cells in the OAF did show reactivity for all molecules, the proportion was always significantly lower than in the NP and IAF (All targets P < 0.05) (Fig. 2).
Immunohistochemical staining and quantification of immunopositive cells
The most prominent aspects of the immunophenotype of non-degenerate discs (histological degeneration scores 0 to 3) included: little immunoreactivity for any of the five molecules in the OAF; low proportions of cells immunopositive for IL-1α, IL-1β, and IL-1RI in the NP and IAF (approximately 20%); the presence of IL-1Ra immunopositive cells in every disc, with high proportions of cells of up to 40% showing immunopositivity in the NP and IAF; and high numbers of cells in the NP and IAF also showing immunopositivity for ICE (60%) (Fig. 2).
In the degenerate IVDs (histological degenerative scores 4 to 12), the immunophenotype of cells differed in two ways from cells in non-degenerate discs (scores 0 to 3). Firstly, the proportion of cells immunopositive for IL-1α, IL-1β, IL-1RI, and ICE in both the NP and IAF was two or three times that in cells from non-degenerate IVDs, and this immunopositivity increased with the severity of degeneration. The difference between the degenerate and non-degenerate samples was significant in the NP and IAF in a number of stages of histological degeneration: IL-1α (NP and IAF: non-degenerate vs degenerate grades 10–12, P < 0.05); IL-1β (NP and IAF: non-degenerate vs three degrees of degeneration (scores 4 to 6, 7 to 9, and 10 to 12), all P < 0.05); IL-1RI (NP: non-degenerate vs three degrees of degeneration, all P < 0.05; IAF: non-degenerate vs severe grades of degeneration (scores 10 to 12), P < 0.05); ICE (NP: non-degenerate vs severe grades of degeneration, P < 0.05; IAF: non-degenerate vs severe grades of degeneration, P < 0.05). Secondly, similar numbers of IL-1Ra-immunopositive disc cells were seen in levels of degeneration scoring 4 to 6 and 7 to 9 and in non-degenerate discs, but in severe degeneration (scores 10 to 12), a significant decrease in the proportion of cells with IL-1Ra-immunopositivity was seen compared toI that seen in non-degenerate discs (P < 0.05) (Fig. 2).
Assessment of redifferentiated state in alginate
NP and AF cells directly extracted from IVD tissue showed similar morphology and phenotypic characteristics. Morphologically, the cells were small and rounded, often (in cells from degenerate discs) localized in clusters. Immunopositivity for aggrecan and collagen type II was seen, but no cells immunopositive for collagen type I were observed (Fig. 3a). In monolayer, these cells adhered and spread, developing a fibroblastic morphology, together with loss of immunopositivity for aggrecan and collagen type II, and they expressed collagen type I protein (Fig. 3b). However, when transferred to alginate and cultured for 4 weeks, these cells regained their rounded morphology and began to produce aggrecan and collagen type II protein, and lost their immunopositivity to collagen type I (Fig. 3c), resembling the immunohistochemical profile of uncultured, directly extracted cells. Gene expression analysis showed a similar pattern to protein production in monolayer and alginate cultures, with 4 weeks' culture in alginate required before gene expression levels returned to that seen in uncultured, directly extracted cells (P > 0.05) (data not shown). No significant difference was observed in the re-differentiation potential of cells extracted from NP or from AF cells, or between cells extracted from non-degenerate or from degenerate IVDs.
Effect of IL-1 on human IVD cells
Interleukin 1 treatment (IL-1α and IL-1β) of the four cell types/origins (degenerate and non-degenerate cells, from AF or NP) resulted in altered in expression of genes for matrix molecules and matrix-degrading enzymes. The responses of cells to IL-1α and IL-1β were similar, and hence only the effects of IL-1β are detailed here. Although it can be generally summarized that IL-1 caused an increase in gene expression for matrix-degrading enzymes, particularly in cells derived from the degenerate NP, and caused a decrease in normal matrix molecule gene expression in cells derived from normal discs, the pattern was complex and dependent upon the origin of the cells (Table 3).
Effect of IL-1 on degradative enzymes
Following treatment with IL-1, an increase in MMP-3 gene expression was seen in the four cell types investigated (though the increase was significant only in cells derived from the non-degenerate NP and AF (P < 0.05)) (Fig. 4a). An increase in MMP-13 gene expression was also observed, but only in cells derived from the NP, with significance achieved in cells from non-degenerate discs (P < 0.05) (Fig. 4b). Aggrecanase (ADAMTS-4 and -5) gene expression was increased in cells derived from the NP of degenerate discs. This was significant only for ADAMTS-4 (P < 0.05). In cells derived from the non-degenerate discs, a slight, nonsignificant decrease in aggrecanase gene expression was observed (Fig. 4c,d).
Effect of IL-1 on matrix molecules
IL-1 treatment of cells derived from non-degenerate discs resulted in a decrease in both SOX6 and SOX9 gene expression. However, this achieved significance only for SOX6 (P < 0.05). No real effect was observed on SOX6 and SOX9 gene expression in cells derived from degenerate discs (Fig. 5a,b). A decrease was also observed in expression of the gene for collagen type I in cells derived from non-degenerate AF and degenerate NP; however this was significant only in cells derived from degenerate NP (P < 0.05) (Fig. 5c). The expression of the genes for collagen type II and aggrecan were decreased by IL-1 treatment of cells derived from the non-degenerate disc, although this decrease was only significant for aggrecan (Fig. 5d,e).
IL-1 regulation
IL-1 treatment of cells derived from the degenerate but not the non-degenerate disc resulted in a 100-fold increase in IL-1α and IL-1β gene expression, which reached significance in cells derived from the NP (P < 0.05) (Fig. 6a,b). No real trend was observed in IL-1Ra gene expression after treatment with IL-1 (Fig. 6c). A 10-fold decrease in IL-1 receptor gene expression was observed in cells derived from the non-degenerate AF, but this was not significant and no effect was observed on the other cell types (Fig. 6d).
Discussion
In this study, we investigated whether in IVD degeneration there is local production of the cytokine IL-1 and whether IL-1 could induce the cellular changes characteristic of IVD degeneration. To date, the production of IL-1 by human IVD cells has been shown only in cells derived from herniated tissue [18,19,29,30,40]. However, herniated tissue is not representative of native disc tissue and is usually contaminated with inflammatory cells. For example, Doita and colleagues localized production of IL-1 to infiltrating mononuclear cells within sequestered and extruded disc tissue but did not show any significant immunodetectable IL-1 in connective tissue cells in the displaced IVDs [29]. The current study is the first to investigate protein production and localization of IL-1 in intact, non-degenerate and degenerate human IVDs themselves, as opposed to herniated disc tissue.
This study has shown that both isoforms of IL-1 (IL-1α and IL-1β) are produced by the chondrocyte-like cells of the NP and IAF (but not blood vessels or fibroblast like cells in the OAF) of non-degenerate and degenerate IVDs. Furthermore, chondrocyte-like cells in non-degenerate IVDs express and produce the active receptor IL-1RI, indicating that they can respond to IL-1. Importantly, in degenerate IVDs there is a significant increase in IL-1RI-immunopositive chondrocyte-like cells by comparison with non-degenerate IVDs, indicating an increased responsiveness to IL-1; and there are increased numbers of chondrocyte-like cells expressing ICE, an enzyme required to convert the inactive pro-IL-1β into its active form [41].
This study demonstrated IL-1Ra protein localization to cells in both non-degenerate and degenerate human IVDs. The production of IL-1Ra in the non-degenerate disc demonstrates a means of regulating IL-1. Within most clinical conditions involving IL-1, an increase in IL-1Ra production is considered an excellent marker of disease, and often a better indicator than IL-1 itself [42]. For example, in rheumatoid arthritis, raised IL-1Ra production is considered to be a natural compensatory mechanism to counter the activity of IL-1 [43]. In the current study, a marked increase in the proportion of cells immunoreactive for IL-1 were found in degenerate than in non-degenerate IVDs, but no similar increase in IL-1Ra-immunopositive cells was observed, indicating an imbalance in the local production of IL-1 and IL-1Ra and failure of the normal compensatory mechanism associated with increasing local production of IL-1. When coupled with an increase in IL-1 receptor and ICE with increasing degeneration, the net effect would be the initiation and perpetuation of an IL-1-mediated response.
Having established a basis for a functional excess of IL-1 in degenerate IVDs, we then investigated the role of IL-1 in the processes that characterize disc degeneration, namely, decreased matrix synthesis and increased production of MMPs and ADAMTS-4 [3-6]. This is the first time such a comprehensive study has been undertaken in human IVD cells. Such limited studies as have been conducted previously on IVD cells have focused on cell monolayers and have not used human cells [24,26,27]. However, it is well known that cells in monolayer culture dedifferentiate and therefore effects may be very different from those in vivo. Culture of cells in 3D gels such as alginate allows the phenotype of IVD chondrocyte-like cells to be maintained [37,44-46]. To date, only two studies have investigated the effects of IL-1 in such systems, one using ovine IVD cells [25] and the other, rabbit IVD cells [28]. This is the first reported study to investigate the effects of IL-1 on human disc cells cultured in 3D gels.
Effect of IL-1 on degradative enzymes
In the current study, MMP-3 mRNA expression was increased in NP and AF cells derived from non-degenerate and degenerate IVDs after IL-1 treatment, a phenomenon reported in rabbit disc cells cultured in monolayer [27] and ovine NP cells cultured in agarose [25]. Therefore, in vitro IL-1 causes an increase expression of MMP-3, an enzyme increased in the degenerate disc [6]
Treatment of NP (but not AF) cells from degenerate and non-degenerate IVDs with IL-1 resulted in significant increases in gene expression of MMP-13 (an MMP with high affinity for type II collagen), a finding not previously reported in disc cells, although it has been shown in articular chondrocytes [16,47,48]. We have previously shown that immunodetectable MMP-13 protein is present in significant amounts in IVDs, with the highest immunopositivity in the NP of degenerate discs [6], an area of the IVD containing the highest concentration of collagen type II.
ADAMTS-5 gene expression was not significantly altered by IL-1 treatment. However, such treatment did result in an increase in the gene expression of the aggrecanase ADAMTS-4 in cells derived from degenerate NP. In vivo, the NP contains the highest concentration of aggrecan in the IVD. The response of cells derived from degenerate NP to IL-1 to up-regulate ADAMTS-4 indicates that in vivo a local increase in the concentration of IL-1 might lead to the dehydration and loss of height characteristic of IVD degeneration, through the production of aggrecanases by local cells. We have previously shown an increase in ADAMTS-4 production by the cells of degenerate discs, especially in the NP [6], which, interestingly, were the same discs shown in this study to produce high levels of IL-1 agonists.
Effect of IL-1 on matrix molecules
Degeneration of the IVD is associated with altered collagen production by IVD cells, with a switch in synthesis from type II to type I collagen in the IAF and NP [7]. Proteoglycan production is also altered, with decreased aggrecan [8] and increased production of versican, biglycan, and decorin [9,10]. IL-1 has been implicated in changes in matrix synthesis during degradation of articular cartilage, with studies showing a down-regulation of the genes for SOX9 and collagen type IIa1 [49], aggrecan, collagen types II and XI, and link proteins [48,50], and inhibition of normal matrix assembly [51]. The few studies performed on IVD cells have shown that IL-1 treatment also causes a decrease in proteoglycan and collagen II production in animal cells [24,26-28]. The current study demonstrates that IL-1 decreases expression of the gene for SOX6 by cells of the non-degenerate IVD. SOX6 (usually in combination with SOX9, which was also decreased by IL-1 [albeit not significantly] in this study) determines the chondrogenic phenotype [52]. Such results suggest that IL-1 can inhibit the innate regulator of the chondrocyte-like cells' chondrogenic phenotype, resulting in IVD cells, particularly in the NP, that develop a less differentiated and more fibroblastic phenotype. This inhibition of the SOX genes might lead to the altered collagen and aggrecan synthesis typical of IVD degeneration [7-10,53]. The current study also demonstrated that IL-1 inhibited expression of the genes for collagen types I and II and for aggrecan. This would mean that within the NP, at least, IL-1 can exert its effects on biosynthesis in the same way as it does in articular cartilage [49].
Interestingly, this study has also shown that the cells derived from degenerate and non-degenerate discs respond differently to IL-1. In particular, cells from degenerate IVDs respond to IL-1 with a further increase in IL-1 gene expression (i.e. there is a positive autocrine effect), while cells from non-degenerate discs showed a decrease, suggesting that the normal homeostatic response to IL-1 is replaced in the degenerate IVD by a positive feedback loop. Such a phenomenon has also been reported in human skin fibroblasts treated with IL-1 [54]. This positive feedback loop in degenerate disc cells clearly distinguishes them from non-degenerate disc cells.
Conclusion
We have shown that IL-1 is produced in the degenerate IVD. It is normally synthesized by the native chondrocyte-like cells, but in the non-degenerate IVD there is a balance between IL-1 and its inhibitor, IL-1Ra, ensuring that matrix homeostasis is maintained. Treatment of human IVD cells with IL-1 disturbs the normal balance of catabolic and anabolic events, with the result that degrading enzymes are increased and the expression of genes for matrix proteins are decreased, responses that correspond to the alterations of cell biology that characterize IVD degeneration. In addition, the immunohistochemical data from this study demonstrated that although numbers of cells with immunopositivity for the IL-1 agonists increased with degeneration, no such increase was seen in the numbers of cells with immunopositivity for IL-1Ra. This finding suggests that the normal inhibitory mechanism fails in disc degeneration, with a loss in the balance of IL-1 agonists to antagonists, allowing IL-1 to elicit and perpetuate a response. We have also shown that cells from non-degenerate and degenerate discs respond differently to IL-1. In particular, IL-1 causes cells from degenerate IVDs to synthesize more IL-1, with the potential to induce accelerating degeneration.
This study has shown how IL-1, a naturally occurring cytokine within the IVD, could, through an imbalance between it and its inhibitor, play a role in the pathogenesis of IVD degeneration and therefore be an important therapeutic target for preventing and reversing disc degeneration.
Abbreviations
ADAMTS = a disintegrin and metalloproteinase with thrombospondin motifs; AF = annulus fibrosus; DMEM + F12 = Dulbecco's modified Eagle's medium and Ham's F12 nutrient medium; EDTA = ethylenediaminetetraacetic acid; GAPDH = glyceraldehyde-3-phosphate dehydrogenase; H&E = haematoxylin and eosin; IAF = inner annulus fibrosus; ICE = IL-1β-converting enzyme; IL-1 = interleukin-1; IL-1Ra = IL-1 receptor antagonist; IL-RI = IL-1 receptor, type I; IVD = intervertebral disc; MMP = matrix metalloproteinase; NP = nucleus pulposus; OAF = outer annulus fibrosus.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
CLM participated in the design of the study, performed all laboratory work and analysis, and drafted the manuscript. AJF conceived the study and participated in its design and coordination. JAH conceived the study, participated in its design and coordination, and assisted in its analysis. All authors read and approved the final manuscript.
Acknowledgements
The authors wish to acknowledge the support of the Wellcome Trust (SHoWCASe awards 057601/Z/99 and 063022/Z/O) and a Back Care grant. The work was undertaken in the Human Tissue Profiling Laboratories of the Division of Laboratory and Regenerative Medicine that receive core support from the ARC (ICAC grant F0551) and MRC (Co-operative Group Grant G9900933) and the joint Research Councils (MRC, BBSRC, EPSRC) UK Centre for Tissue Engineering (34/TIE 13617). The authors wish to thank the surgeons Mr ERS Ross and Mr B Williamson and Mr Balamuri, Hope Hospital, Salford for supply of tissue samples.
Figures and Tables
Figure 1 Examples of imunohistochemical staining for the IL-1 family. IL-1β (row A), IL-1Ra (row B), and IL-1 receptor, type I (row C) in grade-1 non-degenerate discs (column 1) and grade-12 degenerate discs (column 2), IgG controls (row D) were all negative. Immunopositivity is revealed by brown staining. N.B In non-degenerate discs, no cell clusters were seen and little immunopositivity was observed in the single cells. In degenerate discs, a large number of cell clusters were observed, which were predominately immunopositive. Bars = 570 μm.
Figure 2 Immunopositive staining for the IL-1 family in human intervertebral discs. Numbers of cells with immunopositivity for IL-1α (a), IL-1β (b), IL-1 receptor antagonist (c), IL-1 receptor, type I (d), and IL-1β-converting enzyme (e), according to place of origin in the disc and grade of intervertebral disc degeneration (n = 30). Data are presented as means ± 2 standard errors (as a representative of 95%CI). *P < 0.1,; **P < 0.05
Figure 3 Immunopositive staining for phenotypic markers in chondrocyte-like cells from human intervertebral discs. Immunohistochemical staining for collagen type II, aggrecan, and collagen type I in uncultured directly extracted cells (a), cells cultured in monolayer for 2 weeks and cytospun prior to staining (b), and cells cultured in monolayer for 2 weeks prior to transfer to alginate and then cultured for a further 4 weeks (c). Immunopositivityis revealed bybrown staining. Data shown are from cells derived from degenerate discs, but results were similar in non-degenerate discs. Bars = 570 μm. DE, directly extracted.
Figure 4 Effect of IL-1 on MMP and ADAMTS gene expression in cells from human intervertebral discs. Relative gene expression was normalized to that of the GAPDH (glyceraldehyde-3-phosphate dehydrogenase) housekeeping gene and untreated controls (hence control is graphed at 1 on the log scale) for matrix metalloproteinase (MMP)-3 (a), MMP-13 (b), ADAMTS (a disintegrin and metalloproteinase with thrombospondin motifs)-4 (c), and ADAMTS-5 (d) following IL-1β treatment of cells from two regions of non-degenerate (non-deg) (n = 6) and degenerate (n = 24) discs. **P < 0.05. AF, annulus fibrosus; NP, nucleus pulposus.
Figure 5 Effect of IL-1 treatment on matrix gene expression in cells from human intervertebral discs. Relative gene expression was normalized to the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) housekeeping gene and untreated controls (hence control is graphed at 1 on the log scale) for SOX6 (a), SOX9 (b), collagen type I (c), collagen type II (d), and aggrecan (e) following IL-1β treatment of disc cells from two regions of non-degenerate (non-deg) (n = 6) and degenerate (n = 24) discs. **P < 0.05. AF, annulus fibrosus; NP, nucleus pulposus.
Figure 6 Effect of IL-1 treatment on the IL-1 family gene expression in human intervertebral disc cells. Relative gene expression was normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) housekeeping gene and untreated controls (hence control is graphed at 1 on the log scale) for IL-1α (a), IL-1β (b), IL-1 receptor antagonist (IL-1Ra) (c), and IL-1 receptor, type I (IL-RI) (d) following IL-1β treatment of disc cells from two regions of non-degenerate (non-deg) (n = 6) and degenerate (n = 24) discs. **P < 0.05.
Table 1 Patient details and grades of disc degeneration in tissues used for immunohistochemical analysis
Laboratory number Source of tissue Sex Age (y) Clinical diagnosis Disc level Histological grade
1 Post-mortem Male 53 No data L4/5 1
2 Post-mortem Male 53 No data L5/S1 1
3 Surgical Male 44 Relatively normal L4/5 1
4 Surgical Male 47 Relatively normal L4/5 2
5 Post-mortem Male 75 No data L5/S1 3
6 Surgical Male ? Disc degeneration L5/S1 3
7 Surgical Male 48 Disc degeneration L4/5 3
8 Surgical Male 64 Disc degeneration L5/S1 3
9 Surgical Male 46 Disc degeneration L5/S1 4
10 Surgical Male 21 Disc degeneration L5/S1 4
11 Surgical Female 36 Disc degeneration L5/S1 4
12 Surgical Male 39 Disc degeneration L4/5 5
13 Surgical Female 32 Disc degeneration L5/S1 5
14 Surgical Female 36 Disc degeneration L4/5 5
15 Surgical Male 25 Disc degeneration L4/5 5
16 Surgical Female 35 Disc degeneration L4/5 6
17 Surgical Male 40 Disc degeneration L4/5 6
18 Post-mortem Female 73 No data L5/S1 6
19 Surgical Male 25 Disc degeneration L5/S1 6
20 Surgical Female 55 Disc degeneration L5/S1 7
21 Post-mortem Female ? No data L4/5 7
22 Surgical Female 58 Disc degeneration L4/5 7
23 Surgical Male 34 Disc degeneration L4/5 8
24 Surgical Female 24 Disc degeneration L5/S1 8
25 Surgical Female 33 Disc degeneration L5/S1 9
26 Post-mortem Female 73 No data L4/5 9
27 Surgical Male 68 Disc degeneration L5/S1 10
28 Post-mortem ? 47 No data L5/S1 10
29 Post-mortem ? 47 No data L5/S1 11
30 Surgical Male 39 Disc degeneration L4/5 12
?, not known.
Table 2 Real-time PCR probes and details of primers
Target Forward primer Probe Reverse primer Threshold
GAPDH PDAR PDAR PDAR 0.047
Collagen type I 5' CAG CCG CTT CAC CTA CAG C 3' 5' CCG GTG TGA CTC GTG CAG CCA TC 3' 5' TTT TGT ATT CAA TCA CTG TCT TGC C 3' 0.078
Collagen type II 5' GGC AAT AGC AGG TTC ACG TAC A 3' 5' CCG GTA TGT TTC GTG CAG CCA TCC T 3' 5' CGA TAA CAG TCT TGC CCC ACT T 3' 0.100
Aggrecan 5' TCG AGG ACA GCG AGG CC 3' 5' ATG GAA CAC GAT GCC TTT CAC CAC GA 3' 5' TCG AGG GTG TAG CGT GTA GAG A 3' 0.050
SOX9 5' GAC TTC CGC GAC GTG GAC 3' 5' CGA CGT CAT CTC CAA CAT CGA GAC 3' 5' GTT GGG CGG CAG GTA CTG 3' 0.0562
SOX6 5' CCG TGA GAT AAT GAC CAG TGT TAC TT 3' 5' AAC CCC AGA GCG CCG CAA A 3' 5' GTC CAC CAC ATC GGC AAG AC 3' 0.052
IL-1α PDAR PDAR PDAR 0.107
IL-1β PDAR PDAR PDAR 0.122
IL-1Ra 5' CCT GCA GGG CCA AGC A 3' 5' AGC CTC GCT CTT GGC AGG TAC TCA GT 3' 5' GCA CCC AAC ATA TAC AGC ATT CA 3' 0.122
IL-1RI 5' ATT TCT GGC TTC TAG TCT GGT GTT C 3' 5' ACT TGA TTT CAG GTC AAT AAC GGT CCC C 3' 5' AAC GTG CCA GTG TGG AGT GA 3' 0.163
MMP-3 5' TGA AGA GTC TTC CAA TCC TAC TGT TG 3' 5' CGT GGC AGT TTG CTC AGC CTA TCC AT 3' 5' CTA GAT ATT TCT GAA CAA GGT TCA TGC A 3' 0.108
MMP-9 5' CCC GGA GTG AGT TGA ACC A 3' 5' CCA AGT GGG CTA CGT GAC CTA TGA CAT CC 3' 5' CAG GAC GGG AGC CCT AGT C 3' 0.041
MMP-13 5' GGA CAA GTA GTT CCA AAG GCT ACA A 3' 5' CTC CAA GGA CCC TGG AGC ACT CAT GTT 3' 5' CTT TTG CCG GTG TAG GTG TAG ATA G 3' 0.108
ADAMTS-4 5' ACT GGT GGT GGC AGA TGA CA 3' 5' ATG GCC GCA TTC CAC GGT G 3' 5' TCA CTG TTA GCA GGT AGC GCT TT 3' 0.052
ADAMTS-5 5' GGA CCT ACC ACG AAA GCA GAT C 3' 5' CCC AGG ACA GAC CTA CGA TGC CAC C 3' 5' GCC GGG ACA CAC GGA GTA 3' 0.122
ADAMTS, a disintegrin and metalloproteinase with thrombospondin motifs; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; IL-1Ra, IL-1 receptor antagonist; IL1-RI, receptor, type I; MMP, matrix metalloproteinase; PDAR pre-designed assay reagent.
Table 3 Effects of IL-1β on gene expression in cells from non- degenerate or degenerate intervertebral discs
Target General trend Tissues affected Significant changesa (P < 0.05)
Origin of cells Disease state
MMP-3 Increase (5- to 10-fold) NP, AF N, D Non-degenerate NP and AF
MMP-13 Increase (5- to 10-fold) NP N, D Non-degenerate NP
ADAMTS-4 Increase (8-fold) NP D Degenerate NP
ADAMTS-5 No real trend - - None
SOX6 Decrease (3- to 9-fold) NP, AF N Non-degenerate NP
SOX9 Decrease (3-fold) NP, AF N None
Collagen I Decrease (5- to 10-fold) NP, AF N, D Degenerate NP
Collagen II Decrease (5- to 50-fold) AF N, D None
Aggrecan Decrease (3- to 7-fold) NP, AF N, D Non-degenerate NP and AF
IL-1α Increase (100-fold) NP, AF D Degenerate NP
IL-1β Increase (100-fold) NP, AF D Degenerate NP
IL-1Ra No real trend - - None
IL-1RI Decrease (2- to 10-fold) NP, AF N None
aSite of any significant change in gene expression. -, no effect seen; ADAMTS, a disintegrin and metalloproteinase with thrombospondin motifs; AF, annulus fibrosus; D, degenerate intervertebral disc; IL-1Ra, IL-1 receptor antagonist; IL1-RI, receptor, type I; MMP, matrix metalloproteinase; N, non-degenerate intervertebral disc; NP, nucleus pulposus.
==== Refs
Luoma K Riihimaki H Luukkonen R Raininko R Viikari-Juntura E Lamminen A Low back pain in relation to lumbar disc degeneration Spine 2000 25 487 492 10707396 10.1097/00007632-200002150-00016
Freemont AJ Watkins A Le Maitre C Jeziorska M Hoyland JA Current understanding of cellular and molecular events in intervertebral disc degeneration: implications for therapy J Pathol 2002 196 374 379 11920731 10.1002/path.1050
Kang JD Georgescu HI McIntyre-Larkin L Stefanovic-Racic M Donaldson WF IIIEvans CH Herniated lumbar intervertebral discs spontaneously produce matrix metalloproteinases, nitric oxide, interleukin-6, and prostaglandin E2 Spine 1996 21 271 277 8742201 10.1097/00007632-199602010-00003
Roberts S Caterson B Menage J Evans EH Jaffray DC Eisenstein SM Matrix metalloproteinases and aggrecanase: their role in disorders of the human intervertebral disc Spine 2000 25 3005 3013 11145811 10.1097/00007632-200012010-00007
Sztrolovics R Alini M Roughley PJ Mort JS Aggrecan degradation in human intervertebral disc and articular cartilage Biochem J 1997 326 235 241 9337874
Le Maitre CL Freemont AJ Hoyland JA Localization of degradative enzymes and their inhibitors in the degenerate human intervertebral disc J Pathol 2004 204 47 54 15307137 10.1002/path.1608
Boos N Nerlich AG Wiest I von der Mark K Aebi M Immunolocalization of type X collagen in human lumbar intervertebral discs during ageing and degeneration Histochem Cell Biol 1997 108 471 480 9450629 10.1007/s004180050187
Sive JI Baird P Jeziorsk M Watkins A Hoyland JA Freemont AJ Expression of chondrocyte markers by cells of normal and degenerate intervertebral discs Mol Pathol 2002 55 91 97 11950957 10.1136/mp.55.2.91
Cs-Szabo G Ragasa-San Juan D Turumella V Masuda K Thonar EJ An HS Changes in mRNA and protein levels of proteoglycans of the anulus fibrosus and nucleus pulposus during intervertebral disc degeneration Spine 2002 27 2212 2219 12394896 10.1097/00007632-200210150-00006
Inkinen RI Lammi MJ Lehmonen S Puustjarvi K Kaapa E Tammi MI Relative increase of biglycan and decorin and altered chondroitin sulfate epitopes in the degenerating human intervertebral disc J Rheumatol 1998 25 506 514 9517772
Martel-Pelletier J McCollum R Fujimoto N Obata K Cloutier JM Pelletier JP Excess of metalloproteases over tissue inhibitor of metalloprotease may contribute to cartilage degradation in osteoarthritis and rheumatoid arthritis Lab Invest 1994 70 807 815 8015285
Lohmander LS Neame PJ Sandy JD The structure of aggrecan fragments in human synovial fluid. Evidence that aggrecanase mediates cartilage degradation in inflammatory joint disease, joint injury, and osteoarthritis Arthritis Rheum 1993 36 1214 1222 8216415
Fujisawa T Hattori T Takahashi K Kuboki T Yamashita A Takigawa M Cyclic mechanical stress induces extracellular matrix degradation in cultured chondrocytes via gene expression of matrix metalloproteinases and interleukin-1 J Biochem (Tokyo) 1999 125 966 975 10220591
Lemare F Steimberg N Le Griel C Demignot S Adolphe M Dedifferentiated chondrocytes cultured in alginate beads: restoration of the differentiated phenotype and of the metabolic responses to interleukin-1beta J Cell Physiol 1998 176 303 313 9648917 10.1002/(SICI)1097-4652(199808)176:2<303::AID-JCP8>3.0.CO;2-S
Mengshol JA Vincenti MP Brinckerhoff CE IL-1 induces collagenase-3 (MMP-13) promoter activity in stably transfected chondrocytic cells: requirement for Runx-2 and activation by p38 MAPK and JNK pathways Nucleic Acids Res 2001 29 4361 4372 11691923 10.1093/nar/29.21.4361
Koshy PJ Lundy CJ Rowan AD Porter S Edwards DR Hogan A Clark IM Cawston TE The modulation of matrix metalloproteinase and ADAM gene expression in human chondrocytes by interleukin-1 and oncostatin M: a time-course study using real-time quantitative reverse transcription-polymerase chain reaction Arthritis Rheum 2002 46 961 967 11953973 10.1002/art.10212
Sandy JD Thompson V Doege K Verscharen C The intermediates of aggrecanase-dependent cleavage of aggrecan in rat chondrosarcoma cells treated with interleukin-1 Biochem J 2000 351 161 166 10998358 10.1042/0264-6021:3510161
Igarashi T Kikuchi S Shubayev V Myers RR 2000 Volvo Award winner in basic science studies: Exogenous tumor necrosis factor-alpha mimics nucleus pulposus-induced neuropathology. Molecular, histologic, and behavioral comparisons in rats Spine 2000 25 2975 2980 11145807 10.1097/00007632-200012010-00003
Olmarker K Larsson K Tumor necrosis factor alpha and nucleus-pulposus-induced nerve root injury Spine 1998 23 2538 2544 9854752 10.1097/00007632-199812010-00008
Olmarker K Rydevik B Selective inhibition of tumor necrosis factor-alpha prevents nucleus pulposus-induced thrombus formation, intraneural edema, and reduction of nerve conduction velocity: possible implications for future pharmacologic treatment strategies of sciatica Spine 2001 26 863 869 11317106 10.1097/00007632-200104150-00007
Cooper RG Freemont AJ TNF-alpha blockade for herniated intervertebral disc-induced sciatica: a way forward at last? Rheumatology (Oxford) 2004 43 119 121 13130150
Elfervig MK Minchew JT Francke E Tsuzaki M Banes AJ IL-1beta sensitizes intervertebral disc annulus cells to fluid-induced shear stress J Cell Biochem 2001 82 290 298 11527154 10.1002/jcb.1153
Koch H Reinecke JA Meijer H Wehling P Spontaneous secretion of interleukin 1 receptor antagonist (IL-1ra) by cells isolated from herniated lumbar discal tissue after discectomy Cytokine 1998 10 703 705 9770331 10.1006/cyto.1998.0353
Maeda S Kokubun S Changes with age in proteoglycan synthesis in cells cultured in vitro from the inner and outer rabbit annulus fibrosus. Responses to interleukin-1 and interleukin-1 receptor antagonist protein Spine 2000 25 166 169 10685479 10.1097/00007632-200001150-00005
Shen B Melrose J Ghosh P Taylor F Induction of matrix metalloproteinase-2 and -3 activity in ovine nucleus pulposus cells grown in three-dimensional agarose gel culture by interleukin-1beta: a potential pathway of disc degeneration Eur Spine J 2003 12 66 75 12592549
Shinmei M Kikuchi T Yamagishi M Shimomura Y The role of interleukin-1 on proteoglycan metabolism of rabbit annulus fibrosus cells cultured in vitro Spine 1988 13 1284 1290 3264623
Rannou F Corvol MT Hudry C Anract P Dumontier MF Tsagris L Revel M Poiraudeau S Sensitivity of anulus fibrosus cells to interleukin 1 beta. Comparison with articular chondrocytes Spine 2000 25 17 23 10647155 10.1097/00007632-200001010-00005
Takegami K Thonar EJ An HS Kamada H Masuda K Osteogenic protein-1 enhances matrix replenishment by intervertebral disc cells previously exposed to interleukin-1 Spine 2002 27 1318 1325 12065981 10.1097/00007632-200206150-00014
Doita M Kanatani T Harada T Mizuno K Immunohistologic study of the ruptured intervertebral disc of the lumbar spine Spine 1996 21 235 241 8720410 10.1097/00007632-199601150-00015
Takahashi H Suguro T Okazima Y Motegi M Okada Y Kakiuchi T Inflammatory cytokines in the herniated disc of the lumbar spine Spine 1996 21 218 224 8720407 10.1097/00007632-199601150-00011
Solovieva S Kouhia S Leino-Arjas P Ala-Kokko L Luoma K Raininko R Saarela J Riihimaki H Interleukin 1 polymorphisms and intervertebral disc degeneration Epidemiology 2004 15 626 633 15308963 10.1097/01.ede.0000135179.04563.35
Solovieva S Leino-Arjas P Saarela J Luoma K Raininko R Riihimaki H Possible association of interleukin 1 gene locus polymorphisms with low back pain Pain 2004 109 8 19 15082121 10.1016/j.pain.2003.10.020
Konttinen YT Kaapa E Hukkanen M Gu XH Takagi M Santavirta S Alaranta H Li TF Suda A Cathepsin G in degenerating and healthy discal tissue Clin Exp Rheumatol 1999 17 197 204 10342046
Gruber HE Hanley EN Jr Analysis of aging and degeneration of the human intervertebral disc. Comparison of surgical specimens with normal controls [see comments] Spine 1998 23 751 757 9563104 10.1097/00007632-199804010-00001
Walsh L Freemont AJ Hoyland JA The effect of tissue decalcification on mRNA retention within bone for in-situ hybridization studies Int J Exp Pathol 1993 74 237 241 8392858
Streiner DL Maintaining standards: differences between the standard deviation and standard error, and when to use each Can J Psychiatry 1996 41 498 502 8899234
Wang JY Baer AE Kraus VB Setton LA Intervertebral disc cells exhibit differences in gene expression in alginate and monolayer culture Spine 2001 26 1747 1752 11493844 10.1097/00007632-200108150-00003
Le Maitre CL Hoyland JA Freemont AJ Studies of human intervertebral disc cell function in a constrained in vitro tissue culture system Spine 2004 29 1187 1196 15167656 10.1097/00007632-200406010-00006
Livak KJ Schmittgen TD Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method Methods 2001 25 402 408 11846609 10.1006/meth.2001.1262
Ahn SH Cho YW Ahn MW Jang SH Sohn YK Kim HS mRNA expression of cytokines and chemokines in herniated lumbar intervertebral discs Spine 2002 27 911 917 11979160 10.1097/00007632-200205010-00005
Wilson KP Black JA Thomson JA Kim EE Griffith JP Navia MA Murcko MA Chambers SP Aldape RA Raybuck SA Structure and mechanism of interleukin-1 beta converting enzyme Nature 1994 370 270 275 8035875 10.1038/370270a0
Dinarello CA Thompson A From Interleukin-1 The cytokine handbook 1998 3 London Acaedemic press 35 72
Chomarat P Vannier E Dechanet J Rissoan MC Bachereau J Dinarello CA Miossec P The balance of IL-1 receptor antagonist/IL-1beta in rheumatoid synovium and its regulation by IL-4 and IL-10 J Immunol 1995 154 1432 1439 7822808
Baer AE Wang JY Kraus VB Setton LA Collagen gene expression and mechanical properties of intervertebral disc cell-alginate cultures J Orthop Res 2001 19 2 10 11332616 10.1016/S0736-0266(00)00003-6
Chiba K Andersson GB Masuda K Thonar EJ Metabolism of the extracellular matrix formed by intervertebral disc cells cultured in alginate Spine 1997 22 2885 2893 9431624 10.1097/00007632-199712150-00011
Gruber HE Fisher EC JrDesai B Stasky AA Hoelscher G Hanley EN Jr Human intervertebral disc cells from the annulus: three-dimensional culture in agarose or alginate and responsiveness to TGF-beta1 Exp Cell Res 1997 235 13 21 9281347 10.1006/excr.1997.3647
Flannery CR Little CB Caterson B Hughes CE Effects of culture conditions and exposure to catabolic stimulators (IL- 1 and retinoic acid) on the expression of matrix metalloproteinases (MMPs) and disintegrin metalloproteinases (ADAMs) by articular cartilage chondrocytes Matrix Biol 1999 18 225 237 10429942 10.1016/S0945-053X(99)00024-4
Richardson DW Dodge GR Effects of interleukin-1beta and tumor necrosis factor-alpha on expression of matrix-related genes by cultured equine articular chondrocytes Am J Vet Res 2000 61 624 630 10850836
Murakami S Lefebvre V de Crombrugghe B Potent inhibition of the master chondrogenic factor SOX9 gene by interleukin-1 and tumor necrosis factor-alpha J Biol Chem 2000 275 3687 3692 10652367 10.1074/jbc.275.5.3687
Frazer A Bunning RA Thavarajah M Seid JM Russell RG Studies on type II collagen and aggrecan production in human articular chondrocytes in vitro and effects of transforming growth factor-beta and interleukin-1beta Osteoarthritis Cartilage 1994 2 235 245 11550708
Chandrasekhar S Phadke K Interleukin-1-induced alterations in proteoglycan metabolism and matrix assembly Arch Biochem Biophys 1988 265 294 301 3262329 10.1016/0003-9861(88)90131-2
Lefebvre V Li P de Crombrugghe B A new long form of SOX5 (L-SOX5), SOX6 and SOX9 are coexpressed in chondrogenesis and cooperatively activate the type II collagen gene EMBO J 1998 17 5718 5733 9755172 10.1093/emboj/17.19.5718
Antoniou J Steffen T Nelson F Winterbottom N Hollander AP Poole RA Aebi M Alini M The human lumbar intervertebral disc: evidence for changes in the biosynthesis and denaturation of the extracellular matrix with growth, maturation, ageing, and degeneration J Clin Invest 1996 98 996 1003 8770872
Lambert CA Lapiere CM Nusgens BV An interleukin-1 loop is induced in human skin fibroblasts upon stress relaxation in a three-dimensional collagen gel but is not involved in the up-regulation of matrix metalloproteinase 1 J Biol Chem 1998 273 23143 23149 9722543 10.1074/jbc.273.36.23143
| 15987475 | PMC1175026 | CC BY | 2021-01-04 16:02:37 | no | Arthritis Res Ther. 2005 Apr 1; 7(4):R732-R745 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1732 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17331598748010.1186/ar1733Research ArticleEarly rheumatoid arthritis is characterized by a distinct and transient synovial fluid cytokine profile of T cell and stromal cell origin Raza Karim [email protected] Francesco [email protected] S John [email protected] Emma J [email protected] Chi-Yeung [email protected] Arne N [email protected] Janet M [email protected] Caroline [email protected] Christopher D [email protected] Mike [email protected] MRC Centre for Immune Regulation, Division of Immunity and Infection, The University of Birmingham, Birmingham, UK2 Department of Rheumatology, City Hospital, Sandwell and West Birmingham Hospitals NHS Trust, Birmingham, UK3 School of Biosciences, The University of Birmingham, Birmingham, UK4 Department of Radiology, City Hospital, Sandwell and West Birmingham Hospitals NHS Trust, Birmingham, UK5 Department of Immunology and Molecular Pathology, Royal Free and University College Medical School, London, UK2005 7 4 2005 7 4 R784 R795 30 1 2005 10 2 2005 2 3 2005 7 3 2005 Copyright © 2005 Raza et al.; licensee BioMed Central LtdThis is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Pathological processes involved in the initiation of rheumatoid synovitis remain unclear. We undertook the present study to identify immune and stromal processes that are present soon after the clinical onset of rheumatoid arthritis (RA) by assessing a panel of T cell, macrophage, and stromal cell related cytokines and chemokines in the synovial fluid of patients with early synovitis. Synovial fluid was aspirated from inflamed joints of patients with inflammatory arthritis of duration 3 months or less, whose outcomes were subsequently determined by follow up. For comparison, synovial fluid was aspirated from patients with acute crystal arthritis, established RA and osteoarthritis. Rheumatoid factor activity was blocked in the synovial fluid samples, and a panel of 23 cytokines and chemokines measured using a multiplex based system. Patients with early inflammatory arthritis who subsequently developed RA had a distinct but transient synovial fluid cytokine profile. The levels of a range of T cell, macrophage and stromal cell related cytokines (e.g. IL-2, IL-4, IL-13, IL-17, IL-15, basic fibroblast growth factor and epidermal growth factor) were significantly elevated in these patients within 3 months after symptom onset, as compared with early arthritis patients who did not develop RA. In addition, this profile was no longer present in established RA. In contrast, patients with non-rheumatoid persistent synovitis exhibited elevated levels of interferon-γ at initiation. Early synovitis destined to develop into RA is thus characterized by a distinct and transient synovial fluid cytokine profile. The cytokines present in the early rheumatoid lesion suggest that this response is likely to influence the microenvironment required for persistent RA.
==== Body
Introduction
The synovium is the primary site of pathology in rheumatoid arthritis (RA). The rheumatoid synovium contains large numbers of CD4+ T cells. Patients with severe disease frequently express DR4 molecules that share an epitope in the third hypervariable region of the β-chain [1], suggesting a pathogenic role for T cells. However, the presence of only low levels of T cell related cytokines in the synovium and synovial fluid of established RA patients [2,3] led many to question the role of T cells in persistent disease. Nevertheless, this synovial cytokine profile is consistent with the highly differentiated CD45RObrightRBdull phenotype of synovial T cells [4]. A widely accepted model has emerged in which the persistence of inflammation in established RA is driven by interactions between T cells, macrophages and fibroblasts in an abnormal microenvironment [5,6]. The synovial T cell population is maintained through active inhibition of apoptosis, mediated at least in part by fibroblast and macrophage derived type 1 IFNs, and active retention facilitated by fibroblast derived transforming growth factor-β [7-9]. Contact dependent interactions between T cells and macrophages stimulate the production of proinflammatory cytokines, including tumour necrosis factor (TNF)-α, in an antigen independent manner [10-12].
This model of persistence in established disease requires the presence of hyperplastic synovial tissue, which is unlikely to be present at the onset of RA. Consequently, the processes manifest at initiation that lead to persistence are likely to be distinct. Difficulties in accessing patients with very early disease and in sampling those joints involved at clinical onset have proved to be obstacles to addressing these issues. The role of T cells and antigen in the initiation of RA, the mechanisms that drive early fibroblast expansion, and the interplay between T cells and the stromal environment therefore remain obscure.
In order to study mechanisms of very early synovitis potentially leading to RA, we established a rapid access clinic with a wide recruitment base in which patients with synovitis were seen within the first few weeks after symptom onset. Using a multiplex bead based system, allowing simultaneous analysis of over 20 soluble molecules in very small sample volumes [13], we measured a panel of cytokines and chemokines in the synovial fluid of patients with early inflammatory arthritis. Patients whose disease subsequently fulfilled American Rheumatism Association (ARA) criteria for RA had a cytokine profile characterized by a range of T cell, stromal cell and macrophage related cytokines that was not present in long-standing RA. This profile was not seen in patients with other early arthritides. A model was built incorporating these cytokines that distinguished patients who progressed to RA from other early arthritis patients with a high degree of accuracy. These data suggest that the pathological mechanisms operating at the onset of clinically apparent RA are distinct from those in other early inflammatory arthritides, and that these mechanisms are transient. In addition, the present study supports the concept that T cells play a role in disease initiation that is different from their role in maintaining persistent inflammation.
Materials and methods
Patients
Patients were recruited through the rapid access clinic for early inflammatory arthritis at City Hospital, Birmingham, UK. Permission was obtained from the local ethics committee and all patients gave written informed consent. All patients had one or more swollen joints and symptoms (inflammatory joint pain and/or early morning stiffness and/or joint related soft tissue swelling) of duration 3 months or less. Patients with evidence of previous inflammatory joint disease were excluded. Joints were aspirated under either palpation or ultrasound guidance. Where the synovial fluid volume in the inflamed joint was very small (commonly at proximal interphalyngeal, metacarpophalyngeal and wrist joints) direct aspiration was not possible. Although the cellular content of these joints could be sampled using ultrasound guided lavage [14], lavage samples were excluded from this study because the potentially variable dilution made comparisons of the absolute levels of chemokines and cytokines unreliable. Synovial fluid was directly aspirated from the joints of 36 patients with non-crystal-related very early inflammatory arthritis. In addition, for comparison, synovial fluid was obtained from patients in three well defined diagnostic groups: early inflammatory arthritis of crystal origin that resolved (gout [n = 12] and pseudogout [n = 2]), established RA (n = 9) and osteoarthritis (n = 4). Synovial fluid was collected into nonheparinized tubes and spun at 1000 g for 10 min within 30 min of sample collection. The acellular portion of synovial fluid was stored at -70°C before subsequent analysis.
The 36 patients with early inflammatory arthritis were followed for 18 months and then assigned to their final diagnostic groups. Patients were classified as having RA according to the 1987 ARA criteria [15], allowing criteria to be satisfied cumulatively. Although the 1987 ARA criteria have no exclusions, we excluded from the RA category patients with alternative rheumatological diagnoses explaining their inflammatory arthritis. Thus, one patient, with polymyositis related arthritis, who fulfilled criteria for RA was excluded from the RA group and included in the non-rheumatoid persistent group. In addition, one patient, who fulfilled criteria for RA at presentation (but was seronegative for rheumatoid factor [RF] and anti-cyclic citrullinated peptide [CCP] antibody), had transient disease, remained symptom free and off all mediation at follow up, and was included in the resolving group. Patients were diagnosed with reactive arthritis (ReA), psoriatic arthritis (PsA) and a number of miscellaneous conditions according to established criteria. Of the 36 patients with non-crystal-related early inflammatory arthritis, 14 had a resolving disease (sexually acquired ReA [n = 4] and unclassified arthritis [n = 10]) and 22 developed persistent inflammatory arthritis (RA [n = 8], unclassified arthritis [n = 9], PsA [n = 2] and arthritis related to ulcerative colitis [n = 1], polymyositis [n = 1] and Behçet's disease [n = 1]). Three of these RA patients presented with inflammatory arthritis of just the knee or ankle joint(s) though polyarticular synovitis, including involvement of the small joints of the hands, developed subsequently.
Anti-CCP antibody and rheumatoid factor assay
IgG anti-CCP antibody was detected using the DIASTAT™ anti-CCP assay (Axis-Shield Diagnostics Ltd., Dundee, UK) and seropositivity was defined as a titre of ≥5 IU/ml. RF was measured using an ELISA (AUTOSTAT™II Total RF; Hycor Biomedical Ltd., Penicuik, UK).
Cytokine and chemokine assay
Twenty-three cytokines and chemokines were measured simultaneously in synovial fluid samples using a multiplex detection kit (Biosource International, Camarillo, CA, USA): IL-1β, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12, IL-13, IL-15, IL-17, macrophage inflammatory protein (MIP)-1α, MIP-1β, monocyte chemoattractant protein (MCP)-1, RANTES (regulated on activation, normal T expressed and secreted), eotaxin, TNF-α, IFN-γ, granulocyte–macrophage colony-stimulating factor (GM-CSF), granulocyte colony-stimulating factor, epidermal growth factor (EGF), basic fibroblast growth factor (bFGF) and vascular endothelial growth factor. A volume of 50 μl synovial fluid or of the cytokine/chemokine standards was preincubated with 50 μl blocking buffer (40% normal mouse serum [Sigma, Poole, UK], 20% goat serum [DakoCytomation Ltd, Ely, UK] and 20% rabbit serum [DakoCytomation Ltd]; see below) for 30 min. A volume of 50 μl diluted sample, or blocking buffer alone, was incubated with 25 μl of multiplex microspheres for 2 hours. Microspheres were washed with phosphate-buffered saline/0.05% Tween 20 and incubated with 25 μl of biotinylated detection antibody, diluted in 25 μl blocking buffer and 50 μl assay buffer (1% bovine serum albumin [Sigma] in phosphate-buffered saline/0.05% Tween 20), for 1 hour. Microspheres were then washed, incubated with streptavidin–PE at room temperature for 30 min and washed again. Subsequently, the microspheres were resuspended in 100 μl assay buffer and analyzed (Luminex100 LabMAP™ system; Luminex Corporation, Austin, TX, USA). Cytokine and chemokine concentrations were calculated by reference to the standard curve.
False-positive results, caused by cross-linking of capture and detection antibodies by RF, are a common problem with any immunoassay of rheumatoid samples. We tested several strategies to eliminate this effect, including the absorption of RF, and found the optimal approach in this system to be blocking with a combination of 20% normal mouse serum, 10% rabbit serum and 10% goat serum. The validity of the results were assessed in two ways. First, a negative control microsphere was produced by conjugating a microsphere, with a different intrinsic fluorescence to any of those in the assay, to total mouse immunoglobulin (the capture antibodies on the microspheres were all mouse monoclonal antibodies). These microspheres were tested in parallel with those conjugated to specific antibodies. Second, an additional aliquot of each synovial fluid sample was tested with an entirely different detection kit obtained from an alternative source (Upstate Biotechnology, Milton Keynes, UK). This assay uses different antibody combinations to detect 14 of the cytokines and chemokines measured by the primary assay (IL-1β, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12, IL-13, MCP-1, RANTES, eotaxin, TNF-α and IFN-γ).
Analysis
In order to identify cytokines that distinguish patients with early RA from other patients with early synovitis and from patients with established RA, we performed univariate and multivariate analyses. The Wilcoxon rank sum test was used for univariate analysis. The false discovery rate correction was used to correct for the multiple comparisons made [16]. Results are reported as Q values, which represent the likelihood of obtaining a given P value by chance given the multiple tests performed. We used a 1% false discovery rate (Q < 0.01) as our threshold for statistical significance.
Univariate statistical tests can be used to identify cytokines that are differentially expressed between two or more groups of patients. However, such analyses do not address the relative significance of different combinations of cytokines required to discriminate between different conditions. These discriminatory profiles are likely to provide insights to the biological processes underlying the different disease states. To identify groups of cytokines that allow the distinction of potential outcomes in patients with early arthritis, we used a classification algorithm termed Random Forest [17,18]. This method is based on the principal of decision trees and incorporates efficient methods to establish the importance of each cytokine in the classification and to perform an unbiased estimate of classification error. Random Forest does not use cross-validation or a separate training set to obtain an unbiased estimate of classification error. Instead, a collection of decision trees is constructed using a different randomly chosen (bootstrap) sample of two-thirds of the data for each tree. Each branching point, or node, is produced by using the best of a subset of predictors randomly chosen at that node. One-third of the cases are left out of each bootstrap sample and not used in the construction of any given decision tree. Each of these cases left out in the construction of a particular tree is then put back into that tree to test the validity of the classification obtained from the bootstrap sample. In this way, a test set classification is obtained for each case, in about one-third of the trees. The classification error is defined as the proportion of patients who are misclassified. This strategy has previously been shown to perform well compared with other classifiers, including discriminant analysis, support vector machines and neural networks; it is also robust against overfitting [17].
The Random Forest algorithm also estimates the relative importance of each variable in contributing to the classification process. In this study we used Random Forest to rank the contribution of each cytokine to discrimination between outcomes of patients with early synovitis, as a possible index of their biological contribution to this discrimination. This analysis does not emphasize cytokines that are of general importance to inflammation.
The models developed using Random Forest can be visualized graphically by using multidimensional scaling to plot the relative similarity between patients in the trees [19]. This allows the magnitude of the difference between groups, and the utility of the data set in distinguishing different outcomes, to be assessed.
Results
Patient characteristics
Baseline characteristics of patients with early inflammatory arthritis of non-crystal origin are shown in Table 1. Patients who developed RA were significantly older than patients in other groups. There were no significant differences in symptom duration or C-reactive protein (CRP) level at initial presentation between the groups. Of the nine patients with established RA, four were female and the median age was 62 years (interquartile range 43–74 years). There were no significant sex or age differences between the established RA and early RA patients.
Cytokine and chemokine levels
The levels of synovial cytokines in patients with early and established synovitis are shown in Figs 1 and 2, and the statistical significance of differences between cytokine levels in patient groups is shown in Table 2. Patients with early synovitis destined to develop RA exhibited a cytokine profile in synovial fluid that was distinct from that of patients with other early inflammatory arthritides (early synovitis that developed into non-rheumatoid persistent disease plus non-crystal-related resolving arthritis plus crystal-related resolving arthritis). Early RA synovial fluid was characterized by significantly elevated levels of T cell related cytokines (IL-2, IL-4, IL-13 and IL-17) and stromal cell and macrophage related cytokines (EGF, bFGF, IL-1 and IL-15) when compared with synovial fluid from patients with other early synovitis (Figs 1 and 2, and Table 2). Although levels of the Th2-type cytokines IL-4 and IL-13 were elevated in patients with early RA, IFN-γ was never detected in these patients (Fig. 1). In contrast, IFN-γ was detected in five patients with early non-rheumatoid persistent disease (one PsA, one ulcerative colitis related arthritis and three unclassified) and in three patients with early self-limiting disease (two ReA and one unclassified; Fig. 1). In addition, the synovial cytokine profile of patients with early synovitis destined to develop RA was significantly different from that of patients with established RA. Patients with early RA had significantly elevated synovial levels of IL-2, IL-4, IL-13, IL-17, EGF and bFGF when compared with patients with established RA (Figs 1 and 2, and Table 2).
We assessed the relative importance of the cytokines measured in distinguishing patients with early disease destined to develop RA from patients with all other early arthritis together (early disease that progressed to non-rheumatoid persistent disease plus non-crystal-related resolving arthritis plus crystal-related resolving arthritis) and patients with established RA. The relative importance of each cytokine in the accuracy of classification in these models is shown in Fig. 3a,c. IL-13, together with IL-2, IL-4, IL-15, bFGF and EGF, were the most important cytokines in distinguishing early RA patients from other patients with early synovitis, with high levels of these cytokines predicting the development of RA. Similarly, IL-13, together with IL-2, IL-4, IL-17, bFGF and EGF, were the most important cytokines in distinguishing early RA patients from patients with established RA. Using this approach two of the eight early RA patients were misclassified in both models. One of these was an 82-year-old man who presented with a 6-week history of synovitis (RF negative, anti-CCP antibody negative, CRP 51 mg/l); the other was a 70-year-old woman who presented with a 10-week history of synovitis (RF positive, anti-CCP antibody positive, CRP 25 mg/l). The relationships between the different patients in the two models and the ability of the synovial cytokines to distinguish between different patient outcomes are shown in Fig. 3b,d.
The cytokine profile that was seen in patients with early RA was transient. It was not seen in established RA (Figs 1 and 2) or after the first few months of symptoms in patients with early disease that went on to persist (Fig. 4). The transient nature of the elevations in IL-2 and IL-4 in early RA synovial fluid (Fig. 4) was also apparent for IL-13, IL-15, EGF and bFGF.
The validity of the results obtained using the multiplex assay was confirmed in two ways. First, no significant irrelevant staining was observed with any of the synovial fluid samples using the negative control microspheres. The median fluorescence intensity of the negative control microspheres when incubated with the synovial fluid samples was 14 (standard deviation 2.6) and when incubated with assay buffer alone was 16.3 (standard deviation 0.5). Second, the correlations between the results obtained using the two different antibody detection systems were all highly statistically significant and were as follows (expressed as Spearman's rank correlation [rs]): IL-1β, rs = 0.65 (P < 0.0001); IL-2, rs = 0.77 (P < 0.0001); IL-4, rs = 0.63 (P < 0.0001); IL-5, rs = 0.5 (P < 0.0001); IL-6, rs = 0.95 (P < 0.0001); IL-8, rs = 0.8 (P < 0.0001); IL-10, rs = 0.56 (P < 0.0001); IL-12, rs = 0.27 (P = 0.006); MCP-1, rs = 0.72 (P < 0.0001); RANTES, rs = 0.84 (P < 0.0001); TNF-α, rs = 0.47 (P < 0.0001); and IFN-γ, rs = 0.54 (P < 0.0001). Levels of eotaxin and IL-13 were below the detection limit of the second assay.
There was no significant correlation between the RF titre and the levels of any of the cytokines or chemokines measured in patients with early RA. Correlations between CRP and levels of chemokines and cytokines in early inflammatory arthritis synovial fluid were statistically significant only for IL-6 (rs = 0.49 [P = 0.003]), TNF-α (rs = 0.37 [P = 0.03]), IFN-γ (rs = 0.52 [P = 0.001]) and vascular endothelial growth factor (rs = 0.47 [P = 0.004]). These correlations were independent of outcome. An independent analysis of the eight patients with early RA revealed no correlation between CRP and the level of any chemokine or cytokine.
Discussion
The synovium of patients with established RA is expanded and contains large numbers of fibroblasts, macrophages and highly differentiated T cells [20]. Although the mechanisms responsible for persistence of this infiltrate in established disease have been well characterized (for review [5,21,22]), those involved in disease initiation have not. A number of groups have studied RA at a relatively early stage from pathological, radiological and therapeutic perspectives. The maximum duration of symptoms accepted for recruitment to such studies has been highly variable, usually ranging from less than 1 year to less than 3 years [23-26]. However, the observation that spontaneous remission is unusual if inflammatory arthritis has persisted for longer than 6 months [27] suggests that pathological mechanisms driving the switch to persistence are already established by this time. In contrast, the concept that RA patients with a much shorter disease duration may be within a therapeutically distinct window is supported by a study in which remission was more commonly induced in patients with disease of ≤4 months duration compared with longer duration disease [28]. However, the confounding effect of differential rates of spontaneous remission has been difficult to exclude. Whether this very early phase of disease is pathologically distinct remains to be determined.
Within the first 12 weeks after symptom onset, we found that the synovial fluid of patients who eventually developed RA was characterized by a wide range of cytokines and chemokines. Some, such as IL-6, were present in all inflammatory arthritides, suggesting their importance in synovitis per se rather than a specific role in rheumatoid synovitis. In contrast, early RA patients had a distinct and consistent synovial cytokine profile characterized by T cell, macrophage and stromal cell related cytokines (in particular IL-2, IL-4, IL-13, IL-17, IL-1, IL-15, bFGF and EGF), which was not seen in other early arthritides. This profile was transient, and was no longer present in any patients with established RA. Seven of the eight patients whose disease developed into RA already expressed RF and anti-CCP antibody at this early stage, within weeks of symptom onset. Several groups have shown that these antibodies can be found in patients who subsequently develop RA, long before symptoms are apparent [29-31], implying a preclinical pathology. Our data are entirely compatible with the possibility of a preclinical phase of disease in patients with RA. We are clearly unable to address the issue of how the synovial cytokine profile within the first few months of symptoms compares with that which may be present during a preclinical phase of disease. However, our data do suggest that the pathological processes operating in the rheumatoid joint within the first few weeks after symptom onset differ from those processes operating in other early synovial lesions, and that these processes are transient.
In other studies comparing early RA (defined as a symptom duration of <1 year) and established RA, immunohistological analysis of the synovium, including an assessment of TNF-α, IL-1β and IL-6 expression, did not reveal any differences between early (mean disease duration 6 months) and long-standing disease [23]. Expression of IFN-γ, IL-10 and IL-12 mRNA in synovial fluid mononuclear cells were also similar between such early patients and those with established RA [32]. However, very few groups have studied the pathology of RA within the first few weeks of symptom onset. Indeed, because patients with a symptom duration of less than 6 weeks cannot fulfil 1987 ARA classification criteria for RA [15], any study of early disease that limits itself to patients fulfilling these criteria will exclude those with very early synovitis. We studied patients with inflammatory arthritis of duration 3 months or less, recruited through a rapid access clinic. The assignment of patients to specific diagnostic categories was done subsequently at clinical follow up. In one of the few other studies of the pathology of very early inflammatory arthritis, synovial histology was assessed in 24 patients with disease of duration under 2 months [33]. Nine patients had transient synovitis and six developed RA. Interestingly, there were no clear histological differences between these patients. In contrast, our results suggest that pathological mechanisms operating at the onset of clinically apparent RA are different from those in patients with other early arthritides. The results also show a surprising degree of uniformity between RA patients in this early phase of disease.
The concept that T cells play an important role in the initiation of RA, with antigen specific T cells mediating autoimmunity, is not new (for review [34]). However, robust evidence to support such a role for T cells has been lacking. The observation in this report of consistently elevated T cell derived cytokines strongly suggests T cell activity in early RA. The clear difference between the levels of specific T cell related cytokines in early RA and other early inflammatory arthritides, as well as established RA, suggests a pathological process that is specific to early RA and that is transient. The Th2 cytokine pattern, with a marked absence of IFN-γ and predominance of IL-4 and IL-13, in the earliest clinically apparent RA lesions was unexpected. Current paradigms suggest the cytokine balance in established RA to be skewed in favour of a Th1-type response [34]. In established RA, synovial T cells produce IFN-γ after in vitro stimulation [35], and mRNA for IFN-γ but not IL-4 is found in synovial fluid mononuclear cells [32]. However, recent evidence from mouse models suggests that IFN-γ plays a significant role in the resolution of synovial inflammation [36]. In murine models of Leishmania infection, a Th2 polarized response leads to persistent disease, whereas a Th1 polarized delayed-type hypersensitivity response leads to resolution [37]. This is similar to the effects seen in human leprosy [38]. The discrepancy between IFN-γ levels in early RA and other early arthritides may consequently be of relevance to the pathology of the transition to persistent inflammation. In addition, T cells cloned from early rheumatoid synovium have been shown to produce significantly more IL-4 than those from long-standing disease [39]. When T cells were re-cloned from an early patient at a subsequent time point, IL-4 production was significantly diminished [39]. These findings concur with the data reported here, suggesting a transient Th2 phenotype in early RA synovial fluid.
The role of the Th2-type cytokines in early RA is unclear. IL-4 has divergent proinflammatory and anti-inflammatory effects in animal models of inflammatory arthritis [40,41]. IL-13 induces proliferation and CD154 (CD40 ligand) expression in lung fibroblasts [42,43] and is important in inducing fibrosis in Th2 mediated diseases such as schistosomiasis [44]. In addition, both IL-4 and IL-13 protect synoviocytes against nitric oxide induced apoptosis [45]. The pro-survival and proliferative effects of these cytokines may be important in the development of the expanded fibroblast network, which occurs during early disease and which characterizes established RA [46]. The presence of significant levels of the autocrine synovial fibroblast growth factors bFGF and EGF [47] clearly supports this process. The absence of these growth factors in the synovial compartment of patients with self-limiting disease is not surprising, because one would not expect an expanded fibroblast layer (which would mediate the switch to persistence) in such patients. Interestingly, however, these growth factors were absent in non-RA persistent inflammatory arthritis, suggesting a difference between the mechanism of synovial hyperplasia in early RA and other persistent inflammatory arthritides.
Th2-type cytokines have additional effects on synovial fibroblasts that may be relevant in early RA. Cultured synovial fibroblasts have a global gene expression profile that is quite different from that of lymph node and tonsil fibroblasts [48]. However, the addition of IL-4 to synovial fibroblasts dramatically modulates their gene expression profile, which converges with that of fibroblasts from secondary lymphoid tissue [48]. Germinal centre-like structures are seen in the synovium of many RA patients [49]. The synovial environment in early RA may modulate fibroblast function, leading to the production of factors facilitating lymphoid aggregate formation and allowing local RF production.
The distinct T cell related cytokine profile observed in patients with early RA supports the concept that T cells play an important role at the onset of clinically apparent disease. Tissue dendritic cells (DCs) are specialized for high antigen capture and migration into draining lymph nodes, where they are essential for activating naïve T cells. The role played by DCs in the onset of inflammatory arthritis has been explored in a murine model, in which collagen pulsed mature DCs induced inflammatory arthritis when transferred into DBA/1 mice [50]. Characterization of the T cell response in draining lymph nodes revealed significant proliferation of collagen specific T cells and IL-2 production, suggesting that priming of autoreactive T cells by DCs may play a role in disease initiation. Blood monocytes may differentiate into DCs in the presence of GM-CSF, IL-4 and TNF-α [51] and early myeloid DC progenitors in RA synovial fluid differentiate in response to IL-4 or IL-13 in combination with GM-CSF and stem cell factor [52]. The synovial cytokine environment in early RA may thus stimulate mature DC production and consequent T cell activation.
The recruitment of leucocytes into the synovial compartment is regulated, in part, by chemokines. Although levels of the chemokines measured (e.g. MIP-1α, MIP-1β, MCP-1 and RANTES) were elevated in the synovial fluid of many patients with early RA, they were of limited value in distinguishing early RA from other early arthritides or from established RA. The chemokine stromal cell-derived factor 1 (SDF-1; CXC chemokine ligand 12 [CXCL12]) has been implicated in the recruitment and retention of T cells and monocytes into the rheumatoid synovium [9,53]. Unfortunately, no combination of monoclonal antibodies could be found that produced reliable estimates of SDF-1 using the multiplex detection strategy, so we were unable to measure this chemokine.
Several roles can be proposed for the other cytokines found in early RA. IL-17 has pleiotropic effects on leucocytes and stromal cells (for review [54]). For example, it induces IL-6 and IL-8 production by fibroblasts [55] and stimulates macrophage IL-1 and TNF-α production [56]. IL-15 stimulated T cells induce macrophage-mediated TNF-α production [11]. In addition, all common γ-chain cytokines, including IL-2, IL-4 and IL-15, are potent T cell survival factors [57], which may support the persistence of the early rheumatoid lesion.
Conclusion
The data presented herein suggest that the pathology of RA within the first few months after symptom onset is distinct from that of other early inflammatory arthritides and of established RA. The nature of the cytokines present in the synovial fluid of patients with early RA suggests that this response is likely to influence the microenvironment required for persistent disease.
Abbreviations
ARA = American Rheumatism Association; bFGF = basic fibroblast growth factor; CCP = cyclic citrullinated peptide; CRP = C-reactive protein; DC = dendritic cell; EGF = epidermal growth factor; GM-CSF = granulocyte–macrophage colony-stimulating factor; IFN = interferon; IL = interleukin; MCP = monocyte chemoattractant protein; MIP = macrophage inflammatory protein; PsA = psoriatic arthritis; RA = rheumatoid arthritis; RANTES = regulated on activation, normal T expressed and secreted; ReA = reactive arthritis; RF = rheumatoid factor; Th = T-helper (cell); TNF = tumour necrosis factor.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
KR participated in the design of the study, recruited and followed up the early arthritis patients, analyzed and interpreted the data, and drafted the manuscript. FF performed the statistical analysis and was involved in drafting the manuscript. SJC acquired the cytokine and chemokine data, analyzed and interpreted the data, and was involved in drafting the manuscript. ET acquired the cytokine and chemokine data. CYL participated in assessing patients and in performing ultrasound guided joint aspirations. ANA, JML, CG and CB participated in the design of the study and interpretation of data. MS participated in the design of the study and interpretation of data, and was involved in drafting the manuscript. All authors have read and approved the final manuscript.
Acknowledgements
This work was supported by the Arthritis Research Campaign (ARC). We are grateful to GD Kitas and M Breese for measuring anti-CCP antibodies at the Dudley Group of Hospitals NHS Trust, Dudley, UK; to K Kumar for help with metrology and patient assessment; to V Trevino for help with statistical analysis; and to DM Carruthers and RD Situnayake for help with the recruitment of patients.
Figures and Tables
Figure 1 Synovial fluid cytokines in early and established arthritis. Shown are synovial fluid concentrations (pg/ml) of IL-2, IL-4, IL-13, IL-15, epidermal growth factor (EGF), basic fibroblast growth factor (bFGF), eotaxin and IFN-γ. Patient groups: 1, early synovitis that develops into rheumatoid arthritis (RA); 2, early synovitis that develops into non-rheumatoid persistent synovitis; 3, early non-crystal-related resolving synovitis; 4, crystal-related resolving synovitis; 5, established RA; and 6, osteoarthritis.
Figure 2 Synovial fluid cytokines in early and established arthritis. Shown are synovial fluid concentrations (pg/ml) of IL-1β, macrophage inflammatory protein (MIP)-1β, granulocyte–macrophage colony-stimulating factor (GM-CSF), IL-12, MIP-1α, monocyte chemoattractant protein (MCP)-1, IL-17, IL-10, granulocyte colony-stimulating factor (G-CSF), vascular endothelial growth factor (VEGF), tumour necrosis factor (TNF)-α, RANTES (regulated on activation, normal T expressed and secreted), IL-8, IL-6 and IL-5. Patient groups: 1, early synovitis that develops into rheumatoid arthritis (RA); 2, early synovitis that develops into non-rheumatoid persistent synovitis; 3, early non-crystal-related resolving synovitis; 4, crystal-related resolving synovitis; 5, established RA; and 6, osteoarthritis.
Figure 3 Importance of individual cytokines in classifying patient groups. Plots represent models discriminating patients with early inflammatory arthritis who develop rheumatoid arthritis (RA) from (a,b) all other early inflammatory arthritis patients (early disease that progresses to non-rheumatoid persistent disease plus early non-crystal-related resolving arthritis plus crystal-related resolving arthritis) and from (c,d) established RA. Panels a and c show the relative importance of the cytokines in the overall classification. The vertical axes represent individual cytokines arranged according to importance. The horizontal axes represent the average decrease in classification accuracy seen when the values for each cytokine are permuted. Important cytokines are associated with a greater decrease in classification accuracy. The plots shown in panels b and d are metric multidimensional scaling (MDS) representations of the proximity matrices of the Random Forest models demonstrating the relationship between individual patients in the two models. The two axes represent the first and second MDS axes. Closed circles represent patients with early inflammatory arthritis who develop RA in both panels; open circles represent patients with (panel b) all other early inflammatory arthritis and (panel d) established RA.
Figure 4 Longitudinal synovial fluid cytokine concentrations (pg/ml) in patients with early inflammatory arthritis. Results are shown for patients with early inflammatory arthritis who develop rheumatoid arthritis (closed circles) and non-rheumatoid persistent inflammatory arthritis (open circles) for (a) IL-2, (b) IL-4 and (c) IFN-γ.
Table 1 Baseline characteristics of patients with early inflammatory arthritis of non-crystal origin
RA Non-RA persistent Resolving P
Number 8 14 14
Female (n) 4 6 6 NSa
Age (years)b,c 65.5 (55–73) 29.5 (23.5–61) 28.5 (21–45) 0.006d
RA versus non-RA persistent <0.05e
RA versus resolving <0.01e
Symptom duration (weeks)b,c 9 (6–9.5) 7.5 (2.5–11.5) 4 (2–8) NSd
CRP (mg/l)b,c 31 (27.5–52.5) 67 (29–172.5) 24 (3.5–45) NSd
RF ≥30 IU/ml (n) 7 0 2 <0.0001a
Anti-CCP antibodies ≥5 IU/ml (n) 7 0 1 <0.0001a
RF ≥30 IU/ml + Anti-CCP antibodies ≥5 IU/ml (n) 7 0 0 <0.0001a
Initial NSAID use (n) 7 6 8 NSa
Initial DMARD use (n) 0 0 0 NSa
Initial prednisolone use (n) 0 3 0 NSa
Joint aspirated (knee/ankle [n]) 5/3 13/1 11/3 NSa
aχ2 test. bMedian. cInterquartile range. dKruskall–Wallis test. eWhere the medians were significantly different at the 5% level, Dunn's post-test was used to compare individual groups. CCP, cyclic citrullinated peptide; CRP, C-reactive protein; DMARD, disease-modifying antirheumatic drug; NS, not significant; NSAID, nonsteroidal antirheumatic drug; RA, rheumatoid arthritis; RF, rheumatoid factor.
Table 2 Comparisons between synovial fluid cytokineconcentrations in patients with early synovitis that develops into rheumatoid arthritis and other patient groups
Cytokine Early RA versus all other early synovitis Early RA versus established RA
P Q P Q
IL-13 <0.00001 0.00002 0.00008 0.0004
EGF <0.00001 0.00002 0.002 0.007
bFGF <0.00001 0.00002 0.002 0.007
IL-4 <0.00001 0.00002 0.004 0.0097
IL-2 <0.00001 0.00002 0.004 0.0097
Eotaxin 0.00003 0.0002 NS NS
IL-1 0.00003 0.0002 NS NS
IL-17 0.00005 0.0003 0.002 0.007
IL-15 0.0001 0.0005 NS NS
MIP-1β 0.0005 0.002 NS NS
MIP-1α 0.0006 0.002 NS NS
IL-12 0.002 0.006 NS NS
GM-CSF 0.002 0.006 NS NS
G-CSF 0.004 0.0097 NS NS
Comparator groups were, first, patients with all other early synovitis (synovitis that develops into non-rheumatoid persistent synovitis plus non-crystal-related resolving synovitis plus crystal-related resolving synovitis) and, second, patients with established rheumatoid arthritis (RA). Groups were compared using the Wilcoxon rank sum test. The significance of the difference between groups is shown (P). The false discovery rate correction was used to correct for the multiple comparisons made. Results are reported as Q values, which represent the likelihood of obtaining a given P value by chance, given the multiple tests performed. For cytokines not shown there were no significant differences between groups (a 1% false discovery rate [Q < 0.01] was used as the threshold for statistical significance). bFGF, basic fibroblast growth factor; EGF, epidermal growth factor; G-CSF, granulocyte colony-stimulating factor; GM-CSF, granulocyte–macrophage colony-stimulating factor; IL, interleukin; MIP, macrophage inflammatory protein; NS, not significant.
==== Refs
Gregersen PK Silver J Winchester RJ The shared epitope hypothesis. An approach to understanding the molecular genetics of susceptibility to rheumatoid arthritis Arthritis Rheum 1987 30 1205 1213 2446635
Husby G Williams RC Jr Immunohistochemical studies of interleukin-2 and gamma-interferon in rheumatoid arthritis Arthritis Rheum 1985 28 174 181 3918543
Firestein GS Zvaifler NJ Peripheral blood and synovial fluid monocyte activation in inflammatory arthritis. II. Low levels of synovial fluid and synovial tissue interferon suggest that gamma-interferon is not the primary macrophage activating factor Arthritis Rheum 1987 30 864 871 3115274
Matthews N Emery P Pilling D Akbar A Salmon M Subpopulations of primed T helper cells in rheumatoid arthritis Arthritis Rheum 1993 36 603 607 8098214
Buckley CD Pilling D Lord JM Akbar AN Scheel-Toellner D Salmon M Fibroblasts regulate the switch from acute resolving to chronic persistent inflammation Trends Immunol 2001 22 199 204 11274925 10.1016/S1471-4906(01)01863-4
Firestein GS Zvaifler NJ How important are T cells in chronic rheumatoid synovitis?: II. T cell-independent mechanisms from beginning to end Arthritis Rheum 2002 46 298 308 11840432 10.1002/art.502
Salmon M Scheel-Toellner D Huissoon AP Pilling D Shamsadeen N Hyde H D'Angeac AD Bacon PA Emery P Akbar AN Inhibition of T cell apoptosis in the rheumatoid synovium J Clin Invest 1997 99 439 446 9022077
Pilling D Akbar AN Girdlestone J Orteu CH Borthwick NJ Amft N Scheel-Toellner D Buckley CD Salmon M Interferon-beta mediates stromal cell rescue of T cells from apoptosis Eur J Immunol 1999 29 1041 1050 10092109 10.1002/(SICI)1521-4141(199903)29:03<1041::AID-IMMU1041>3.0.CO;2-#
Buckley CD Amft N Bradfield PF Pilling D Ross E Arenzana-Seisdedos F Amara A Curnow SJ Lord JM Scheel-Toellner D Persistent induction of the chemokine receptor CXCR4 by TGF-beta 1 on synovial T cells contributes to their accumulation within the rheumatoid synovium J Immunol 2000 165 3423 3429 10975862
Sebbag M Parry SL Brennan FM Feldmann M Cytokine stimulation of T lymphocytes regulates their capacity to induce monocyte production of tumor necrosis factor-alpha, but not interleukin-10: possible relevance to pathophysiology of rheumatoid arthritis Eur J Immunol 1997 27 624 632 9079801
McInnes IB Leung BP Sturrock RD Field M Liew FY Interleukin-15 mediates T cell-dependent regulation of tumor necrosis factor-alpha production in rheumatoid arthritis Nat Med 1997 3 189 195 9018238 10.1038/nm0297-189
Brennan FM Hayes AL Ciesielski CJ Green P Foxwell BM Feldmann M Evidence that rheumatoid arthritis synovial T cells are similar to cytokine-activated T cells: involvement of phosphatidylinositol 3-kinase and nuclear factor kappaB pathways in tumor necrosis factor alpha production in rheumatoid arthritis Arthritis Rheum 2002 46 31 41 11822409 10.1002/1529-0131(200201)46:1<31::AID-ART10029>3.0.CO;2-5
Curnow SJ Scheel-Toellner D Jenkinson W Raza K Durrani OM Faint JM Rauz S Wloka K Pilling D Rose-John S Inhibition of T cell apoptosis in the aqueous humor of patients with uveitis by IL-6/soluble IL-6 receptor trans-signaling J Immunol 2004 173 5290 5297 15470075
Raza K Lee CY Pilling D Heaton S Situnayake RD Carruthers DM Buckley CD Gordon C Salmon M Ultrasound guidance allows accurate needle placement and aspiration from small joints in patients with early inflammatory arthritis Rheumatology (Oxford) 2003 42 976 979 12730511
Arnett FC Edworthy SM Bloch DA McShane DJ Fries JF Cooper NS Healey LA Kaplan SR Liang MH Luthra HS The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis Arthritis Rheum 1988 31 315 324 3358796
Benjamini Y Hochberg Y Controlling the false discovery rate: a practical and powerful approach to multiple testing J R Stat Soc B 1995 57 289 300
Breiman L Random Forrests Machine Learning 2001 45 5 32 10.1023/A:1010933404324
Breiman L Manual on Setting Up, Using, and Understanding Random Forests v31 Last accessed 18 March 2005.
Cox TF Cox MAA Multidimensional scaling 2000 Chapman and Hall / CRC
Zvaifler NJ The immunopathology of joint inflammation in rheumatoid arthritis Adv Immunol 1973 16 265 336 4599390
Akbar AN Salmon M Cellular environments and apoptosis: tissue microenvironments control activated T-cell death Immunol Today 1997 18 72 76 9057357 10.1016/S0167-5699(97)01003-7
Pap T Muller-Ladner U Gay RE Gay S Fibroblast biology. Role of synovial fibroblasts in the pathogenesis of rheumatoid arthritis Arthritis Res 2000 2 361 367 11094449 10.1186/ar113
Tak PP Smeets TJ Daha MR Kluin PM Meijers KA Brand R Meinders AE Breedveld FC Analysis of the synovial cell infiltrate in early rheumatoid synovial tissue in relation to local disease activity Arthritis Rheum 1997 40 217 225 9041933
Conaghan PG O'Connor P McGonagle D Astin P Wakefield RJ Gibbon WW Quinn M Karim Z Green MJ Proudman S Elucidation of the relationship between synovitis and bone damage: a randomized magnetic resonance imaging study of individual joints in patients with early rheumatoid arthritis Arthritis Rheum 2003 48 64 71 12528105 10.1002/art.10747
Boers M Verhoeven AC Markusse HM van de Laar MA Westhovens R van Denderen JC van Zeben D Dijkmans BA Peeters AJ Jacobs P Randomised comparison of combined step-down prednisolone, methotrexate and sulphasalazine with sulphasalazine alone in early rheumatoid arthritis Lancet 1997 350 309 318 9251634 10.1016/S0140-6736(97)01300-7
Bathon JM Martin RW Fleischmann RM Tesser JR Schiff MH Keystone EC Genovese MC Wasko MC Moreland LW Weaver AL A comparison of etanercept and methotrexate in patients with early rheumatoid arthritis N Engl J Med 2000 343 1586 1593 11096165 10.1056/NEJM200011303432201
Visser H le Cessie S Vos K Breedveld FC Hazes JM How to diagnose rheumatoid arthritis early: a prediction model for persistent (erosive) arthritis Arthritis Rheum 2002 46 357 365 11840437 10.1002/art.10117
Mottonen T Hannonen P Korpela M Nissila M Kautiainen H Ilonen J Laasonen L Kaipiainen-Seppanen O Franzen P Helve T Delay to institution of therapy and induction of remission using single-drug or combination-disease-modifying antirheumatic drug therapy in early rheumatoid arthritis Arthritis Rheum 2002 46 894 898 11953964 10.1002/art.10135
Kurki P Aho K Palosuo T Heliovaara M Immunopathology of rheumatoid arthritis. Antikeratin antibodies precede the clinical disease Arthritis Rheum 1992 35 914 917 1379430
Berglin E Padyukov L Sundin U Hallmans G Stenlund H van Venrooij WJ Klareskog L Dahlqvist SR A combination of autoantibodies to cyclic citrullinated peptide (CCP) and HLA-DRB1 locus antigens is strongly associated with future onset of rheumatoid arthritis Arthritis Res Ther 2004 6 R303 R308 15225365 10.1186/ar1187
Nielen MM Van Schaardenburg D Reesink HW van de Stadt RJ van der Horst-Bruinsma IE de Koning MH Habibuw MR Vandenbroucke JP Dijkmans BA Specific autoantibodies precede the symptoms of rheumatoid arthritis: a study of serial measurements in blood donors Arthritis Rheum 2004 50 380 386 14872479 10.1002/art.20018
Bucht A Larsson P Weisbrot L Thorne C Pisa P Smedegard G Keystone EC Gronberg A Expression of interferon-gamma (IFN-gamma), IL-10, IL-12 and transforming growth factor-beta (TGF-beta) mRNA in synovial fluid cells from patients in the early and late phases of rheumatoid arthritis (RA) Clin Exp Immunol 1996 103 357 367 8608632
Schumacher HR Kitridou RC Synovitis of recent onset. A clinicopathologic study during the first month of disease Arthritis Rheum 1972 15 465 485 4117789
Firestein GS Evolving concepts of rheumatoid arthritis Nature 2003 423 356 361 12748655 10.1038/nature01661
Morita Y Yamamura M Kawashima M Harada S Tsuji K Shibuya K Maruyama K Makino H Flow cytometric single-cell analysis of cytokine production by CD4+ T cells in synovial tissue and peripheral blood from patients with rheumatoid arthritis Arthritis Rheum 1998 41 1669 1676 9751101 10.1002/1529-0131(199809)41:9<1669::AID-ART19>3.0.CO;2-G
Seo SK Choi JH Kim YH Kang WJ Park HY Suh JH Choi BK Vinay DS Kwon BS 4-1BB-mediated immunotherapy of rheumatoid arthritis Nat Med 2004 10 1088 1094 15448685 10.1038/nm1107
Heinzel FP Sadick MD Mutha SS Locksley RM Production of interferon gamma, interleukin 2, interleukin 4, and interleukin 10 by CD4+ lymphocytes in vivo during healing and progressive murine leishmaniasis Proc Natl Acad Sci USA 1991 88 7011 7015 1908085
Modlin RL Th1-Th2 paradigm: insights from leprosy J Invest Dermatol 1994 102 828 832 8006444 10.1111/1523-1747.ep12381958
Gerli R Bistoni O Russano A Fiorucci S Borgato L Cesarotti ME Lunardi C In vivo activated T cells in rheumatoid synovitis. Analysis of Th1- and Th2-type cytokine production at clonal level in different stages of disease Clin Exp Immunol 2002 129 549 555 12197898 10.1046/j.1365-2249.2002.01913.x
Joosten LA Lubberts E Helsen MM Saxne T Coenen-de Roo CJ Heinegard D van den Berg WB Protection against cartilage and bone destruction by systemic interleukin-4 treatment in established murine type II collagen-induced arthritis Arthritis Res 1999 1 81 91 11056663 10.1186/ar14
Jacobs MJ van den Hoek AE van Lent PL van de Loo FA van de Putte LB van den Berg WB Role of IL-2 and IL-4 in exacerbations of murine antigen-induced arthritis Immunology 1994 83 390 396 7835964
Jakubzick C Choi ES Kunkel SL Joshi BH Puri RK Hogaboam CM Impact of interleukin-13 responsiveness on the synthetic and proliferative properties of Th1- and Th2-type pulmonary granuloma fibroblasts Am J Pathol 2003 162 1475 1486 12707030
Kaufman J Sime PJ Phipps RP Expression of CD154 (CD40 ligand) by human lung fibroblasts: differential regulation by IFN-gamma and IL-13, and implications for fibrosis J Immunol 2004 172 1862 1871 14734771
Chiaramonte MG Donaldson DD Cheever AW Wynn TA An IL-13 inhibitor blocks the development of hepatic fibrosis during a T-helper type 2-dominated inflammatory response J Clin Invest 1999 104 777 785 10491413
Relic B Guicheux J Mezin F Lubberts E Togninalli D Garcia I van den Berg WB Guerne PA IL-4 and IL-13, but not IL-10, protect human synoviocytes from apoptosis J Immunol 2001 166 2775 2782 11160344
Salmon M Pilling D Borthwick NJ Akbar AN Inhibition of T cell apoptosis – a mechanism for persistence in chronic inflammation The Immunologist 1997 5 87 92
Bucala R Ritchlin C Winchester R Cerami A Constitutive production of inflammatory and mitogenic cytokines by rheumatoid synovial fibroblasts J Exp Med 1991 173 569 574 1997647 10.1084/jem.173.3.569
Parsonage G Falciani F Burman A Filer A Ross E Bofill M Martin S Salmon M Buckley CD Global gene expression profiles in fibroblasts from synovial, skin and lymphoid tissue reveals distinct cytokine and chemokine expression patterns Thromb Haemost 2003 90 688 697 14515190
Weyand CM Goronzy JJ Ectopic germinal center formation in rheumatoid synovitis Ann N Y Acad Sci 2003 987 140 149 12727633
Leung BP Conacher M Hunter D McInnes IB Liew FY Brewer JM A novel dendritic cell-induced model of erosive inflammatory arthritis: distinct roles for dendritic cells in T cell activation and induction of local inflammation J Immunol 2002 169 7071 7077 12471143
Zhou LJ Tedder TF CD14+ blood monocytes can differentiate into functionally mature CD83+ dendritic cells Proc Natl Acad Sci USA 1996 93 2588 2592 8637918 10.1073/pnas.93.6.2588
Santiago-Schwarz F Anand P Liu S Carsons SE Dendritic cells (DCs) in rheumatoid arthritis (RA): progenitor cells and soluble factors contained in RA synovial fluid yield a subset of myeloid DCs that preferentially activate Th1 inflammatory-type responses J Immunol 2001 167 1758 1768 11466401
Blades MC Ingegnoli F Wheller SK Manzo A Wahid S Panayi GS Perretti M Pitzalis C Stromal cell-derived factor 1 (CXCL12) induces monocyte migration into human synovium transplanted onto SCID Mice Arthritis Rheum 2002 46 824 836 11920421 10.1002/art.10102
Miossec P Interleukin-17 in rheumatoid arthritis: if T cells were to contribute to inflammation and destruction through synergy Arthritis Rheum 2003 48 594 601 12632409 10.1002/art.10816
Fossiez F Djossou O Chomarat P Flores-Romo L Ait-Yahia S Maat C Pin JJ Garrone P Garcia E Saeland S T cell interleukin-17 induces stromal cells to produce proinflammatory and hematopoietic cytokines J Exp Med 1996 183 2593 2603 8676080 10.1084/jem.183.6.2593
Jovanovic DV Di Battista JA Martel-Pelletier J Jolicoeur FC He Y Zhang M Mineau F Pelletier JP IL-17 stimulates the production and expression of proinflammatory cytokines, IL-beta and TNF-alpha, by human macrophages J Immunol 1998 160 3513 3521 9531313
Akbar AN Borthwick NJ Wickremasinghe RG Panayoitidis P Pilling D Bofill M Krajewski S Reed JC Salmon M Interleukin-2 receptor common gamma-chain signaling cytokines regulate activated T cell apoptosis in response to growth factor withdrawal: selective induction of anti-apoptotic (bcl-2, bcl-xL) but not pro- apoptotic (bax, bcl-xS) gene expression Eur J Immunol 1996 26 294 299 8617294
| 15987480 | PMC1175027 | CC BY | 2021-01-04 16:02:37 | no | Arthritis Res Ther. 2005 Apr 7; 7(4):R784-R795 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1733 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17341598747410.1186/ar1734Research ArticleDynamic magnetic resonance of the wrist in psoriatic arthritis reveals imaging patterns similar to those of rheumatoid arthritis Cimmino Marco A [email protected] Massimiliano [email protected] Stefania [email protected] Giulia [email protected] Simone [email protected] Enzo [email protected] Giacomo [email protected] Clinica Reumatologica, Dipartimento di Medicina Interna e Specialità Mediche, Università di Genova, Italy2 ESAOTE Biomedica, Genova, Italy3 Sezione di Diagnostica Radiologica, Dipartimento di Medicina Sperimentale, Università di Genova, Italy2005 1 4 2005 7 4 R725 R731 14 12 2004 17 1 2005 2 3 2005 7 3 2005 Copyright © 2005 Cimmino et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
This dynamic magnetic resonance imaging (MRI) study is concerned with a prospective evaluation of wrist synovitis in patients with psoriatic arthritis (PsA) in comparison with patients with rheumatoid arthritis (RA) and healthy controls. Fifteen consecutive patients with PsA, 49 consecutive patients with RA, 30 RA patients matched for disease severity with those with PsA, and 8 healthy controls were studied. MRI was performed with a low-field (0.2T), extremity-dedicated machine. After an intravenous bolus injection of gadolinium-diethylenetriaminepentaacetic acid, 20 consecutive fast spin-echo axial images of the wrist were obtained every 18 s. The enhancement ratio was calculated both as rate of early enhancement (REE), which shows the slope of the curve of contrast uptake per second during the first 55 s, and as relative enhancement (RE), which indicates the steady state of enhancement. The REE was 1.0 ± 0.6 in patients with PsA, 1.6 ± 0.7 in consecutive patients with RA, and 0.1 ± 0.1 in controls (p <0.001). The RE was 87.1 ± 39.2 in patients with PsA, 125.8 ± 48.0 in consecutive RA patients, and 15.5 ± 19.2 in controls (p <0.001). However, the same figures in matched RA patients were 1.3 ± 0.7 and 107.3 ± 48.2, respectively (not significant in comparison with PsA). Rheumatoid-like PsA and oligoarticular PsA did not differ from each other in terms of synovial enhancement. Dynamic MRI shows the same pattern of synovitis in patients with PsA and RA when the two groups are matched for disease severity. This technique cannot be used to differentiate PsA from RA. However, REE and RE were significantly higher in PsA than in normal controls, with only one instance of overlap between values found for the two groups.
==== Body
Introduction
Psoriatic arthritis (PsA), defined as the occurrence of seronegative arthritis and psoriasis, is a debated entity. The reasons for this uncertainty include disparity in the subgroups identified in the original description [1], lack of a validated case definition of PsA, possible inclusion of patients with enthesitis but without clear arthritis, the elusive link between skin and joint diseases, and similarities with rheumatoid arthritis (RA). It is not known if PsA and RA are completely distinct diseases, or if PsA is a form of RA modified by the coexisting psoriasis. This second hypothesis is brought into doubt by the original observation that articular involvement is chiefly symmetric and polyarticular in RA, but oligoarticular and asymmetric in PsA.
However, subsequent studies have shown that the most common subset of PsA is a symmetric polyarthritis resembling RA [2]. In particular, oligoarticular disease may be a characteristic of PsA mainly at presentation [3], and symmetry is a function of the number of joints involved but not of the type of arthritis [4]. On average, however, PsA is characterized by a milder degree of synovitis than RA, with only 8% of patients developing erosions of the hand. Therefore, therapy with disease-modifying anti-rheumatic drugs is rarely needed [5].
Magnetic resonance imaging (MRI) has helped in the evaluation of PsA by suggesting that the primary site of inflammation is extrasynovial and that synovial inflammation may be a secondary phenomenon [6]. In addition to morphologic studies of the joint, MRI may be used also to evaluate synovial membrane inflammation through a dynamic, contrast-enhanced technique. By this method, we evaluated gadolinium perfusion of the synovial membrane to differentiate active RA from RA in remission [7]. In the present paper, the same technique has been used to study patients with PsA. The main goals of the study were to investigate whether synovitis differs between PsA and RA, and, if so, whether the difference is intrinsic to the type of arthritis or is due to the severity and duration of the disease.
Materials and methods
Patients
Fifteen consecutive patients with PsA, defined as the simultaneous occurrence of active arthritis and psoriasis, were prospectively studied. Eight patients had a polyarticular rheumatoid-like pattern of arthritis and seven had monoarthritis or oligoarthritis, associated in one patient with axial involvement. Dynamic MRI data for these patients were compared with those of 49 consecutive patients with active RA (group I), diagnosed according to the criteria of the American Rheumatism Association [8]. An additional group of 30 patients with RA, matched with those with PsA in terms of age, disease duration, and number of involved joints, was considered (group II). These patients were identified in ongoing follow-up studies. Because of difficulties in matching, nine of these patients were recruited from groupI RA patients. All the patients considered in this study were seen at the Rheumatological Clinic of the University of Genoa, Italy, either in the clinical ward or as outpatients, and had clinical inflammatory involvement of at least one wrist. Clinical parameters, such as duration of arthritis, early-morning stiffness, fatigue, number of tender and swollen joints, and type of treatment, were evaluated before MRI. At the same time, blood was drawn for the determination of the erythrocyte sedimentation rate (Westergren method), C-reactive protein, and IgM rheumatoid factor by standard laboratory methods. In addition, a control group of eight healthy volunteers, who did not report any history of joint disease and were negative for signs of arthritis on clinical examination, was also studied. All subjects gave their informed consent to the protocol, which had been approved by the ethics committee of the Department of Internal Medicine of the University of Genoa. Demographic and clinical characteristics of the three groups of patients and the group of controls are reported in Table 1.
Methods
MRI of the wrist was performed with a low-field (0.2T), extremity-dedicated machine (Artoscan™, Esaote, Genoa, Italy) equipped with a permanent magnet and with a dedicated hand-and-wrist coil 13 cm in diameter, as previously described [6]. The hand was fixed in neutral position and the fingers in extended position with the thumb up, by the application of several cushions. The field of view was 120 mm and allowed the evaluation of the carpal bones, the metacarpal bases, and the distal radius and ulna. Slice thickness was 5 mm and the interslice gap was 0.3 mm. The sequence used was a spin echo (TR/TE = 100/16 ms, matrix = 160 × 128, FOV = 150 × 150), which was acquired in the axial plane. In patients with arthritis, the more severely affected wrist was examined. In patients with equal involvement of the wrists, and in normal controls, the right wrist was examined. Patients and controls were instructed to avoid intense activity involving the wrists in the 24 hours preceding the examination.
After the wrist was positioned in the gantry, the first image was acquired. Then an intravenous bolus injection of 0.2 ml/kg of Gd-DTPA (gadolinium-diethylenetriaminepentaacetic acid) (Omniscan, Schering, Germany) was given manually in 30 s through a 21-mm butterfly needle into a cubital vein. Twenty consecutive fast images of three slices of the wrist, the first of which was positioned tangential to the radius, were repeated every 18 s thereafter. The rate of enhancement was evaluated by a radiologist on the slice that showed the highest visual enhancement. It was calculated as Δ on a small, elliptical region of interest (ROI) of synovial membrane of approximately 25 mm2 positioned in the area of highest visual enhancement (Fig. 1). Entheses and synovial sheaths were not included in the ROI. In healthy controls, the synovial membrane was more difficult to identify. In these cases, the area where the synovium was thought to be located was chosen on the basis of anatomic landmarks and comparison of pre- and post-enhancement images. For this reason, the elliptical ROI was usually smaller in controls than in patients. This corresponded usually to the median and dorsal area of the wrist.
The images were processed blind to the clinical and laboratory findings. The enhancement ratio was calculated both as rate of early enhancement (REE) per second during the first 55 s according to the formula
REE55 = (S55 - S0)/(S0 × 55) × 100%
and as relative enhancement (RE) at t seconds according to the formula
REt = (St - S0)/S0 × 100%
where S0 and St are the signal-to-noise ratios, before and t seconds after contrast injection, calculated as ratio between the signal measured in the ROI and the standard deviation of the background noise. The study of enhancement after 55 s was chosen because it showed maximal enhancement difference between knees with clinically inactive or active disease in a previous study [9]. In addition, the signal was normalized to the bone to reduce noise. The REE shows the slope of the curve of contrast uptake tangential to the α angle and is steeper if inflammation is higher. The RE indicates the steady state of enhancement. The intra- and inter-observer mean percentage variations for REE were 3.9% (range 0.5% to 14.3%) and 2.8% (range 0% to 5.1%), respectively, in 18 wrists. Intra- and inter observer mean percentage variation for RE were 1.9% (range 0% to 9.3%) and 1.9% (range 0.05% to 6.4%) (manuscript in preparation).
Statistical evaluation
Means were compared by the Student's t-test or by one-way analysis of variance (ANOVA) if their distribution was normal and by the Wilcoxon test with Mann–Whitney correction or Kruskal–Wallis ANOVA when the distribution was nonparametric. Frequencies were compared using the Fisher exact test. Correlations were calculated by the Pearson or Spearman rank tests. P values less than 0.05 were considered significant.
Results
Comparison between PsA and consecutive RA patients
The demographic and clinical data for patients and controls are reported in Table 1 in comparison with RA patients of group I. PsA patients had a less skewed female-to-male ratio, had a lower tender joint count, and were less frequently positive for IgM rheumatoid factor. All the other clinical characteristics were similar in the two groups. Controls were younger than the patients of the other two groups. Dynamic MRI was performed in all subjects without causing any discomfort or adverse events. The mean duration of the complete examination was 15 min. REE and RE were significantly different in the three groups (p <0.001) (Table 1; Fig. 2). The values were highest for the group with RA, followed by that with PsA and controls. In PsA and RA patients, the REE ranged from 0.14% to 2.13% per second, and from 0.08% to 3.49% per second, respectively. In controls, it ranged from -0.02% to 0.37% per second.
Nonsteroidal anti-inflammatory drugs were being used by 14 (93.3%) of 15 PsA patients, and by 40 (81.6%) of 49 RA patients (ns). Sulphasalazine was used in 7 (47.7%) of 15 PsA patients and in 17 (34.7%) of 49 RA patients (ns). Conversely, patients with PsA were treated less frequently with prednisone (4 of 15, or 26.7%, vs 36 of 49, or 73.5%; P = 0.002) or methotrexate (2 of 15, or 13.3%, vs 21 of 49, or 42.9%, P = 0.06) than patients with RA. Dosages were similar in the two groups of patients (data not shown).
Comparison between PsA and matched RA patients
To exclude the possibility that a lower disease activity in PsA patients could account for the observed dynamic MRI difference, 2 RA patients were matched for age and number of tender and swollen joints to each of the 15 patients with PsA. The clinical and laboratory characteristics of the PsA patients and of the new RA control group (group II) were similar, with no statistically significant difference. Only positivity for IgM rheumatoid factor was higher in RA patients (P < 0.05) (Table 1).
Treatment in the matched RA group (II) was similar to that seen in the groupII RA patients. The use of nonsteroidal anti-inflammatory drugs and sulphasalazine was not different in PsA patients from that in matched RA patients. Patients with PsA were treated less frequently with prednisone (26.7% vs 70%; P = 0.01) and with methotrexate (13.3% vs 46.6%, P = 0.046) than patients with RA, but dosages were similar.
REE and RE were not different between PsA and RA of similar severity (Table 1; Fig. 2). Figure 3 compares the mean values of the curves in the three subgroups, that is, patients with PsA, groupII RA patients, and healthy controls. Before Gd-DTPA infusion and for the first 36 s after infusion, the three curves were almost identical. However, a highly significant difference in enhancement was seen by ANOVA at all the following time points (P = 0.003 at t = 156 s, p <0.001 at t = 174 s, and P < 0.001 thereafter). The curves identified two groups of patients, one being patients with PsA or groupII RA and the other being controls.
Correlations between dynamic MRI and clinical and laboratory findings
REE was 0.8 ± 0.5 in patients with rheumatoid-like PsA and 1.2 ± 0.6 in those with monoarthritis or oligoarthritis (ns). Values for RE were 78 ± 28.2 and 97.5 ± 49.2, respectively. This difference was not significant. REE and RE were not correlated with clinical and laboratory findings in PsA. There was a tendency to an association between REE and number of swollen joints in the RA patients of group II (P = 0.07).
Discussion
Dynamic MRI is a promising method for the investigation of patients with arthritis. In our previous experience, it could differentiate patients with active RA from those in remission and from controls [7]. In another study of arthritic knees, dynamic gadolinium-enhanced MRI showed increased contrast diffusion in comparison with controls [9]. These results are in keeping with the well-known ability of MRI to detect synovitis in diseased joints [10]. We therefore decided to evaluate patients with PsA and to compare them with RA, in view of the existing debate on similarities and differences between the two diseases.
After intravenous injection, Gd-DTPA, a relatively small molecule, rapidly diffuses to highly vascularized tissues, such as the inflamed synovial membrane, and the rapidity and amount of diffusion seem to be related to the number, size, and permeability of synovial vessels as well as to the volume of the synovial membrane [11]. When PsA patients were compared with an unselected group of consecutive RA patients, the REE and RE were significantly lower in those with PsA. This observation could lead to the interpretation that inflammation of the synovial membrane is lower in PsA, a finding supported by several epidemiological and clinical studies [5]. However, the suspicion arose that the two groups of patients might not be comparable, because of higher disease activity in patients with RA. In fact, the number of tender joints was significantly higher in RA patients (Table 1), although the other disease characteristics were only slightly higher. We had the opportunity to match each patient with PsA with two RA patients for number of tender and swollen joints. As a result, a new RA group (group II) was formed that included some of the consecutive RA patients and several new patients drawn from ongoing follow-up studies. The two groups of patients were clinically fully comparable. Of course, the percentage of patients with IgM rheumatoid factor could not be easily matched in this type of study. Another difference between groups included the more frequent administration of prednisone and methotrexate in RA patients. In fact, matching also for treatment was not possible due to the relatively small number of RA patients. It could be argued that prednisone and methotrexate could affect dynamic MRI per se by directly acting on neovascularization. However, these two drugs were used more frequently in both RA groups, which behave differently from PsA as far as REE and RE are concerned. It is therefore unlikely that their direct effect on dynamic MRI could have influenced our results, which are, rather, explained by a difference in inflammation.
After matching, the REE and RE were not significantly different in the two groups of patients (Fig. 3). This finding indicates that, at comparable levels of disease severity, synovitis revealed by dynamic MRI presents a pattern in PsA that is similar to that of RA. This finding contradicts the common belief that PsA, on the whole, is a mild form of arthritis. The amount of contrast agent transported to the inflamed synovial membrane is probably a result of the number, size, and permeability of vessels and volume of the synovial membrane itself. A greater number of synovial vessels per squaremillimetre of tissue has been demonstrated in PsA than in RA [12]. Conversely, significantly less lining-layer hyperplasia was demonstrated in PsA in the same study [12]. The net effect of these two contrasting features on Gd-DTPA diffusion is not known and could be assessed only by comparing dynamic MRI and synovial membrane histology in the same joints. Other vessel characteristics of PsA synovial membrane that could play a role in contrast agent diffusion are the marked thickening of the vessel wall [13] and the peculiar, tortuous vascular pattern [14].
Dynamic MRI highlights the similarity of the synovial membrane in PsA and RA and supports the view that the two conditions may be more similar than is usually believed, at least as far as disease activity is concerned. This observation is in keeping with the fact that the same types of treatment, including sulphasalazine, methotrexate, leflunomide, and anti-tumor-necrosis-factor-α compounds, are effective in RA and PsA. As a result, dynamic MRI cannot be used to differentiate the two diseases. However, both REE and RE data were significantly higher in PsA than in healthy controls, with only one case of overlap between the two conditions.
A more efficient way of differentiating PsA from RA by MRI is to study the pattern of joint involvement. Jevtic and colleagues [15] showed that inflammation is localized within the joint capsule in the small joints of the hand of RA patients, whereas PsA patients also show extracapsular involvement, with thickened collateral ligaments and oedema of the neighbouring soft tissues. In another study of the knee, focal perientheseal high signal outside the joint, and bone marrow oedema at entheseal insertions were typical features of patients with spondyloarthropathy [16]. Our study, in which we aimed to compare the degree of synovitis in the two forms of arthritis, did not take into consideration damage to entheses and synovial sheaths, areas that could be more effective in the differential diagnosis and deserve a separate investigation.
Our results were obtained with a low-field extremity-dedicated MRI device. The lack of discrimination between RA and PsA with a 0.2-T MRI device could be overcome with other MRI protocols, such as different pulse sequences, field strengths, or magnet types. High-field MRI machines have a better signal-to-noise ratio and could be hypothetically more sensible in the evaluation of enhancement. Evaluation of the sensitivity of dynamic MRI is still in its early stages. Within RA, we showed that this technique can discriminate between different degrees of clinical activity [7]. These considerations suggest that differences in dynamic MRI between RA and PsA, if present, should be relatively small. There are no papers directly comparing dynamic MRI obtained with low- and high-field machines. Results of another recent study by Palosaari and colleagues on wrist RA [17], made using a low-field MRI, are difficult to compare with ours, because technical features such as type of sequence, imaging parameters, acquisition plane, number of sequences, and amount of contrast agent were different. In addition, the cohort of RA patients in that study, being affected by early disease, was different. The absolute values of signal enhancement were higher in Palosaari's study. By contrast, a third study [18] on the rheumatoid knee performed with a 1.5-T unit showed absolute enhancement values similar to those that we obtained.
The intraobserver/interobserver agreement in evaluating dynamic enhancement was very high. This may be surprising, in view of the fact that the examination process included selection of slice, of maximal enhancing area, and of the size of the ROI. We feel that high reproducibility may have been facilitated by a significant association between enhancement figures of elliptical areas in the three sequentially acquired wrist slices [unpublished observations in [7]]. This makes the choice of the slice less important. In addition, the area of maximum enhancement, exclusive of entheses and tendon sheaths, is very often located on the dorsal side of the wrist and is relatively small, another constraint of choice for the examiner. Elliptical ROIs, although apparently less logical from a pathophysiological point of view than ROIs outlining the enhanced synovial membrane, were chosen to improve reproducibility and standardization. A recent paper on contrast-enhanced dynamic MRI of coronal slices of the wrist also showed high intraobserver reliability [17].
Only one PsA patient was positive for rheumatoid factor. She did not show a rheumatoid-like pattern of joint involvement, but had monoarthritis of the right wrist and dactylitis, which is not typical of RA. Of the 15 PsA patients, 8 had a rheumatoid-like pattern of arthritis and 7 had oligomonoarthritis. Dynamic MRI was not significantly different between these two groups, reinforcing the suggestion that severity of the arthritis, and not its type or subtype, is associated with MRI findings.
The patients were selected on the basis of clinical involvement of the wrist. This prerequisite was set because, if all consecutive patients had been enrolled, many more RA patients than PsA patients would have had wrist involvement, thus making the two groups more difficult to compare. In a previous study by our team [7], RA patients in remission and without wrist arthritis had dynamic MRI values significantly lower than those with active disease and wrist involvement. We do not know if the results obtained in unaffected wrists of otherwise active arthritis patients reflect local (wrist) or general disease activity. No correlation was found between indexes of severity and dynamic MRI in either PsA and RA. This last finding is surprising in view of our previous results on the high correlation between clinical and laboratory indexes of RA inflammation and dynamic MRI. However, the inclusion of severely active arthritis only – with exclusion of patients in remission – and the relatively low mean number of affected joints in the patients with PsA and in the matched RA controls may account for the difference. Nonetheless, a tendency to correlation between number of swollen joints and REE, which in our experience is the more sensitive measure, was observed.
Conclusion
We have shown that dynamic MRI gives similar results in PsA and RA, suggesting that the type and degree of inflammatory process is similar in the two diseases.
Abbreviations
ANOVA = analysis of variance; Gd-DTPA = gadolinium-diethylenetriaminepentaacetic acid; MRI = magnetic resonance imaging; ns = not significant; PsA = psoriatic arthritis; RA = rheumatoid arthritis; RE = relative enhancement; REE = rate of early enhancement; ROI = region of interest.
Competing interests
SI is an employee of ESAOTE, the manufacturer of the magnetic resonance device.
Authors' contributions
MAC and MP contributed to the conception and design of the study, to the clinical and MRI evaluation of the patients, to the analysis and interpretation of data, to the drafting of the article, and to the critical revision of the article for important intellectual content; SI contributed to the analysis and interpretation of data and to the critical revision of the article for important intellectual content; ES and GG contributed to the conception and design of the study, to the analysis and interpretation of data, and to the critical revision of the article for important intellectual content; GS and SB contributed to the conception and design of the study, to the clinical and MRI evaluation of the patients, to the analysis and interpretation of data, and to the critical revision of the article for important intellectual content. All authors read and approved the final manuscript.
Acknowledgements
This work was supported in part by a grant from the University of Genoa, Italy.
Figures and Tables
Figure 1 Dynamic gadolinium-diethylenetriamine pentaacetic acid (Gd-DTPA)-enhanced MRI of the wrist in a patient with psoriatic arthritis. Sequence (a) shows the precontrast image; sequences (b–d) show images acquired after 36, 90, and 180 s, respectively. The region of interest on which the enhancement curve has been calculated is outlined. Gd-DTPA, dynamic gadolinium-diethylenetriamine pentaacetic acid; MRI, magnetic resonance image.
Figure 2 Individual values of REE (left) and RE (right) in patients with arthritis and controls. (A) Patients with psoriatic arthritis, (B) group-I rheumatoid arthritis, or (C) groupII rheumatoid arthritis; (D) controls. Triangles indicate the mean values. Vertical bars indicate standard deviations. RE, relative enhancement; REE, rate of early enhancement.
Figure 3 Slope of the mean enhancement curves in patients with arthritis and in controls. Patients with psoriatic arthritis (squares) or with rheumatoid arthritis (diamonds) matched for demographic characteristics and disease severity; controls (triangles). The arrow indicates the time of gadolinium-diethylenetriamine pentaacetic acid infusion.
Table 1 Demographic, clinical, and laboratory characteristics of patients with psoriatic arthritis, rheumatoid arthritis, and controls
Characteristic Psoriatic arthritis Rheumatoid arthritis (group I) Matched rheumatoid arthritis (group II) Controls
Number of patients 15 49 30 8
Age (years) 55.7 ± 10.7 57.6 ± 14.6 56.3 ± 16.3 38.1 ± 21.9
Sex (women/men)* 8/7 42/7 20/10 4/4
Disease duration (months) 63.5 ± 62.7 93.5 ± 98.7 78.2 ± 86.2 NA
Morning stiffness (minutes) 70.4 ± 55.7 96.2 ± 96.3 76.8 ± 92.1 NA
Number of tender joints* 8.0 ± 5.5 13.7 ± 9.4 9.1 ± 7.3 0
Number of swollen joints 5.3 ± 4.4 8.0 ± 5.8 5.0 ± 3.9 0
Ritchie index 8.3 ± 5.7 11.6 ± 6.9 7.0 ± 5.1 0
ESR (mm/h) 34.7 ± 21.8 49.1 ± 31.6 45.5 ± 35.7 ND
CRP (mg/l) 30.4 ± 39.4 30.3 ± 34.4 26.6 ± 36.1 ND
Number of patients with IgM rheumatoid factor** 1 (6.7%) 33 (67.3%) 18 (60%) ND
Rate of early enhancement§ 1.0 ± 0.6 1.6 ± 0.7 1.3 ± 0.7 0.1 ± 0.1
Relative enhancement § 87.1 ± 39.2 125.8 ± 48.0 107.3 ± 48.2 15.5 ± 19.2
Values are expressed as means ± standard deviations except for sex and IgM rheumatoid factor, for which the percentage of positive patients is reported. *P < 0.05 for the comparison between PsA and groupI RA; **P < 0.001 for the comparison between PsA and groupI RA and P < 0.05 for the comparison between PsA and groupII RA; §P < 0.001 by one-way analysis of variance between PsA patients, groupI RA patients, and controls.
==== Refs
Moll JMH Wright V Psoriatic arthritis Semin Arthritis Rheum 1973 3 55 78 4581554 10.1016/0049-0172(73)90035-8
Helliwell P Marchesoni A Peters M Barker M Wright V A re-evaluation of the osteoarticular manifestations of psoriasis Br J Rheumatol 1991 30 339 345 1913001
Marsal S Armadans-Gil L Martinez M Gallardo D Ribera A Lience E Clinical, radiographic and HLA associations as markers for different patterns of psoriatic arthritis Rheumatology 1999 38 332 337 10378711 10.1093/rheumatology/38.4.332
Helliwell PS Hetthen J Sokoll K Green M Marchesoni A Lubrano E Veale D Emery P Joint symmetry in early and late rheumatoid and psoriatic arthritis Arthritis Rheum 2000 43 865 871 10765932 10.1002/1529-0131(200004)43:4<865::AID-ANR18>3.0.CO;2-W
Shbeeb M Uramoto KM Gibson LE O'Fallon WM Gabriel SE Epidemiology of psoriatic arthritis in Olmsted County, Minnesota, USA, 1982–1991 J Rheumatol 2000 27 1247 1250 10813295
McGonagle D Conaghan PG Emery P Psoriatic arthritis: a unified concept twenty years on Arthritis Rheum 1999 42 1080 1086 10366099 10.1002/1529-0131(199906)42:6<1080::AID-ANR2>3.0.CO;2-7
Cimmino MA Innocenti S Livrone F Magnaguagno F Silvestri E Garlaschi G Dynamic gadolinium-enhanced magnetic resonance imaging of the wrist in patients with rheumatoid arthritis can discriminate active from inactive disease Arthritis Rheum 2003 48 1207 1213 12746893 10.1002/art.10962
Arnett FC Edworthy SM Bloch DA McShane DJ Fries JF Cooper NS Healey LA Kaplan SR Liang MH Luthra HS The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis Arthritis Rheum 1988 31 315 324 3358796
Østergaard M Lorenzen I Henriksen O Dynamic gadolinium-enhanced MR imaging in active and inactive immunoinflammatory gonarthritis Acta Radiol 1994 35 275 281 8192967
Cimmino MA Bountis C Silvestri E Garlaschi G Accardo S An appraisal of magnetic resonance imaging of the wrist in rheumatoid arthritis Semin Arthritis Rheum 2000 30 180 195 11124282 10.1053/sarh.2000.9204
Gaffney K Cookson J Blades S Coumbe A Blake D Quantitative assessment of the rheumatoid synovial microvascular bed by gadolinium-DTPA enhanced magnetic resonance imaging Ann Rheum Dis 1998 57 152 157 9640130
Veale D Yanni G Rogers S Barnes L Bresnihan B Fitzgerald O Reduced synovial membrane macrophage numbers, ELAM-1 expression, and lining layer hyperplasia in psoriatic arthritis as compared with rheumatoid arthritis Arthritis Rheum 1993 36 893 900 7686370
Espinoza LR Vasey FB Espinoza CG Bocanegra TS Germain BF Vascular changes in psoriatic synovium. A light and electron microscopic study Arthritis Rheum 1982 25 677 684 7092966
Cañete JD Rodríguez JR Salvador G Gómez-Centeno A Muñoz-Gómez J Sanmartí R Diagnostic usefulness of synovial vascular morphology in chronic arthritis. A systematic survey of 100 cases Semin Arthritis Rheum 2003 32 378 387 12833246 10.1053/sarh.2002.50004
Jevtic V Watt J Rozman B Kos-Golja M Demsar F Jarh O Distinctive radiological features of small hand joints and seronegative spondyloarthritis demonstrated by contrast-enhanced (Gd-DTPA) magnetic resonance imaging Skeletal Radiol 1995 24 351 355 7570155 10.1007/BF00197064
McGonagle D Gibbon W O'Connor P Green M Pease C Emery P Characteristic magnetic resonance imaging entheseal changes of knee synovitis in spondyloarthropathy Arthritis Rheum 1998 41 694 700 9550479 10.1002/1529-0131(199804)41:4<694::AID-ART17>3.0.CO;2-#
Palosaari K Vuotila J Takalo R Jartti A Niemelä R Haapea M Soini I Tervonen O Hakala M Contrast-enhanced dynamic and static MRI correlates with quantitative 99Tcm-labelled nanocolloid scintigraphy. Study of early rheumatoid arthritis patients Rheumatology 2004 43 1364 1373 15238644 10.1093/rheumatology/keh302
Østergaard M Stoltenberg M Henriksen O Lorenzen I Quantitative assessment of synovial inflammation by dynamic gadolinium-enhanced magnetic resonance imaging. A study of the effect of intra-articular methylprednisolone on the rate of early synovial enhancement Br J Rheumatol 1996 35 50 59 8624624
| 15987474 | PMC1175028 | CC BY | 2021-01-04 16:02:37 | no | Arthritis Res Ther. 2005 Apr 1; 7(4):R725-R731 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1734 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17351598747910.1186/ar1735Research ArticleMithramycin downregulates proinflammatory cytokine-induced matrix metalloproteinase gene expression in articular chondrocytes Liacini Abdelhamid 1Sylvester Judith 1Li Wen Qing 1Zafarullah Muhammad [email protected] Département de Médecine and Centre de Recherche du Centre Hospitalier de l'Université de Montréal (CHUM), Hôpital Notre-Dame du CHUM, Montréal, Québec, Canada2005 4 4 2005 7 4 R777 R783 18 7 2003 15 8 2003 21 2 2005 7 3 2005 Copyright © 2005 Zafarullah et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Interleukin-1 (IL-1), IL-17 and tumor necrosis factor alpha (TNF-α) are the main proinflammatory cytokines implicated in cartilage breakdown by matrix metalloproteinase (MMPs) in arthritic joints. We studied the impact of an anti-neoplastic antibiotic, mithramycin, on the induction of MMPs in chondrocytes. MMP-3 and MMP-13 gene expression induced by IL-1β, TNF-α and IL-17 was downregulated by mithramycin in human chondrosarcoma SW1353 cells and in primary human and bovine femoral head chondrocytes. Constitutive and IL-1-stimulated MMP-13 levels in bovine and human cartilage explants were also suppressed. Mithramycin did not significantly affect the phosphorylation of the mitogen-activated protein kinases, extracellular signal-regulated kinase, p38 and c-Jun N-terminal kinase. Despite effective inhibition of MMP expression by mithramycin and its potential to reduce cartilage degeneration, the agent might work through multiple unidentified mechanisms.
==== Body
Introduction
A major pathological manifestation of patients with osteoarthritis (OA) and rheumatoid arthritis is the degeneration of articular cartilage [1,2]. Matrix metalloproteinases (MMPs) such as MMP-3 and MMP-13 are known to cleave collagens and aggrecan of cartilage extracellular matrix [3-5]. The concentrations of several MMPs are increased in cartilage, synovial membrane and synovial fluid of patients with arthritis [6,7]. Indeed, cartilage-specific overexpression of active human MMP-13 causes OA in mice [8]. Proinflammatory cytokines, interleukin-1 (IL-1), IL-17 and tumor necrosis factor (TNF)-α are also increased in arthritic joints and are known to induce catabolic pathways leading to an enhanced expression of MMPs [9-11]. Inhibition of these proteases is regarded as an important approach for reducing damage in arthritic tissues [12].
AP-1 binding sites found in the promoter regions of the genes encoding MMP-3 and MMP-13 are essential for the expression of these genes [13,14]. Sp1 transcription factor is a zinc-finger type transcription factor whose binding sites are found in numerous housekeeping and inducible genes [15]. Human MMP-13 promoter has one putative Sp1 consensus site [16]. Mithramycin is an aureolic acid anti-neoplastic antibiotic that is used for treating cancer-related hypercalcemia [17]. Previous work has revealed that it inhibits bone resorption in vitro, possibly by interfering with bone cell lysosomal enzymes [18]. It also prevents the binding of Sp1 transcription factor to its cognate site in DNA by modifying the CG sequences [19]. Here we have studied the impact of mithramycin on proinflammatory cytokine-induced MMP expression. We show for the first time that mithramycin potently suppresses MMP induction by IL-1, IL-17 and TNF-α in chondrocytic cells without impairing the activation of mitogen-activated protein kinases (MAPKs).
Materials and methods
Primary cultures of human and bovine chondrocytes, SW1353 cells and treatments
Human cartilage was acquired from the femoral heads of OA patients who underwent hip-replacement surgery at the Notre-Dame Hospital. Normal bovine cartilage was obtained from the knee and hip joints of adult animals from a local abattoir. Chondrocytes were released by 90 min pronase and 9 hours digestion with collagenase (Sigma type IA). The cells were washed with PBS and grown in DMEM containing 10% FCS as high-density primary monolayer cultures until confluent growth. Cells were distributed in six-well plates, grown to confluence, washed with PBS and kept in serum-free DMEM for 24 hours; mithramycin (from Sigma-Aldrich Canada Ltd, Oakville, Ontario; dissolved in water as a 10 mM solution) was then added without medium change at final concentrations of 100 and 150 nM (doses known to inhibit Sp1 binding [20]) for 30 min before treatment for 24 hours with human recombinant IL-1β (10 ng/ml), TNF-α (20 ng/ml) and IL-17 (20 ng/ml) (R&D systems, Minneapolis, MN). The human chondrosarcoma cell line SW1353 was obtained from the American Type Culture Collection (ATCC, Manassas, VA) and treated as described for primary chondrocytes.
Northern hybridization analysis
Total cellular RNA was extracted by the guanidinium procedure [21] and aliquots of 3 to 5 μg were analyzed by electrophoretic fractionation in 1.2% formaldehyde-agarose gels. The integrity and quantity of RNA were verified by ethidium bromide staining of the 28S and 18S ribosomal RNA bands. The RNA was transferred onto Zeta-probe nylon membrane with a Bio-Rad Transblot in the presence of 0.5 × TAE (Tris-acetate-EDTA) buffer at a current of 500 mA for 12 hours. Northern blots were hybridized with a human stromelysin cDNA probe generously provided by Dr Richard Breathnach (Nantes, France). This probe was a 1.6-kilobase EcoRI cDNA fragment cloned in the plasmid pGEM-4Z (Promega Biotech, Madison, WI) and the vector was linearized with NarI. A 491-base-pair RT-PCR-generated [22] and cloned collagenase-3 cDNA was linearized with EcoRI. The human 28S ribosomal RNA plasmid (ATCC) was digested with XbaI. All antisense RNA probes were synthesized with T7 polymerase in accordance with the protocols of Promega Biotech and were labeled to high specific radioactivity (108 c.p.m./μg) with [α-32P]CTP (3,000 Ci/mmol; Perkin Elmer Life Sciences Inc., Boston, MA).
Western immunoblot analysis
Total secreted proteins from the 2 to 3 ml of conditioned medium of the chondrocytes or SW1353 cells were concentrated by precipitation with trichloroacetic acid, quantified with the Bio-Rad protein assay system and different amounts of protein aliquots adjusted to 15 μl with 4 × sample buffer comprising 62.5 mM Tris-HCl, 20% glycerol, 0.032% bromophenol blue, 5% mercaptoethanol and 2% SDS. Along with the prestained broad-range molecular mass standards (Bio-Rad), samples were fractionated by a 4% stacking and 10% SDS-PAGE mini gel (Bio-Rad, Mississauga, ON) and transferred to nitrocellulose membrane by electroblotting at 200 mA in a buffer comprising 25 mM Tris-HCl, 192 mM glycine, 0.04% SDS and 20% ethanol. The membranes were rinsed with distilled water, incubated for 1 hour in PBS pH 7.4 with 5% Carnation non-fat milk to block non-specific interactions, and washed five times (twice for 5 min, once for 15 min and twice for 5 min) with PBS containing 0.1% Tween. They were then reacted overnight sequentially in the same buffer at 4°C with 1 to 2 μg/ml anti-human MMP-3 (developed in mouse) and MMP-13 (hinge region, developed in rabbit) antibodies (from Sigma-Aldrich). Subsequently, membranes were washed at 22°C five times with PBS containing 0.1% Tween, incubated with the anti-rabbit or anti-mouse secondary peroxidase-conjugated IgG (300 mU/ml), and washed seven times with PBS containing 0.1% Tween. To reveal the MMP-3 and MMP-13 bands, membranes were incubated with 10 μl of solution A and 990 μl of solution B of the chemiluminescence detection system of Roche Biochemicals (Laval, Québec) and exposed to film for 2 to 15 min.
For Western blots of MAPKs, human femoral-head chondrocytes were pretreated with mithramycin for 30 min and then stimulated with IL-1β for 20 min; total cellular protein extracts (20 μg) in lysis buffer (62.5 mM Tris-HCl pH 6.8, 10% glycerol, 1% Triton X-100, 50 mM dithiothreitol, 2% SDS, 0.01% bromophenol blue) were resolved by 10% SDS-PAGE, transferred to nitrocellulose membranes by electroblotting and incubated overnight at 4°C with primary phosphorylation-state-specific antibodies for phosphorylated extracellular signal-regulated kinase (p-ERK), phospho-p38 and phosphorylated c-Jun N-terminal kinase (p-JNK) (from Cell Signaling Technology Inc., Beverley, MA) at 1:1,000 dilution in 5% BSA, 1 × Tris-buffered saline and 0.1% Tween. Proteins were detected with the enhanced chemiluminescence system from Pharmacia-Amersham. Subsequently, the membranes were stripped with a buffer (containing 100 mM 2-mercaptoethanl, 2% SDS and 62.5 mM Tris-HCl, pH 6.8) at 55°C and reprobed with the antibodies detecting total ERK, p38 and JNK.
Cartilage explants
Human or bovine femoral head cartilage explants were maintained for 1 week in DMEM with 10% FCS, medium was then changed with 0.01% serum-containing DMEM for 3 days until treatments. Explants were treated with mithramycin and IL-1 vehicles as control (water and PBS-0.1% BSA) or exposed to mithramycin (150 nM) and IL-1 (33 ng/ml) for 15 days with replacement of the fresh reagents every 2 days; the secreted media were concentrated by precipitation with 10% trichloroacetic acid and equal amounts of protein (16 μg per lane for human explants and 20 μg per lane for bovine explants) were subjected to Western immunoblotting as described above. All the experiments described in this paper were performed at least two (primary human chondrocytes and cartilage) or three (SW1353 and bovine chondrocytes) times and the results were reproducible.
Results
Mithramycin blocks IL1-stimulated expression of MMP-3 and MMP-13 in human and bovine chondrocyte cell lines
IL-1β potently induced expression of the genes encoding MMP-3 and MMP-13 in the human chondrocytic cell line SW1353 and in primary human femoral head chondrocytes. Mithramycin, a hypocalcemic antibiotic, potently blocked the induction of MMP-3 and MMP-13 mRNA by IL-1β without affecting the control 28S rRNA levels (Fig. 1a,b). The two MMP-13 mRNA bands correspond to transcripts produced by differential use of polyadenylation sites at the 3' end of the gene, the upper band being the longest transcript as reported previously [23]. Induction of MMP-13 protein was also similarly inhibited (Fig. 1a,b). To examine whether mithramycin could also affect MMP gene expression in articular chondrocytes from other species, adult bovine chondrocytes (an important model system in cartilage research) were exposed to different concentrations of mithramycin and then stimulated with IL-1β. This cytokine induced MMP-3 and MMP-13 mRNA expression above basal levels, and pretreatment with mithramycin reduced both constitutive and induced expression in a fashion similar to that of human cells (Fig. 1c). The induction of MMP-13 protein (the main collagen-degrading MMP) by IL-1 was also inhibited. The double MMP-13 protein bands are due to a better resolution of the upper proenzyme and lower active MMP-13 forms.
IL-17-induced MMP gene expression is suppressed by mithramycin
IL-17 is a major proinflammatory cytokine and an inducer of MMP expression in chondrocytes and macrophages [11,24]. As hown in Fig. 2, IL-17 stimulated the basal MMP-3 and MMP-13 mRNA and MMP-13 protein expression in bovine and human OA chondrocytes. Mithramycin dose-dependently diminished these inductions.
TNF-α-induced MMP gene expression is inhibited by mithramycin
TNF-α is another prominent inflammatory cytokine that increased the constitutive MMP-3 and MMP-13 mRNA expression in chondrocytic SW1353 cells, primary human chondrocytes and bovine chondrocytes. Exposure to mithramycin followed by stimulation with TNF-α resulted in decreased constitutive and induced MMP-3 and MMP-13 gene expression (Fig. 3). In some cases, a concentration of 100 nM caused maximal inhibition; a 150 nM dose did not have any additional effect (e. g. mRNA in 3B).
Mithramycin inhibits IL-1-stimulated expression of MMPs in human and bovine cartilage
To study the impact of mithramycin on the production of MMP-13 by chondrocytes in their native cartilage matrix, human and bovine cartilage explants were maintained in low-serum (0.01%) medium and exposed to mithramycin (150 nM) and IL-1β for 15 days with changes of reagents every 2 days. Human cartilage had somewhat elevated constitutive levels of MMP-3 and MMP-13. Mithramycin drastically reduced the secreted basal and IL-1-induced MMP-3 and MMP-13 protein levels in human cartilage (Fig. 4a) and MMP-13 in bovine cartilage (Fig. 4b) as measured by Western blotting of the conditioned media. MMP-3 levels were too low to be measurable in bovine explants. Therefore mithramycin diminishes IL-1-stimulated MMP production in cartilage explants.
Mithramycin does not affect the phosphorylation of ERK, p38 and JNK
Because MAPKs are important mediators of proinflammatory cytokine signal transduction [25], we investigated whether mithramycin affected these signaling cascades. As reported previously [25], TNF-α induced the phosphorylation of the ERK, p38 and JNK subclasses of MAPKs without affecting their total protein levels. Mithramycin did not significantly influence their phosphorylation levels (Fig. 5).
Discussion
We have shown here that mithramycin downregulates basal and proinflammatory cytokine-stimulated MMP-3 and MMP-13 gene expression in chondrocytes and cartilage. This inhibition might be via multiple mechanisms. Sp1 is a ubiquitous transcription factor generally associated with the constitutive expression of genes. However, serum and growth-promoting conditions can stimulate its phosphorylation at specific carboxy-terminal serine residues and can affect the expression of several genes [15,20,26]. Mithramycin is a GC-specific DNA-binding drug, which prevents the binding of Sp1 to its cognate DNA [19]. MMP-3 and MMP-13 induction by the three major inflammatory cytokines and inhibition by mithramycin imply that interference with Sp1 binding might be one of the possible mechanisms. The putative Sp1 site in the MMP-13 promoter [16] might be the target of mithramycin. Because no obvious Sp1-binding site has been found in the MMP-3 promoter [13], the mechanism of MMP-3 inhibition is not known. Suppression by mithramycin might also involve indirect mechanisms. These could include blocking the transcription of other Sp1-responsive MMP regulatory genes such as ets-1, which has Sp1-binding sites in its promoter [27]. Analogously to our results, a requirement for Sp1 activity was demonstrated for the induction of monocyte chemoattractant protein-1 (MCP-1) by TNF-α, and a possible interaction between Sp1 and NF-κB was suggested [28]. Another possibility is that TNF-α-induced c-Jun (a component of AP-1) might superactivate Sp1, and their physical and functional interaction [29] might upregulate MMP promoters. An interaction of Sp1 and c-Jun has also been observed in the gene encoding atrial natriuretic factor [30]. ERK2 was shown to phosphorylate Sp1 [31]. IL-1 can increase the phosphorylation and activity of Sp1 in synovial fibroblasts [32]. However, in our experience, mithramycin had no effect on the IL-1-induced activation of ERK1/2, p38 and JNK MAPKs. Further, a calcium-influx-reducing agent (bis-(o-aminophenoxy)ethane-N, N, N', N' -tetra-acetic acid acetoxymethyl ester (BAPTA-AM)) did not mimic the inhibition of MMP expression by mithramycin (results not shown). Thus, inhibition by mithramycin does not seem to involve MAPKs or a decrease in calcium concentration. Mithramycin might work through the aforementioned mechanisms or by interfering with Sp1/AP-1, ets-1/Sp1 and Sp1/NF-κB interactions, which are important regulators of MMPs. These hypotheses will be tested in future.
The inhibition of MMP gene expression by mithramycin is not unique to this antibiotic. Interestingly, a tetracycline analogue, doxycycline, downregulated the TNF-α-induced expression of MMP-13 RNA in human chondrocytes [33]. Similarly, tetracycline also reduced the IL-1-induced accumulation of stromelysin mRNA [34] as well as that of MMP-1 and MMP-3 in bovine chondrocytes [35]. Subsequent studies revealed that inhibition occurred by decreasing IL-1 and increasing transforming growth factor-β and its receptors, which could downregulate MMP gene expression [36]. It is not known whether mithramycin works through similar mechanisms. Mithramycin also has an interesting property of blocking bone resorption [18], which could be through the suppression of MMP gene expression. Indeed, osteoblast-derived interstitial collagenase initiates bone resorption by the generation of collagen fragments, which in turn activate bone-resorbing osteoclasts [37]. Thus, the ability of mithramycin to block the resorption of bone and cartilage (as implied here) can be advantageous in treating arthritis, in which both tissues are damaged by MMPs. Alternatively, it might work through multiple mechanisms attributed to bisphosphonates, which also prevent cartilage and bone loss and might have utility in treating arthritis [38,39]. Mithramycin is known to have several side effects in patients, including bleeding in the stomach [17], so its benefits in arthritis in vivo are questionable, requiring the development of safer and more specific analogues.
Conclusion
We have shown that the upregulation of MMP-3 and MMP-13 gene expression by IL-1, IL-17 and TNF-α can be inhibited by mithramycin. The mechanisms of inhibition remain to be deciphered but do not seem to involve MAPKs. Multiple mechanisms of action similar to those of bisphosphonate may be operative. It is worth exploring whether this knowledge could lead to the development of novel therapies for blocking tissue damage in arthritis.
Abbreviations
BSA = bovine serum albumin; DMEM = Dulbecco's modified Eagle's medium; ERK = extracellular signal-related kinase; FCS = fetal calf serum; IL = interleukin; JNK = c-Jun N-terminal kinase; MAPK = mitogen-activated protein kinase; MMP = matrix metalloproteinase; OA = osteoarthritis; PBS = phosphate-buffered saline; PCR = polymerase chain reaction; RT = reverse transcriptase; TNF = tumor necrosis factor.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
AL performed most of the tissue culture work and Western blotting experiments. JS conducted several Northern blotting and hybridization experiments. WQL cloned and tested the MMP-13 probe. MZ designed the experimental plan, coordinated the research and drafted the manuscript. All authors read and approved the final manuscript.
Acknowledgements
This work was supported by the Canadian Institutes of Health Research, Arthritis Society and the Canadian Arthritis Network. We thank Dr Faramaze Dehnade, Dr Julio Fernandes and Dr Nicolas Duval for human cartilage, and Ms Anna Chelchowska for preparing the figures.
Figures and Tables
Figure 1 Repression of interleukin (IL)-1β-inducible matrix metalloproteinase (MMP)-3 and MMP-13 RNA expression by mithramycin. Quiescent human chondrosarcoma (a), primary human chondrocytes (b) or bovine chondrocytes (c) were pretreated with different concentrations of mithramycin for 30 min, followed by additional treatment with IL-1β for 24 hours. The MMP-3, MMP-13 and 28S RNA levels were measured by Northern hybridization, and MMP-13 protein levels were measured by Western blot analysis. For protein gels, 3 μg (a) or 4 μg (b, c) of protein was applied to each lane. The resulting autoradiograms indicating the respective gene products are shown.
Figure 2 Decrease in interleukin (IL)-17-inducible matrix metalloproteinase (MMP)-3 and MMP-13 gene expression by mithramycin. Quiescent bovine chondrocytes (a) or primary human chondrocytes (b) were pretreated with different doses of mithramycin for 30 min and treated further with IL-17 for 24 hours. The MMP-3, MMP-13 and 28S RNA levels were measured by Northern hybridization, and MMP-13 protein levels were measured by Western blot analysis. For protein gels, 4 μg of protein was applied to each lane. The resulting autoradiograms indicating the respective gene products are shown.
Figure 3 Downregulation of tumor necrosis factor (TNF)-α-inducible matrix metalloproteinase (MMP)-3 and MMP-13 RNA expression by mithramycin. Human SW1353 condrosarcoma cells (a), primary human femoral head chondrocytes (b) and bovine chondrocytes (c) were pre-exposed to the indicated concentrations of mithramycin for 30 min, followed by additional treatment with TNF-α for 24 hours. The MMP-3, MMP-13 and 28S RNA levels were measured by Northern hybridization, and MMP-3 protein levels were measured by Western blot analysis. For protein gels, 3 μg (a) or 4 μg (b) of protein was applied to each lane. The resulting autoradiograms indicating the respective products are shown.
Figure 4 Downregulation of interleukin (IL)-1β-inducible matrix metalloproteinase (MMP) protein expression by mithramycin in cartilage explants. Human (a) or bovine (b) cartilage explants maintained in DMEM with 0.01% serum were either treated with mithramycin and IL-1β vehicles (water and PBS containing 0.1% BSA) or exposed to mithramycin (150 nM) and IL-1β (33 ng/ml) for 15 days, with renewal of the reagents every 2 days. The secreted media were concentrated by precipitation, and equal amounts of protein (human, 16 μg per lane; bovine, 20 μg per lane) were subjected to Western blotting. The MMP-3 and MMP-13 protein bands are shown.
Figure 5 Impact of mithramycin on interleukin (IL)-1β-induced phosphorylation of mitogen-activated protein kinases. Primary human chondrocytes were pretreated with the indicated doses of mithramycin for 30 min and then stimulated with IL-1β for 20 min. Protein extracts (20 μg per lane) were analyzed by Western blotting with phosphorylation-specific and total antibodies. The resulting bands are shown. ERK, extracellular signal-regulated kinase; JNK, c-Jun N-terminal kinase.
==== Refs
Poole AR An introduction to the pathophysiology of osteoarthritis Front Biosci 1999 4 D662 D670 10525481
Goldring MB The role of the chondrocyte in osteoarthritis Arthritis Rheum 2000 43 1916 1926 11014341 10.1002/1529-0131(200009)43:9<1916::AID-ANR2>3.0.CO;2-I
Nagase H Woessner JF Jr Matrix metalloproteinases J Biol Chem 1999 274 21491 21494 10419448 10.1074/jbc.274.31.21491
Caterson B Flannery CR Hughes CE Little CB Mechanisms involved in cartilage proteoglycan catabolism Matrix Biol 2000 19 333 344 10963994 10.1016/S0945-053X(00)00078-0
Billinghurst RC Dahlberg L Ionescu M Reiner A Bourne R Rorabeck C Mitchell P Hambor J Diekmann O Tschesche H Enhanced cleavage of type II collagen by collagenases in osteoarthritic articular cartilage J Clin Invest 1997 99 1534 1545 9119997
Konttinen YT Ainola M Valleala H Ma J Ida H Mandelin J Kinne RW Santavirta S Sorsa T Lopez-Otin C Analysis of 16 different matrix metalloproteinases (MMP-1 to MMP-20) in the synovial membrane: different profiles in trauma and rheumatoid arthritis Ann Rheum Dis 1999 58 691 697 10531073
Reboul P Pelletier JP Tardif G Cloutier JM Martel-Pelletier J The new collagenase, collagenase-3, is expressed and synthesized by human chondrocytes but not by synoviocytes. A role in osteoarthritis J Clin Invest 1996 97 2011 2019 8621789
Neuhold LA Killar L Zhao W Sung ML Warner L Kulik J Turner J Wu W Billinghurst C Meijers T Postnatal expression in hyaline cartilage of constitutively active human collagenase-3 (MMP-13) induces osteoarthritis in mice J Clin Invest 2001 107 35 44 11134178
Tetlow LC Adlam DJ Woolley DE Matrix metalloproteinase and proinflammatory cytokine production by chondrocytes of human osteoarthritic cartilage: associations with degenerative changes Arthritis Rheum 2001 44 585 594 11263773 10.1002/1529-0131(200103)44:3<585::AID-ANR107>3.0.CO;2-C
Westacott CI Sharif M Cytokines in osteoarthritis: mediators or markers of joint destruction? Semin Arthritis Rheum 1996 25 254 272 8834014 10.1016/S0049-0172(96)80036-9
Shalom-Barak T Quach J Lotz M Interleukin-17-induced gene expression in articular chondrocytes is associated with activation of mitogen-activated protein kinases and NF-κB J Biol Chem 1998 273 27467 27473 9765276 10.1074/jbc.273.42.27467
Bigg HF Rowan AD The inhibition of metalloproteinases as a therapeutic target in rheumatoid arthritis and osteoarthritis Curr Opin Pharmacol 2001 1 314 320 11712757 10.1016/S1471-4892(01)00055-8
Buttice G Quinones S Kurkinen M The AP-1 site is required for basal expression but is not necessary for TPA-response of the human stromelysin gene Nucleic Acids Res 1991 19 3723 3731 1906606
Pendas AM Balbin M Llano E Jimenez MG Lopez-Otin C Structural analysis and promoter characterization of the human collagenase-3 gene (MMP13) Genomics 1997 40 222 233 9119388 10.1006/geno.1996.4554
Black AR Black JD Azizkhan-Clifford J Sp1 and kruppel-like factor family of transcription factors in cell growth regulation and cancer J Cell Physiol 2001 188 143 160 11424081 10.1002/jcp.1111
Wu N Opalenik S Liu J Jansen ED Giro MG Davidson JM Real-time visualization of MMP-13 promoter activity in transgenic mice Matrix Biol 2002 21 149 161 11852231 10.1016/S0945-053X(01)00192-5
Zojer N Keck AV Pecherstorfer M Comparative tolerability of drug therapies for hypercalcaemia of malignancy Drug Saf 1999 21 389 406 10554053
Kiang DT Effect of mithramycin on bone beta-glucuronidase and resorption Calcif Tissue Res 1978 26 209 213 155493
Blume SW Snyder RC Ray R Thomas S Koller CA Miller DM Mithramycin inhibits SP1 binding and selectively inhibits transcriptional activity of the dihydrofolate reductase gene in vitro and in vivo J Clin Invest 1991 88 1613 1621 1834700
Ihn H Tamaki K Increased phosphorylation of transcription factor Sp1 in scleroderma fibroblasts: association with increased expression of the type I collagen gene Arthritis Rheum 2000 43 2240 2247 11037883 10.1002/1529-0131(200010)43:10<2240::AID-ANR11>3.0.CO;2-2
Chomczynski P Sacchi N Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction Anal Biochem 1987 162 156 159 2440339 10.1016/0003-2697(87)90021-2
Shlopov BV Lie WR Mainardi CL Cole AA Chubinskaya S Hasty KA Osteoarthritic lesions: involvement of three different collagenases Arthritis Rheum 1997 40 2065 2074 9365097
Freije JM Diez-Itza I Balbin M Sanchez LM Blasco R Tolivia J Lopez-Otin C Molecular cloning and expression of collagenase-3, a novel human matrix metalloproteinase produced by breast carcinomas J Biol Chem 1994 269 16766 16773 8207000
Jovanovic DV Martel-Pelletier J Di Battista JA Mineau F Jolicoeur FC Benderdour M Pelletier JP Stimulation of 92-kd gelatinase (matrix metalloproteinase 9) production by interleukin-17 in human monocyte/macrophages: a possible role in rheumatoid arthritis Arthritis Rheum 2000 43 1134 1144 10817568 10.1002/1529-0131(200005)43:5<1134::AID-ANR24>3.0.CO;2-#
Liacini A Sylvester J Li WQ Huang W Dehnade F Ahmad M Zafarullah M Induction of matrix metalloproteinase-13 gene expression by TNF-α is mediated by MAP kinases, AP-1, and NF-κB transcription factors in articular chondrocytes Exp Cell Res 2003 288 208 217 12878172 10.1016/S0014-4827(03)00180-0
Black AR Jensen D Lin SY Azizkhan JC Growth/cell cycle regulation of Sp1 phosphorylation J Biol Chem 1999 274 1207 1215 9880488 10.1074/jbc.274.3.1207
Oka T Rairkar A Chen JH Structural and functional analysis of the regulatory sequences of the ets-1 gene Oncogene 1991 6 2077 2083 1945412
Ping D Boekhoudt G Zhang F Morris A Philipsen S Warren ST Boss JM Sp1binding is critical for promoter assembly and activation of the MCP-1 gene by tumor necrosis factor J Biol Chem 2000 275 1708 1714 10636866 10.1074/jbc.275.3.1708
Kardassis D Papakosta P Pardali K Moustakas A c-Jun transactivates the promoter of the human p21(WAF1/Cip1) gene by acting as a superactivator of the ubiquitous transcription factor Sp1 J Biol Chem 1999 274 29572 29581 10506225 10.1074/jbc.274.41.29572
McDonough PM Hanford DS Sprenkle AB Mellon NR Glembotski CC Collaborative roles for c-Jun N-terminal kinase, c-Jun, serum response factor, and Sp1 in calcium-regulated myocardial gene expression J Biol Chem 1997 272 24046 24053 9295358 10.1074/jbc.272.38.24046
Merchant JL Du M Todisco A Sp1 phosphorylation by Erk 2 stimulates DNA binding Biochem Biophys Res Commun 1999 254 454 461 9918860 10.1006/bbrc.1998.9964
Ray A Schatten H Ray BK Activation of Sp1 and its functional co-operation with serum amyloid A-activating sequence binding factor in synoviocyte cells trigger synergistic action of interleukin-1 and interleukin-6 in serum amyloid A gene expression J Biol Chem 1999 274 4300 4308 9933631 10.1074/jbc.274.7.4300
Shlopov BV Smith GN JrCole AA Hasty KA Differential patterns of response to doxycycline and transforming growth factor β1 in the down-regulation of collagenases in osteoarthritic and normal human chondrocytes Arthritis Rheum 1999 42 719 727 10211886 10.1002/1529-0131(199904)42:4<719::AID-ANR15>3.0.CO;2-T
Jonat C Chung FZ Baragi VM Transcriptional downregulation of stromelysin by tetracycline J Cell Biochem 1996 60 341 347 8867810 10.1002/(SICI)1097-4644(19960301)60:3<341::AID-JCB6>3.0.CO;2-W
Sadowski T Steinmeyer J Effects of tetracyclines on the production of matrix metalloproteinases and plasminogen activators as well as of their natural inhibitors, tissue inhibitor of metalloproteinases-1 and plasminogen activator inhibitor-1 Inflamm Res 2001 50 175 182 11339506 10.1007/s000110050742
Shlopov BV Stuart JM Gumanovskaya ML Hasty KA Regulation of cartilage collagenase by doxycycline J Rheumatol 2001 28 835 842 11327259
Holliday LS Welgus HG Fliszar CJ Veith GM Jeffrey JJ Gluck SL Initiation of osteoclast bone resorption by interstitial collagenase J Biol Chem 1997 272 22053 22058 9268345 10.1074/jbc.272.35.22053
Catterall JB Cawston TE Drugs in development: bisphosphonates and metalloproteinase inhibitors Arthritis Res Ther 2003 5 12 24 12716443 10.1186/ar604
Hayami T Pickarski M Wesolowski GA McLane J Bone A Destefano J Rodan GA Duong le T The role of subchondral bone remodeling in osteoarthritis: reduction of cartilage degeneration and prevention of osteophyte formation by alendronate in the rat anterior cruciate ligament transection model Arthritis Rheum 2004 50 1193 1206 15077302 10.1002/art.20124
| 15987479 | PMC1175029 | CC BY | 2021-01-04 16:02:37 | no | Arthritis Res Ther. 2005 Apr 4; 7(4):R777-R783 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1735 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17401598748110.1186/ar1740Research ArticleAcute phase reactants add little to composite disease activity indices for rheumatoid arthritis: validation of a clinical activity score Aletaha Daniel [email protected] Valerie PK 1Stamm Tanja 1Uffmann Martin 3Pflugbeil Stephan 4Machold Klaus 1Smolen Josef S [email protected] Department of Rheumatology, Medical University of Vienna, Vienna, Austria2 National Institute of Arthritis and Musculoskeletal and Skin Diseases, National Institutes of Health, Bethesda, Maryland, USA3 Department of Radiology, Medical University of Vienna, Vienna, Austria4 2nd Department of Medicine, Lainz Hospital, Vienna, Austria2005 7 4 2005 7 4 R796 R806 15 12 2004 7 2 2005 16 2 2005 10 3 2005 Copyright © 2005 Aletaha et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Introduction
Frequent assessments of rheumatoid arthritis (RA) disease activity allow timely adaptation of therapy, which is essential in preventing disease progression. However, values of acute phase reactants (APRs) are needed to calculate current composite activity indices, such as the Disease Activity Score (DAS)28, the DAS28-CRP (i.e. the DAS28 using C-reactive protein instead of erythrocyte sedimentation rate) and the Simplified Disease Activity Index (SDAI). We hypothesized that APRs make limited contribution to the SDAI, and that an SDAI-modification eliminating APRs – termed the Clinical Disease Activity Index (CDAI; i.e. the sum of tender and swollen joint counts [28 joints] and patient and physician global assessments [in cm]) – would have comparable validity in clinical cohorts.
Method
Data sources comprised an observational cohort of 767 RA patients (average disease duration 8.1 ± 10.6 years), and an independent inception cohort of 106 patients (disease duration 11.5 ± 12.5 weeks) who were followed prospectively.
Results
Our clinically based hypothesis was statistically supported: APRs accounted only for 15% of the DAS28, and for 5% of the SDAI and the DAS28-CRP. In both cohorts the CDAI correlated strongly with DAS28 (R = 0.89–0.90) and comparably to the correlation of SDAI with DAS28 (R = 0.90–0.91). In additional analyses, the CDAI when compared to the SDAI and the DAS28 agreed with a weighted kappa of 0.70 and 0.79, respectively, and comparably to the agreement between DAS28 and DAS28-CRP. All three scores correlated similarly with Health Assessment Questionnaire (HAQ) scores (R = 0.45–0.47). The average changes in all scores were greater in patients with better American College of Rheumatology response (P < 0.0001, analysis of variance; discriminant validity). All scores exhibited similar correlations with radiological progression (construct validity) over 3 years (R = 0.54–0.58; P < 0.0001).
Conclusion
APRs add little information on top (and independent) of the combination of clinical variables included in the SDAI. A purely clinical score is a valid measure of disease activity and will have its greatest merits in clinical practice rather than research, where APRs are usually always available. The CDAI may facilitate immediate and consistent treatment decisions and help to improve patient outcomes in the longer term.
==== Body
Introduction
Rheumatoid arthritis (RA) is a progressive inflammatory disease, which causes damage and disability [1-5] that can be prevented by promptly initiated and effective therapy [6-9]. To ensure that therapy is effective, frequent clinical assessments are needed [10-12]. For the purpose of disease activity assessment, valid assessment tools using the well established ACR/EULAR/WHO–ILAR (American College of Rheumatology/European League Against Rheumatism/World Health Organization–International League of Associations for Rheumatology) core set variables of disease activity [13-15] are available, such as the Disease Activity Score (DAS) [16]. Also available are the mathematical modifications to the DAS, namely the DAS28 (based on 28-joint counts) and the DAS28-CRP (i.e. the DAS28 using C-reactive protein [CRP] instead of erythrocyte sedimentation rate [ESR]) [17,18], and the recently introduced Simplified Disease Activity Index (SDAI) [19].
However, these scores are rarely used to follow patients in clinical practice because they either employ extensive joint counts (DAS), their computation requires the use of calculators (DAS, DAS28, DAS28-CRP), or their results are not accessible for immediate decision making at the time of patient–physician interaction because of missing laboratory results (DAS, DAS28, DAS28-CRP and SDAI). Although the inclusion of CRP and ESR is fully justified by their face and content validity, the delay associated with their assessment might be one reason why many physicians do not apply composite scores to guide their clinical decisions.
We hypothesized that an abbreviating modification to the SDAI that omits CRP would be a useful score in clinical practice. Our hypothesis was based on the following factors. First, laboratory test results are frequently missing at patient visits, and thus the long-term benefit of a therapeutic approach that is guided by consistent, regular and immediate assessments of disease activity could be jeopardized. Second, simple scores that can be performed 'on the spot' are more likely to be successfully adopted. Third, the principle of numerical summation has been proven and validated to be equivalent to more complex methods of computation [19-23]. Fourth. acute phase reactants (APRs) correlate with each of the other core set variables, especially those employed in the composite indices, suggesting that they may not add importantly to a composite score [24]. Finally, the ACR response criteria consist of an invariable part (joint counts) and a variable part [25], the latter of which employs the APR as one of five measures. Because only three of these measures need to change by more than 20%, the APR is not necessarily required to assess changes in disease activity according to the ACR response criteria; nevertheless, the ACR response criteria agree well with the DAS28 and the SDAI response in data from clinical trials [19,26].
In the present study we established that our initial hypothesis was valid by showing that the contributions made by CRP and ESR to various composite scores are low. We subsequently assessed the correlational, discriminant, and construct validity of a clinical activity index omitting APR in comparison with established scores.
Method
Datasets
One source of data employed was a large observational cohort of RA outpatients, who were seen on a regular basis, usually every 3 months. At each visit clinical, functional and laboratory core set variables [18-20] and disease activity according to the composite scores DAS28 and SDAI were documented. Clinical assessments including joint counts were performed by independent, trained assessors who were not involved in treatment decisions. In July 2004, data on 998 patients followed in our clinics had been entered into the database. Each patient's first visit with complete documentation of clinical data was included to assemble the 'routine' cohort. There were 767 patients with at least one complete observation, and the first of these complete observations was used for the analyses. Of all 5070 patient observations that were initially documented, 2564 (50.6%) had missing data. Among these incomplete observations, 45% (n = 1150) had missing ESR and/or CRP values.
The second source of data was an independent cohort of newly diagnosed RA patients ('inception' cohort), whose visits were documented in the same manner as described above but starting from their first presentation to the clinic. The referral pattern and detailed follow up of these patients were described elsewhere [9,27]. Radiographs of the hands and feet were obtained every 1–2 years, and were scored using the Larsen method [28] by a team of two experienced readers; they were presented to the readers in chronological order. Reassessment of a random subgroup of 40 radiographs of hands and feet revealed good agreement (R = 0.86, 95% confidence interval [CI] 0.81–0.91). All patients in the inception cohort received disease-modifying antirheumatic drugs, such as methotrexate, as soon as the diagnosis was made, with a few exceptions in patients who refused to take such therapy immediately.
The demographic and disease activity characteristics of patients in both cohorts are summarized in Table 1. Because several baseline variables were not normally distributed (see below), we present the median along with the first and third quartiles as robust descriptive measures.
Distribution of study variables and appropriateness of test statistics
Whenever variables were normally distributed, as assessed using the Kolmogorov–Smirnov test, we performed parametric test statistics (such as Pearson correlation, or one-way analysis of variance [ANOVA]). In several cases, skewed distributions required the use of nonparametric tests (such as Spearman rank correlation). However, the exploratory analysis on the contribution of APRs to the various composite scores was based on a linear regression model despite non-normal distributions of several variables, given the large numbers of observations in the routine cohort (n = 767; Table 1), which is sufficient to invoke the central limit theorem.
Analysis of the contributions of acute phase reactants to current composite scores
Calculations of the DAS28 and SDAI are based on the following: numbers of swollen and tender joints (swollen joint count [SJC] and tender joint count [TJC]), employing the 28 joint count; evaluator's and/or patient's global assessment of disease activity (EGA, PGA); and CRP or ESR. The following formulae are the basis for their calculation [16,19]:
DAS28 = (0.56 × TJC1/2) + (0.28 × SJC1/2) + (0.7 × ln [ESR]) + (0.014 × PGA [in mm])
SDAI = SJC + TJC + PGA (visual–analogue scale [VAS; in cm]) + EGA (VAS [in cm]) + CRP (in mg/dl)
In addition, we calculated a version of the DAS28 that, like the SDAI, employs CRP rather than ESR, and is obtained as follows [18]:
DAS28-CRP = (0.56 × TJC1/2) + (0.28 × SJC1/2) + (0.36 × ln [CRP; in mg/l])+1) + (0.014 × PGA [in mm]) + 0.96
To determine whether our clinical hypothesis that CRP makes a small contribution to the SDAI would withstand statistical analysis, we first evaluated the contributions made by individual component variables to the SDAI. We constructed a perfect fit regression model to predict the score by its items, using cross-sectional patient observations from the routine dataset (n = 767). For each variable contained in the SDAI, we determined the contribution made to the SDAI (R2) when the variable was introduced as first (zero order) or last (final) variable in the model. In addition, we determined each item's colinearity, presented as the proportion of its variance that was explained by the other items in the score, which equals the term (1 - tolerance) × 100. The three parameters (zero order and final model contributions, and colinearity diagnostics) provide overlapping information, and were used to assess the statistical characteristics of CRP in the SDAI.
We then followed an analogous sequence of analyses to determine model contribution and colinearity of individual component variables to the DAS28 and the DAS28-CRP. To allow construction of the perfect fit model, introduction of items into the regression model employed transformed values according to the respective formulae (SJC and TJC as square roots, and ESR and CRP as their natural logs).
Clinical activity index and assessment of comparative validity
Next, we calculated the Clinical Disease Activity Index (CDAI) as follows:
CDAI = SJC + TJC + PGA (in cm) + EGA (in cm)
We determined several aspects of validity of the CDAI [29]: correlational validity refers to comparison with other measures of disease activity; discriminant validity in this setting relates to the correlation of changes in the scale with changes in other measures of disease activity; and construct validity considers correlation with important outcomes of the disease, such as radiological progression.
Correlational validity
Correlational validity between CDAI, DAS28 and SDAI was assessed in patients from the routine cohort (n = 767) using Spearman's rank statistics. In addition, we calculated 95% CIs using Fisher's approximation. Next, we used the Health Assessment Questionnaire Disability Index (HAQ) score as an additional comparator in the correlation analysis with these indices (n = 720). As a functional measurement, the HAQ is determined by accumulated joint damage but also by disease activity [30-32]. Moreover, the HAQ is an independent comparator that does not include joint counts, global assessments, or APRs, in contrast to composite scores, which are all based on similar sets of variables. We then validated these results in an independent group of patients at their first presentation using the inception cohort (n = 105). In this manner, the results from a cohort with, on average, moderate disease activity were validated in another one with high disease activity (Table 1).
In addition to the presented correlation coefficients, we sought to determine the agreement of the different scores in individual patients. We therefore created 10 patient groups of equal size based on the patients' DAS28 ranks within the cohort. The groups were ordered (i.e. the first group comprised the 10% of patients with the lowest DAS28, and the last group comprised the 10% with the highest DAS28 values). Then, we grouped the patients in the same way based on their CDAI, SDAI and DAS28-CRP ranks. Based on these groups, we used weighted kappa statistics to assess agreement of different scores on individual patients.
Discriminant validity
For the assessment of discriminant validity we characterized patients by their degree of improvement according to the ACR response criteria within 1 year after entering the inception cohort (n = 91 with complete baseline and 1 year data). We divided ACR responses into three groups: lack of response (<20% improvement by ACR response criteria), major response (≥ 70% improvement), or moderate response (≥ 20% but <70% improvement). Using one-way ANOVA, we analyzed whether changes in the various continuous scores were greater in higher ACR responder groups, and whether these differences were statistically significant at the group level. We then used post hoc t-tests with Bonferroni adjustment to determine which groups were statistically different in pairwise comparisons. Also, effect sizes for each group were calculated as changes in scores divided by their baseline standard deviations [33].
In addition to the comparison with ACR response, we correlated changes in the continuous scores with respective changes in HAQ scores during the first year of disease (n = 91), using Spearman rank correlation and Fisher's approximation as described above. However, in this early cohort, HAQ score mainly reflects disease activity rather than being a measure of functional outcome (i.e. a surrogate for construct validity).
Construct validity
Radiographic data were available for the majority of the 80 patients in the inception cohort who were followed for 3 years (n = 56), which constituted a clinically meaningful time frame in which to detect major changes in damage. We performed linear (Pearson) correlation of time averaged disease activity (equivalent to area under the curve) for DAS28, SDAI and CDAI with changes in Larsen score over 3 years. Again, we calculated 95% CIs as above. For simplicity, we did not employ more sophisticated methods, such as longitudinal modelling (e.g. by generalized estimating equations), in this validation analysis.
Results
Contribution of acute phase reactants to composite scores
Figure 1 depicts the results from the perfect fit regression models. Items are ordered according to their contribution when introduced into the model as first variable (zero-order R2 contribution; dark bars); the independent variables that best accounted for the SDAI (Fig. 1a) were TJC (R2 = 65.1%) and EGA (R2 = 63.4%), and the variable with the smallest R2 was CRP (21.5%). When individual variables were introduced as last items into the model (final R2; grey bars), the contribution was least for EGA (0.7%) and PGA (1.8%), according to their colinearity (>50% for each). The final R2 for CRP was only 5.1%, despite the practical absence of colinearity (7.7%). These analyses indicate that CRP was adding independent information to the score (low colinearity), but that its changes were not likely to be substantially reflected in changes in the SDAI (low model contribution).
The results of analogous analyses for the DAS28 and its items are shown in Fig. 1b. Similar to the results of the SDAI analysis, ESR (as the APR) made the smallest independent contribution to the DAS28 (R2 = 34.0%), and the smallest colinearity (6.6%), although the final contribution was somewhat higher (14.8%). As in the SDAI, there was a significant level of colinearity of residual items (white bars), but to a somewhat lesser degree.
To determine whether the difference in final contributions between ESR in the DAS28 and CRP in the SDAI (14.8% versus 5.1%), given similar degrees of colinearity (6.6% versus 7.7%), was score related (i.e. DAS28 versus SDAI) or item related (i.e. ESR versus CRP), we analyzed the newly proposed modification to the DAS28 [18], which includes CRP instead of ESR but otherwise identical variables (DAS28-CRP; Fig. 1c). Here, despite the differences in construction of and component weighing in the two scores, the contributions made by CRP (zero order 24.5%, final 4.8%) reached similar levels to CRP in the SDAI (21.5% and 5.1%, respectively). The low colinearity of APRs in all three scores indicates that they provide information distinct from the clinical measures. However, the low model contribution of CRP (about 5%) indicates that only a very small proportion of variance in the respective indices remains unexplained without CRP, which is in accord with the small numerical value of CRP in the SDAI and the DAS28-CRP. Likewise, ESR made a relatively low model contribution to the DAS28, which is line with a significant correlation between these two APRs in the studied cohort (R = 0.63; P < 0.001). hypothesized that APRs make limited contribution to the SDAI
Because our initial hypothesis – that CRP makes a limited contribution to the SDAI, and that excluding CRP from the SDAI will yield a simple and immediately calculable score – was supported by these statistical analyses, we next validated the CDAI using the cross-sectional 'routine' cohort and the independent, longitudinal inception cohort of patients with RA. The quartiles and ranges for the CDAI and for all other mentioned scores are shown in Table 2 for both patient cohorts.
Cross-sectional correlation and validation of composite scores and Health Assessment Questionnaire disability index
We next analyzed the correlation between the DAS28, SDAI and CDAI, as well as the correlation between these scores and the HAQ disability index in the routine cohort, which revealed similar correlation coefficients for CDAI and SDAI when compared with DAS28 (Fig. 2, upper diagonal half; n = 767). This correlation was fully validated by virtually identical coefficients obtained in the analysis of the inception cohort (Fig. 2, lower diagonal half), in which patients had higher disease activity. Likewise, Spearman rank correlations with the HAQ revealed comparable results for DAS28, SDAI and CDAI within each of the patient cohorts. The comparable correlation coefficients obtained for all three scores in two independent cohorts strengthens the results obtained from the cross-sectional analysis, because they were not influenced by the level of disease activity, the patients' disease duration, or treatment status, which were all different between patients in the routine and those in the inception cohort. Although there were differences in the degree of correlation with the HAQ between the two cohorts, this pertained to all three disease activity scores in a similar manner.
In a further analysis, based on the cohort ranks of each patient's DAS28, DAS28-CRP, SDAI and CDAI values, we divided the patients into 10 ordered groups for each of the four scores (from the group comprising the 10% of patients with the lowest activity to that consisting of the 10% with the highest activity, by respective score). We then analyzed the agreement of these categorizations between scores using weighted kappa statistics [34]. Kappa values range from 0 (agreement as expected by chance) to 1 (maximum possible agreement beyond chance). For this analysis of individual patient allocation into the different groups, there was good agreement of the CDAI with the DAS28-CRP and the DAS28 (κ = 0.79 and 0.70, respectively). The results were similar when the DAS28 and its derivative, the DAS28-CRP, were compared (κ = 0.80). Not surprisingly, there was excellent agreement between CDAI and SDAI (κ = 0.89).
Changes in composite scores in relation to American College of Rheumatology response and to changes in Health Assessment Questionnaire scores
In the inception cohort, ACR20 responses were achieved by 69% of patients at the end of the first year, ACR50 by 59%, and ACR70 by 47%. To allow comparison of changes in composite scores in individuals with ACR responses, we grouped patients' improvements into the following categories: non-response (ACR response <20%; n = 28, 30.8%), moderate response (20–69% improvement; n = 20, 22.0%) and major response (≥ 70% improvement; n = 43, 47.3%). The high rate of ACR70 responders in this clinic cohort treated with traditional disease-modifying antirheumatic drugs is in accordance with previous observations in similar patient cohorts [12,35]. At the group level, score responses of the DAS28 (Fig. 3a), SDAI (Fig. 3b) and CDAI (Fig. 3c) increased with respect to the ACR response categories (P < 0.0001, one-way ANOVA). Post hoc Bonferroni-adjusted pairwise t-tests revealed significant differences for the comparison of the ACR ≥ 70% responders with the other groups (P < 0.0001). The ACR <20 and ACR 20–69 groups were statistically different only in the CDAI analysis (P = 0.032; Fig. 3c). These findings indicate that, at the group level, the DAS28, SDAI and CDAI were sensitive in discriminating between different response categories. This is further supported by calculating the effect size for the three scores after 1 year of observation: for the DAS28 the effect size in the ACR20–69 responders was 2.4 times higher than in the ACR nonresponders; likewise, the effect size of the ACR70 responders was 4.4 times that in the nonresponders. The same analyses revealed an effect size increases of 2.7-fold and 4.1-fold, respectively, for the SDAI, and of 3.3-fold and 6.5-fold, respectively, for the CDAI. Thus, using effect size calculations, all three scores discriminated various degrees of ACR responsiveness from ACR nonresponsiveness to a similar extent. Also, the 1 year changes in the HAQ in these 91 patients were similarly correlated with the 1 year changes in all three scores: for DAS28, R = 0.32 (95% CI 0.12–0.49; P = 0.001); for SDAI, R = 0.38 (95% CI 0.19–0.54; P < 0.001); and for CDAI, R = 0.39 (95% CI 0.20–0.55; P < 0.001).
Radiological outcome
To compare construct validity between the composite scores, we performed a linear correlation analysis between time-averaged DAS28, SDAI, CDAI and changes in Larsen scores over 3 years (n = 56). The R coefficients were 0.58 (95% CI 0.37–0.73), 0.59 (95% CI 0.39–0.74) and 0.54 (95% CI 0.32–0.70), respectively. All correlations were significant (P < 0.0001). Figure 4a–c permits visual judgement of this relationship for each score, and a line of best fit has been added based on the given observations. Moreover, there was significant correlation between time integrated CRP with changes in Larsen scores (Fig. 4d), as was previously reported by others [36-38].
Discussion
In this study we showed that the CDAI, a simple composite index obtained by numerical summation of four solely clinical variables, is a valid instrument with which to follow patients with RA. Our hypothesis was originally based on feasibility arguments, namely the frequent lack of immediate access to laboratory results in the clinic, but was further strengthened by statistical arguments related to the low contribution made by the acute phase response to the composite scores. In fact, all data obtained support our clinically derived hypothesis that APRs provide little information on actual disease activity on top of that provided by the combination of several clinical components. This was the case for all analyzed RA activity scores, despite the differences in their construction and component weighing.
For many rheumatologists, this lack of additional information provided by APR may be intriguing because CRP and ESR are among the most commonly used laboratory tests in the evaluation of RA disease activity [39], and their importance as surrogates of the disease process, as well as predictors of disease outcome, are well recognized and irrefutable [36-38]. However, APRs did not seem to contribute information to composite scores that was sufficiently important to change judgement of disease activity, in addition to merely using clinical measures. In fact, when we divided all patients into 10 groups based on their disease activity ranks within the cohort, as measured using the different scores, we found statistical agreement that was indicative of high clinical conformity of classifications by different scores. All of these findings indicate that content validity of the CDAI is well maintained despite the absence of CRP as a component.
In accordance with these notions is the observation that as much as 85% of the variance in the DAS28 was explained without ESR; 95% of the variances in the SDAI and the DAS28-CRP were explained by their composing clinical variables (i.e. without CRP). The similarity in these results between the DAS28-CRP and the SDAI further supports previous indications that transformation and/or weighing of the clinical variables does not confer an advantage compared with their simple numerical summation [21-23,40]. However, it should be borne in mind that the DAS28-CRP has only recently been made public and must be regarded with caution until it has been more widely studied; in fact, the present investigation may represent the first validation of the DAS28-CRP. Interestingly, our analyses reveal a high degree of colinearity between the two global assessments employed in the SDAI and CDAI. Because both patient and physician global assessment are parts of the widely applied and validated ACR/EULAR/WHO-ILAR core set variables of RA disease activity assessment, it would not be intuitive to eliminate any one of them, especially as, in contrast to the APRs, they do not correlate with structural damage independently. In addition, because these two variables are usually assessed jointly, the elimination of any one of them would not increase the feasibility of calculating the score.
In a cross-sectional analysis of a large number of patients, the CDAI not only had correlational validity compared with the SDAI, from which it was derived, but also compared with the DAS28 and the DAS28-CRP. Also, there was no difference beyond chance in the correlation of CDAI with the HAQ as compared with the respective correlations of SDAI and DAS28 with the HAQ. This finding is especially noteworthy because the HAQ is a functional measure, which is not based on or constructed with core set elements used in the DAS28 or SDAI. Moreover, when related to different degrees of ACR response, the results obtained using CDAI were graded with statistically significant and clinically meaningful differences between all group means, and were very similar to those seen for the respective DAS28 and SDAI groups. Also, for the CDAI, effect sizes appeared to be even more graded between the different ACR responder groups.
Thus, although none of the comparators in this study represents a 'gold standard' for disease activity measurement, the validity of the new score was proven not only with respect to other composite scores but also with respect to the HAQ, which is a completely distinct construct. In addition, the CDAI was shown to have very good agreement with other composite indices on the categorization of individual patients, which is an important aspect in the clinical use of this score. Furthermore, all mentioned correlation analyses were successfully validated in a second, completely independent cohort of newly diagnosed patients with RA who overall had a higher level of disease activity and were untreated at baseline. The different characteristics of the two cohorts, and the similar correlation coefficients for the three indices obtained within each cohort indicate that the application of our findings might not be confined to patient cohorts with particular characteristics, such as disease duration or disease activity.
A limitation of the CDAI is that many physicians do not perform detailed joint counts in the assessment of RA disease activity [38]. On the other hand, joint counts are also required for other composite disease activity scores, and the CDAI allows elimination of at least one variable that is frequently missed at patient visits – the APR. Although a considerable number of measurements was missing in the overall source dataset, these missing data were random. This was also evident from the similar clinical characteristics of patients with and without available APR measurements. Therefore, and given the large number of complete patient observations, an unbiased analysis was assured. Like for the DAS28 [41], a possible criticism of the CDAI is that it does not include assessment of joints in the feet; however, in the course of proving the reliability of the 28 joint count [42,43], it was found that this reduced joint count reflects overall joint involvement very well and that, in the presence of low joint counts, the joints of the feet rarely add a significant number of additionally involved joints – a finding that we have also observed in our database (data not shown). It might also be regarded as a further limitation that the CDAI was not developed by factor and/or discriminant analysis of individual variables. However, the value of all core set variables has been shown repeatedly [10-12] and their responsiveness has likewise been demonstrated [26]. In addition, there are several conceptual and methodological advantages of composite scores compared with individual items [29]. Moreover, the SDAI, from which the CDAI was derived, has also been validated and shown to have practicability, discriminant capacity and sensitivity to change in several studies [21-23,37]. Likewise, as a composite score, the test–retest reliability of the CDAI is based on the reliability of its individual components, which, although not assessed here, has proven to be good.
Despite omission of the APR from the formula, the CDAI maintained a clinically important ability to predict outcome, measured as radiological progression over 3 years. This stability of construct validity across the scores is therefore also in accord with our initial hypothesis. Interestingly, the deletion of the APR from the SDAI did not change the correlation of the score with radiographic progression; there was a similar degree of correlation with radiographic changes whether DAS28 (using ESR), SDAI (using CRP), or CDAI (using no APR) were employed. Furthermore, the observation that the APR alone also was associated with radiographic progression is not only in accordance with previous reports [36-38] but also suggests an independent relationship with structural damage of both clinical variables (as reflected by the composite CDAI score) and APR. We did not use HAQ scores as an outcome measure because in this early cohort the links between damage and function are expected to be small [31,44]. HAQ scores in this cohort would therefore be a surrogate of disease activity, rather than an independent measure of irreversible loss of function [30,31].
Our introduction of the CDAI was not intended to suggest that the acute phase response does not represent an important measure in the follow up of RA, or that it should be deleted from existing indices such as the DAS28 and the SDAI. In particular, the ESR contributes 15% to the DAS28 composition, which is not an irrelevant amount of information. However, the validity of the CDAI, as revealed here by multiple statistical analyses in two different cohorts, shows that the APR is not an absolute requirement in the context of disease activity scores. In fact, we would urge physicians to continue to obtain an APR measure regularly during follow up because, like the CDAI, it reflects disease activity and correlates with long-term outcome. However, as stated above, the APR can be employed as an independent measure as well as being a part of a composite index.
Because calculation of the SDAI (and of the DAS28) is frequently limited at the time of the patient's visit either by a wait for laboratory results or their unavailability, omitting the APR from the score allows unlimited and immediate assessment of disease activity by including only variables that are available by physical examination and patient questioning at the time of interaction with the patient. Therapeutic decisions will then be possible without further delay. Of course, clinic settings can be revised to have laboratory results delivered at the time of patient visits, although this may not be easy in all situations, and in reality is often not the case. Thus, using a purely clinical score facilitates consistent patient assessment, which might be more attractive for routine application to many physicians, who currently base their treatment decisions on more general and subjective impressions rather than on standardized assessments. The fact that joint counts are frequently not assessed routinely does not diminish these notions; deletion of joint counts from composite scores cannot be justified for a disease of the joints, and joint counts can always be performed at the time of patient visit to the clinic by the physician or another assessor. Moreover, in this age of expensive therapies, consistent assessment of disease activity might soon become compulsory from the payer's perspective. Thus, the ability to adopt a simple but valid score will potentially have great implications with respect to implementation of new therapeutic concepts. At the same time, less frequent laboratory investigations do not appear to impair the physician's ability to detect adverse treatment effects, but can reduce the overall costs of care considerably [45-47].
Of course, further validation of the CDAI will be required to fully confirm its value. Such additional investigations should include analyses of construct validity with regard to radiographic damage and predictive value with regard to long-term functional outcome in larger cohorts of patients. In addition, cutoffs for disease activity categories, including remission, as well as changes that reflect important responses must be determined. Such analyses are currently underway.
Conclusion
Our findings indicate that the CDAI – a composite score that employs only clinical variables and omits assessment of an APR – has similar validity to other currently employed composite indices for following patients with RA. Also, using numerical summation, this score is very easy to calculate. For these reasons, the CDAI should facilitate decision making by physicians and avoid lags in efficient treatment adaptation for patients with RA. According to current knowledge, such intensified and prompt patient care can be expected to reduce the individual [12,48] and socioeconomic impact of the disease in the longer term.
Abbreviations
ACR = American College of Rheumatology; ANOVA = analysis of variance; APR = acute phase reactant; CDAI = Clinical Disease Activity Index; CI = confidence interval; CRP = C-reactive protein; DAS = Disease Activity Score; ESR = erythrocyte sedimentation rate; EULAR = European League Against Rheumatism; HAQ = Health Assessment Questionnaire Disability Index; EGA = evaluator global assessment; PGA= patient global assessment; RA = rheumatoid arthritis; SDAI = Simplified Disease Activity Index; SJC = swollen joint count; TJC = tender joint count; VAS = visual–analogue scale (100 mm); WHO–ILAR = World Health Organization–International League of Associations for Rheumatology.
Acknowledgements
We thank Dr Michael Ward for his thoughtful comments on the manuscript.
Figures and Tables
Figure 1 Contribution of individual variables to composite scores. Explanation of score variability for (a) the Simplified Disease Activity Index (SDAI), (b) the Disease Activity Score (DAS)28, and (c) the DAS28-CRP for the respective clinical and acute phase reactant (APR) variables, at zero-order (i.e. R2 if the variable was introduced as the first one; black bars) or finally (i.e. R2 if variable was introduced in the model as the last one; grey bars), and item colinearity within the respective composite index (1 - tolerance, expressed as percentage; white bars; n = 767). CRP, C-reactive protein; ESR, erythrocyte sedimentation rate; PGA/EGA, patient/evaluator global assessment of disease activity (100 mm visual analogue scale); TJC/SJC, tender/swollen joint count (28 joints).
Figure 2 Cross-sectional correlation of composite scores and correlation with HAQ scores. Matrix displaying Spearman rank coefficients (95% confidence intervals) for cross-sectional correlations of Disease Activity Score (DAS)28, Simplified Disease Activity Index (SDAI), Clinical Disease Activity Index (CDAI), and Health Assessment Questionnaire (HAQ) in the routine cohort (upper diagonal half; n = 720 for correlations with HAQ, otherwise n = 767) and the inception cohort (lower diagonal half; n = 104 for correlation with HAQ, otherwise n = 105).
Figure 3 Changes in composite scores in relation to ACR response. Changes in (a) Disease Activity Score (DAS)28, (b) Simplified Disease Activity Index (SDAI) and (c) Clinical Disease Activity Index (CDAI) in relation to the achieved American College of Rheumatology (ACR) response of 91 patients in the inception cohort. ACR ranges were defined as ACR <20 (n = 28, 30.8%), ACR 20–69 (n = 20, 22.0%) and ACR ≥ 70 (n = 43, 47.2%), allowing analysis of independent observations. Error bars span the 95% confidence interval of the mean. Differences in group changes were statistically significant for all three scores (P < 0.0001, one-way analysis of variance). Presented P values for post hoc pairwise group comparisons are subjected to Bonferroni adjustment. *P < 0.0001 for ≥ ACR70 group compared with other groups.
Figure 4 Association of composite scores with radiological outcome. Correlation with changes in Larsen scores within 3 years from entering the inception cohort (n = 56) of time-averaged (a) Disease Activity Score (DAS)28 (R = 0.58, 95% confidence interval [CI] 0.37–0.73), (b) Simplified Disease Activity Index (SDAI; R = 0.59, 95% CI 0.39–0.74), and (c) Clinical Disease Activity Index (CDAI; R = 0.54, 95% CI 0.32–0.70). All correlations are significant (P < 0.0001). (d) C-rectaive protein (CRP; R = 0.28, 95% CI 0.02 to 0.51; P = 0.025).
Table 1 Characteristics of patients in routine and inception cohorts
Characteristic Routine cohort (cross-sectional) Inception cohort (longitudinal)
Patients (n) 767 106
Age (years; mean ± SD) 54.1 ± 14.9 50.5 ± 15.6
Sex (% female) 79.9 75.2
Rheumatoid factor (% positive) 55.3 78.1
Disease duration at baseline (mean ± SD) 8.1 ± 10.6 years 11.5 ± 12.5 weeks
Duration of follow up (years; mean ± SD; range) - 3.2 ± 1.3; 1–7.25
Disease activity characteristics (median [1st;3rd quartile]) At cross-section At baseline
Swollen joint count (0–28) 3 (1;7) 7 (4;13)
Tender joint count (0–28) 2 (0;6) 8 (3;16)
ESR (mm; normal <20) 23 (14;55) 49 (24;70)
CRP (mg/dl; normal <1.0) 1.1 (0.5;2.7) 5.1 (1.9;17.0)
Patient assessment of pain (mm; 0–100) 37 (19;53) 50 (32;66)
Patient global assessment of activity (mm; 0–100) 37 (18;58) 51 (33;66)
Evaluator global assessment of activity (mm; 0–100) 34 (19;49) 44 (31;58)
HAQ (0–3) 0.875 (0.25;1.5) 0.75 (0;1.5)
Larsen score - 1 (0;7)
Completeness of data for analysis
Cross-sectional correlation between composite indices (n [%]) 767/767 (100)a 105/106 (99.1%)b
Cross-sectional correlation with HAQ scores (n [%]) 720/767 (93.9) 104/106 (98.1)b
Discriminant validity, 1-year follow up (n [%]) - 91/100 (91.0%)
Construct validity, 3-year follow upc (n [%]) - 56/80 (70.0%)
aCompleteness of data was the prerequisite for inclusion. bUsed to validate the results from the cross-sectional analyses in the routine cohort. cIncluding complete radiological data. CRP, C-reactive protein; ESR, erythrocyte sedimentation rate; HAQ, Health Assessment Questionnaire; SD, standard deviation.
Table 2 Values of composite indices in the two cohorts.
Composite scores (rangea) Routine cohort (n = 767) Inception cohort (n = 105)
Median 1st;3rd Quartile Range Median 1st;3rd Quartile Range
DAS28 (0.5–9.1) 4.09 2.99;5.17 0.50–8.56 5.62 4.81;6.44 2.84–8.28
DAS28-CRP (1.0–8.5) 3.78 2.71;4.82 1.60–8.28 4.67 4.04;5.50 2.35–7.42
SDAI (0–86) 16.7 8.1;26.7 0.5–78.9 29.0 20.1;41.6 7.5–77.0
CDAI (0–76) 14.8 6.5;23.3 0–67.8 25.6 17.1;37.9 6.3–70.2
aMaximum possible ranges of acute phase reactants assumed: 5–100 mm for erythrocyte sedimentation rate; 0–10 mg/dl for C-reactive protein (CRP). CDAI, Clinical Disease Activity Index; DAS, Disease Activity Score; SDAI, Simplified Disease Activity Index.
==== Refs
Plant MJ Williams AL O'Sullivan MM Lewis PA Coles EC Jessop JD Relationship between time-integrated C-reactive protein levels and radiologic progression in patients with rheumatoid arthritis Arthritis Rheum 2000 43 1473 1477 10902748 10.1002/1529-0131(200007)43:7<1473::AID-ANR9>3.0.CO;2-N
Welsing PM Landewe RB van Riel PL Boers M van Gestel AM van der Linden S Swinkels HL van der Heijde DM The relationship between disease activity and radiologic progression in patients with rheumatoid arthritis: a longitudinal analysis Arthritis Rheum 2004 2082 2093 15248205 10.1002/art.20350
Fuchs HA Kaye JJ Callahan LF Nance EP Pincus T Evidence of significant radiographic damage in rheumatoid arthritis within the first 2 years of disease J Rheumatol 1989 16 585 591 2754663
Lee DM Weinblatt ME Rheumatoid arthritis Lancet 2001 358 903 911 11567728 10.1016/S0140-6736(01)06075-5
Scott DL Symmons DP Coulton BL Popert AJ Long-term outcome of treating rheumatoid arthritis: results after 20 years Lancet 1987 1 1108 1111 2883443 10.1016/S0140-6736(87)91672-2
Klareskog L van der Heijde D de Jager JP Gough A Kalden J Malaise M Martin Mola E Pavelka K Sany J Settas L Therapeutic effect of the combination of etanercept and methotrexate compared with each treatment alone in patients with rheumatoid arthritis: double-blind randomised controlled trial Lancet 2004 363 675 681 15001324 10.1016/S0140-6736(04)15640-7
Boers M Verhoeven AC Markusse HM van de Laar MA Westhovens R van Denderen JC van Zeben D Dijkmans BA Peeters AJ Jacobs P van den Brink HR Schouten HJ Randomised comparison of combined step-down prednisolone, methotrexate and sulphasalazine with sulphasalazine alone in early rheumatoid arthritis Lancet 1997 350 309 318 9251634 10.1016/S0140-6736(97)01300-7
Lard LR Visser H Speyer I vander Horst-Bruinsma IE Zwinderman AH Breedveld FC Hazes JM Early versus delayed treatment in patients with recent-onset rheumatoid arthritis: comparison of two cohorts who received different treatment strategies Am J Med 2001 111 446 451 11690569 10.1016/S0002-9343(01)00872-5
Nell VP Machold KP Eberl G Stamm TA Uffmann M Smolen JS Benefit of very early referral and very early therapy with disease-modifying anti-rheumatic drugs in patients with early rheumatoid arthritis Rheumatology (Oxford) 2004 43 906 914 15113999
Ward MM Lubeck D Leigh JP Longterm health outcomes of patients with rheumatoid arthritis treated in managed care and fee-for-service practice settings J Rheumatol 1998 25 641 649 9558163
Smolen JS Sokka T Pincus T Breedveld FC A proposed treatment algorithm for rheumatoid arthritis: aggressive therapy, methotrexate, and quantitative measures Clin Exp Rheumatol 2003 S209 S210 14969080
Grigor C Capell H Stirling A McMahon AD Lock P Vallance R Kincaid W Porter D Effect of a treatment strategy of tight control for rheumatoid arthritis (the TICORA study): a single-blind randomised controlled trial Lancet 2004 364 263 269 15262104 10.1016/S0140-6736(04)16676-2
Felson DT Anderson JJ Boers M Bombardier C Chernoff M Fried B Furst D Goldsmith C Kieszak S Lightfoot R The American College of Rheumatology preliminary core set of disease activity measures for rheumatoid arthritis clinical trials. The Committee on Outcome Measures in Rheumatoid Arthritis Clinical Trials Arthritis Rheum 1993 36 729 740 8507213
Smolen JS The work of the EULAR Standing Committee on International Clinical Studies Including Therapeutic Trials (ESCISIT) Br J Rheumatol 1992 31 219 220 1555036
Boers M Tugwell P Felson DT van Riel PL Kirwan JR Edmonds JP Smolen JS Khaltaev N Muirden KD World Health Organization and International League of Associations for Rheumatology core endpoints for symptom modifying antirheumatic drugs in rheumatoid arthritis clinical trials J Rheumatol 1994 86 89
van der Heijde DM van 't Hof MA van Riel PL Theunisse LA Lubberts EW van Leeuwen MA van Rijswijk MH van de Putte LB Judging disease activity in clinical practice in rheumatoid arthritis: first step in the development of a disease activity score Ann Rheum Dis 1990 49 916 920 2256738
Prevoo ML 't Hof MA Kuper HH van Leeuwen MA van de Putte LB van Riel PL Modified disease activity scores that include twenty-eight-joint counts. Development and validation in a prospective longitudinal study of patients with rheumatoid arthritis Arthritis Rheum 1995 38 44 48 7818570
DAS Score NL Disease Activity Score in Rheumatoid Arthritis (last accessed 23 March 2005).
Smolen JS Breedveld FC Schiff MH Kalden JR Emery P Eberl G van Riel PL Tugwell P A simplified disease activity index for rheumatoid arthritis for use in clinical practice Rheumatology (Oxford) 2003 42 244 257 12595618
Aletaha D Stamm T Smolen JS Validation of the Simplified Disease Activity Index (SDAI) in an observational cohort of patients with rheumatois arthritis Ann Rheum Dis 2004 111 14672908
Wong AL Harker JO Park GS Paulus HE Longitudinal measurement of RA disease activity in a clinical practice setting: usefulness of the SDAI [abstract] Arthritis Rheum 2004 s386 s387
Smolen JS Breedveld FC Schiff MH Kalden JR Emery P Segurado OG Assessment of Disease Activity and Treatment Response to Adalimumab (HUMIRA) with the Simplified Disease Activity Index (SDAI) [abstract] Arthritis Rheum 2004 s190 s191
Soubrier M Zerkak D Dougados M Should we revisit the definition of higher disease activity state in rheumatoid arthritis (RA)? [abstract] Arthritis Rheum 2004 s387
Dessein PH Joffe BI Stanwix AE High sensitivity C-reactive protein as a disease activity marker in rheumatoid arthritis J Rheumatol 2004 31 1095 1097 15170920
Felson DT Anderson JJ Boers M Bombardier C Furst D Goldsmith C Katz LM Lightfoot R JrPaulus H Strand V American College of Rheumatology. Preliminary definition of improvement in rheumatoid arthritis Arthritis Rheum 1995 38 727 735 7779114
van Gestel AM Anderson JJ van Riel PL Boers M Haagsma CJ Rich B Wells G Lange ML Felson DT ACR and EULAR improvement criteria have comparable validity in rheumatoid arthritis trials. American College of Rheumatology European League of Associations for Rheumatology J Rheumatol 1999 26 705 711 10090187
Machold KP Stamm TA Eberl GJ Nell VK Dunky A Uffmann M Smolen JS Very recent onset arthritis: clinical, laboratory, and radiological findings during the first year of disease J Rheumatol 2002 29 2278 2287 12415582
Larsen A Dale K Eek M Radiographic evaluation of rheumatoid arthritis and related conditions by standard reference films Acta Radiol Diagn (Stockh) 1977 18 481 491 920239
Bombardier C Tugwell P A methodological framework to develop and select indices for clinical trials: statistical and judgmental approaches J Rheumatol 1982 9 753 757 6983581
Smolen JS Aletaha D Patients with rheumatoid arthritis in clinical care Ann Rheum Dis 2004 63 221 225 14962951 10.1136/ard.2003.012575
Drossaers-Bakker KW Zwinderman AH Vlieland TP Van Zeben D Vos K Breedveld FC Hazes JM Long-term outcome in rheumatoid arthritis: a simple algorithm of baseline parameters can predict radiographic damage, disability, and disease course at 12-year followup Arthritis Rheum 2002 47 383 390 12209484 10.1002/art.10513
Leigh JP Fries JF Predictors of disability in a longitudinal sample of patients with rheumatoid arthritis Ann Rheum Dis 1992 51 581 587 1535493
Kazis LE Anderson JJ Meenan RF Effect sizes for interpreting changes in health status Med Care 1989 S178 S189 2646488
Altman DG Practical Statistics for Medical Research 1991 Boca Raton, FL: Chapman & Hall
Sokka T Pincus T Most patients receiving routine care for rheumatoid arthritis in 2001 did not meet inclusion criteria for most recent clinical trials or american college of rheumatology criteria for remission J Rheumatol 2003 30 1138 1146 12784382
Dawes PT Fowler PD Clarke S Fisher J Lawton A Shadforth MF Rheumatoid arthritis: treatment which controls the C-reactive protein and erythrocyte sedimentation rate reduces radiological progression Br J Rheumatol 1986 25 44 49 3942847
van Leeuwen MA van der Heijde DM van Rijswijk MH Houtman PM van Riel PL van de Putte LB Limburg PC Interrelationship of outcome measures and process variables in early rheumatoid arthritis. A comparison of radiologic damage, physical disability, joint counts, and acute phase reactants J Rheumatol 1994 21 425 429 7516430
Combe B Dougados M Goupille P Cantagrel A Eliaou JF Sibilia J Meyer O Sany J Daures JP Dubois A Prognostic factors for radiographic damage in early rheumatoid arthritis: a multiparameter prospective study Arthritis Rheum 2001 44 1736 1743 11508423 10.1002/1529-0131(200108)44:8<1736::AID-ART308>3.0.CO;2-I
Donald F Ward MM Evaluative laboratory testing practices of United States rheumatologists Arthritis Rheum 1998 41 725 729 9550483 10.1002/1529-0131(199804)41:4<725::AID-ART21>3.0.CO;2-Z
Aletaha D Stamm T Smolen JS Validation of the Simplified Disease Activity Index (SDAI) in an observational cohort of patients with rheumatoid arthritis [abstract] Ann Rheum Dis 2004 111 14672908
Landewé RB van der HD Voskuyl A Boers M Condensed 28-joint counts jeopardise the construct validity of DAS28-remission: a comparison with the original DAS [abstract] Arthritis Rheum 2004 s547
Fuchs HA Pincus T Reduced joint counts in controlled clinical trials in rheumatoid arthritis Arthritis Rheum 1994 37 470 475 8147923
Smolen JS Breedveld FC Eberl G Jones I Leeming M Wylie GL Kirkpatrick J Validity and reliability of the twenty-eight-joint count for the assessment of rheumatoid arthritis activity Arthritis Rheum 1995 38 38 43 7818569
Scott DL Pugner K Kaarela K Doyle DV Woolf A Holmes J Hieke K The links between joint damage and disability in rheumatoid arthritis Rheumatology (Oxford) 2000 39 122 132 10725061
Liang MH Fries JF Containing costs in chronic disease: monitoring strategies in the gold therapy of rheumatoid arthritis J Rheumatol 1978 5 241 244 748547
Aletaha D Smolen JS Laboratory testing in rheumatoid arthritis patients taking disease-modifying antirheumatic drugs: clinical evaluation and cost analysis Arthritis Rheum 2002 47 181 188 11954012 10.1002/art.10266
Yazici Y Erkan D Paget SA Monitoring by rheumatologists for methotrexate-, etanercept-, infliximab-, and anakinra-associated adverse events Arthritis Rheum 2003 48 2769 2772 14558081 10.1002/art.11277
De Vries-Bouwstra JK Goekoop-Ruitermen YPM Van Zeben D Breedveld FC Dijkmans BAC Hazes JMW A comparison of clinical and radiological outcomes of four treatment strategies for early rheumatoid arthritis: results of the BEST trial [abstract] Ann Rheum Dis 2004 58
| 15987481 | PMC1175030 | CC BY | 2021-01-04 16:02:38 | no | Arthritis Res Ther. 2005 Apr 7; 7(4):R796-R806 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1740 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17431598748210.1186/ar1743Research ArticleThe differential contribution of tumour necrosis factor to thermal and mechanical hyperalgesia during chronic inflammation Inglis Julia J [email protected] Ahuva [email protected] Delphine M [email protected] Stephen P [email protected] Yuti [email protected] Bruce L [email protected] Bone and Joint Research Unit, Barts and The London School of Medicine and Dentistry, John Vane Science Centre, London, UK2 Experimental Therapeutics, Barts and The London School of Medicine and Dentistry, John Vane Science Centre, London, UK3 Department of Anatomy and Developmental Biology, University College London, London, UK2005 12 4 2005 7 4 R807 R816 5 10 2004 3 12 2004 21 2 2005 16 3 2005 Copyright © 2005 Inglis et al.; licensee BioMed Central LtdThis is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Therapies directed against tumour necrosis factor (TNF) are effective for the treatment of rheumatoid arthritis and reduce pain scores in this condition. In this study, we sought to explore mechanisms by which TNF contributes to inflammatory pain in an experimental model of arthritis. The effects of an anti-TNF agent, etanercept, on behavioural pain responses arising from rat monoarthritis induced by complete Freund's adjuvant were assessed and compared with expression of TNF receptors (TNFRs) by dorsal root ganglion (DRG) cells at corresponding time points. Etanercept had no effect on evoked pain responses in normal animals but exerted a differential effect on the thermal and mechanical hyperalgesia associated with rat arthritis induced by complete Freund's adjuvant (CFA). Joint inflammation was associated with increased TNFR1 and TNFR2 expression on DRG cells, which was maintained throughout the time course of the model. TNFR1 expression was increased in neuronal cells of the DRG bilaterally after arthritis induction. In contrast, TNFR2 expression occurred exclusively on non-neuronal cells of the macrophage–monocyte lineage, with cell numbers increasing in a TNF-dependent fashion during CFA-induced arthritis. A strong correlation was observed between numbers of macrophages and the development of mechanical hyperalgesia in CFA-induced arthritis. These results highlight the potential for TNF to play a vital role in inflammatory hyperalgesia, both by a direct action on neurons via TNFR1 and by facilitating the accumulation of macrophages in the DRG via a TNFR2-mediated pathway.
==== Body
Introduction
Pain and disability are the principal clinical features associated with chronic arthritis. Inflammation is associated with sensitisation of specialised sensory neurons that comprise the nociceptive (pain) pathway, leading to enhanced pain sensations in response to both noxious and non-noxious stimuli (termed hyperalgesia and allodynia, respectively) [1]. Neural sensitisation is thought to arise in response to the actions of mediators on both peripheral and central nociceptive neurons. Whereas acute inflammatory mediators such as prostaglandins and bradykinin have been shown to exert important effects on neuronal sensitisation in the short term, the longer-term influences of these and other mediators on nociceptive neurons remain less clear.
Therapies directed against tumour necrosis factor (TNF) have proved highly effective in treating rheumatoid arthritis. In addition to exerting an anti-inflammatory effect and slowing the progression of rheumatoid arthritis, anti-TNF therapy produces a profound and rapid analgesia [2]. The mechanism of this latter action remains uncertain, but the action suggests an important role for TNF in persistent inflammatory (neuroplastic) pain.
In acute situations, TNF has been reported to sensitize nociceptive neurons indirectly via the induction of a proinflammatory cytokine cascade involving IL-1β, IL-6, and IL-8, resulting in the release of prostaglandins and other mediators from immune cells [3-5]. Evidence for more direct TNF actions comes from electrophysiological studies showing that low-dose subcutaneous injections of TNF-α induce ectopic activity in nociceptive neurons within 2 minutes, with higher doses producing significant mechanical and thermal hyperalgesia by 15 minutes [6-8]. Furthermore, application of TNF-α enhances calcium currents and increases neuron sensitivity to the neurotoxin capsaicin in cultures of sensory neurons [9-11].
TNF acts via two receptors, including the p55 TNF receptor type 1 (TNFR1) and the p75 TNF receptor 2 (TNFR2). Both receptors have been reported as being present within the rat dorsal root ganglion (DRG), although the cellular distribution remains controversial [11-13]. Recently, increased neuronal TNFR1 expression was reported in association with intraperitoneal lipopolysaccharide [13], suggesting a direct TNF effect on nociceptive pathways via TNFR1. Neuronal expression of TNFR2 has been reported by some investigators [12,14,15] but not others [13]. While these studies have been performed in normal animals or within hours of the onset of inflammation, the situation during more chronic phases of inflammatory disease remains unclear.
The antihyperalgesic actions of an anti-TNF drug, etanercept, have been investigated in an experimental neuropathic model [16], but there are no reports of the antihyperalgesic effects of these drugs in inflammatory models. Nociceptive mechanisms in neuropathic conditions differ from those present in inflammatory disorders. Therefore, our aim was to assess the effects of etanercept on various behavioural pain measures in an experimental model of persistent arthritis and to compare these with the sequential expression and cellular distribution of TNF receptors by DRG cells, particularly during the later and clinically more relevant phases of inflammatory disease.
Materials and methods
Animals
Adult (180 to 250 g) male Wistar Rats were kept in groups of between 3 and 5 animals in cages maintained at 20°C with a 12-hour light/dark cycle and food and water ad libitum. All experimental procedures were approved by the UK Home Office and followed guidelines issued by the International Association for the Study of Pain. Inflammation was induced by a single intraplantar injection (100 μl) of complete Freund's adjuvant (CFA) (Becton Dickinson, Franklin Lakes, NJ, USA) into the right hind footpad of each animal, prepared as a 10-mg/ml suspension of heat-attenuated Mycobacterium tuberculosis in sterile paraffin oil (Sigma, St Lois, MO, USA). For subsequent in situ hybridisation and real-time PCR analysis, animals were killed with CO2, and for immunohistochemical analysis they were killed with an overdose of phenobarbitone, followed by intracardiac perfusion with 250 ml saline and 500 ml periodate–lysine–paraformaldehyde fixative [17]. L4/L5 DRG tissues were removed for assessment.
Drugs
Etanercept (Wyeth Pharmaceuticals, Madison, WI, USA) (0.5 mg/kg reconstituted in 0.5 ml sterile water) was administered subcutaneously on alternate days starting either 1 day before or 3 days after the induction of inflammation with CFA. Control animals received denatured etanercept that had previously been heated to 95°C for 10 min.
Behavioural assessments
Footpad diameter was assessed with calipers before and 3 and 7 days after CFA injection (n = 4 each time). Behavioural pain measures were assessed at the same time points. Nociceptive thresholds to mechanical stimulation were determined using von Frey hairs of increasing gauge (0.6 to 12.6 g), as previously described [18]. Animals were placed in boxes of wire mesh and von Frey hairs were applied to the plantar surface of the hind paw. The lowest-weight von Frey hair to evoke a withdrawal from three consecutive applications was considered to indicate the threshold. Thermal hyperalgesia was assessed with the Hargreaves algometer test [19]. Animals were placed in a Perspex (methyl methacrylate) box and an increasing thermal stimulus was delivered to the plantar surface of the hind paw. The time interval to lifting of the paw was recorded. All assessments were performed with the assessor blinded with respect to treatment and were repeated on at least two separate occasions.
In situ mRNA hybridisation
Primers were designed against fragments of the rat TNFR1 and TNFR2 cDNA [20,21] (Invitrogen, San Diego, CA, USA). TNFR1 (5'-CGG AAT TCC AAA GAG GTG GAG GGT GAA GGA-3' (bp 999 to 1020), 5'-CCA TCG ATC AGT GTC AAG CCG TTG TTG CTG-3' (bp 1297 to 1318)); TNFR2 (5'-CGG AAT TCC CCA GGA TGC AGT AGG CCT TGA-3' (bp 11 to 32), 5'-CCA TCG ATC AGA CGT TCA CGA TGC AGG TGA-3' (bp 230 to 253)). 320- and 242-bp fragments of the TNFR1 and TNFR2 receptors from rat brain and monocyte cDNA, respectively (gifts from Adrian Bristow, National Institute for Biological Standards and Control, UK) were cloned into PCR2.1, using a TA cloning kit (Invitrogen). Plasmids were linearised and were transcribed with DIG-UTP RNA mix (Roche, Basel, Switzerland).
In situ mRNA hybridisation was performed on 10 μM sections of DRG from naive rats 3 and 7 days after the induction of inflammation with CFA [22]. Labelling was detected with sheep anti-DIG-AP antibody (Roche), and NBT/BCIP (Roche).
Real-time RT-PCR
Forward and reverse primers were designed for the TNFR1 and TNFR2 mRNA using Primer Express software (Applied Biosystems, Foster City, CA, USA). The primers chosen for TNFR1 and TNFR2 (TNFR1: 5'-CTC TTG GTG ACC GGG AGA AG-3' (bp 98-117), 5'-GGT TCC TTT GTG GCA CTT GGT-3' (bp 203-183); and TNFR2: 5'-CAT CCC TGT GTC CTT GGG-3' (bp 709–808), 5'-CCC GTG ATG CTT GGT TCA-3' (bp 839-820)) gave products of 101 and 51 bp, respectively.
Total RNA was extracted from frozen L4/L5 DRGs using Qiagen RNeasy mini-kit with on-column DNase digestion (Qiagen, Hilden, Germany), and reverse transcription was performed using a Promega Reverse Transcription kit according to the manufacturer's instructions (Promega, Madison, WI, USA). Real-time PCR was carried out with SYBR Green PCR mastermix (Applied Biosystems) containing AmpliTaq Gold [23]. Taqman real-time PCR was performed on 10 ng of each sample and the standard curve (5 to 20 ng) for 18S RNA [23] (Applied Biosystems) in a 20-μl reaction volume. The TNFR mRNA level was expressed as a ratio to 18S RNA.
For cloning and expression of the rat TNFR2, total RNA was isolated from rat spleen using a Qiagen RNeasy mini-kit with on-column DNAse digestion (Qiagen). RNA was reverse-transcribed using Promega Reverse Transcription kit (Promega). Primers were generated to amplify the full coding region of the receptor, according to the corresponding cDNA sequence in the mouse [24] (Invitrogen) (5'-CTG GGT ACC ACC ATG GCG CCC GCC GCC CTC-3', 5'-GGC CAC TTT GAC TGC AAT CT-3'). The 1384-bp fragment was cloned into PCDNA6 His B (Invitrogen). TNFR2-PCDNA6, or empty vector, was transfected into human embryonic kidney 293T cells using a CaCl2–glycerol shock protocol described previously [25]. The cloned receptor was sequenced and submitted to GenBank (17-07-2003, NCBI Accession number AY344841).
Production and characterisation of an anti-TNFR2 single-chain variable fragment (ScFv)
Selection was carried out as described by White and colleagues [26], with minor modifications. 293T cells were seeded in 6-well plates at 5 × 105. One well per plate was transfected with TNFR2, while two were left untransfected. Forty-eight hours after transfection, 500 μl Tomlinson J-library phage [27], prepared as described at , (2 × 1011TU) was depleted on untransfected 293T cells, and unbound phage were collected. Unbound phage was then applied to a transfected culture and bound phage was eluted by incubation with 500 μl 100 mM triethylamine, pH 12. Three rounds of selection were performed. After the final round, phage was infected into HB2151 bacteria, and monoclonal ScFv ELISA was performed against TNFR2-transfected cell lysate and vector-transfected lysate. A strongly binding clone was produced.
To ensure specificity, double immunohistochemistry was performed on TNFR2-transfected 293T cells, first with anti-His FITC (Qiagen) in accordance with the manufacturer's instructions. Subsequently, cells were stained with ScFv and detected using anti-myc-CY-3 antibody (Sigma, St Louis, MO, USA). To further assess specificity, immunoprecipitation was performed against 293T cells expressing His-tagged TNFR1, or TNFR2 using the produced ScFv, and an anti-His antibody (Qiagen). ScFv or mouse anti-His antibody (5 μg) was mixed with 20 μl 50% (w/v) Protein A beads (Amersham, Little Chalfont, UK) and 100 μg TNFR1- or TNFR2-transfected 293T lysate. Protease cocktail inhibitors were added, and solutions were incubated overnight with agitation at 4°C. Protein A beads were centrifuged and washed three times with 1 ml PBS to remove non-bound protein. Beads were then resuspended in 30 μl SDS loading dye and heated for 10 min at 95°C. Samples were then run on a 10% SDS gel and probed with an anti-His antibody.
Immunohistochemistry and histology
L4/L5 DRGs were removed and cryoprotected overnight in a 30% sucrose solution. 10 μM sections of the DRGs were stained with mouse anti-ED1 (Serotec, Oxford, UK), rabbit anti-ATF-3 (Santa Cruz, Santa Cruz, CA, USA), rabbit anti-GFAP (Dako Cytomation, Glostrup, Denmark), guinea pig anti VR-1 (Chemicon, Temecula, CA, USA), or human antirat TNFR2 ScFv. Detection was performed with fluorescently labelled conjugates (Jackson Immunolabs, West Grove, PA, USA). An anti-His Alexa Fluor 488 (Qiagen) was used for detection of ScFv binding. ED-1-positive cells adjacent to neuronal somata were counted, and the results were expressed as the ratio of the number of positive cells to the number of neuronal somata [28]. Macrophages with a nucleus were counted; hence measurements are conservative. Joints were decalcified, embedded in paraffin wax, sectioned, and stained with haematoxylin and eosin. Joints were assessed for inflammatory infiltrate, bone necrosis, and cartilage degradation [17]. Joints were scored from 0, indicating no damage, to 4, indicating severe damage or infiltration. Assessments were performed in a blinded fashion by a trained investigator (BLK).
Data analysis
Statistical analysis was performed using one-way analysis of variance with post hoc Bonferroni multiple range testing and unpaired t-test for comparing two means. Linear regression and Pearson's correlation were performed to assess the relation between macrophage numbers and mechanical hyperalgesia.
Results
Characteristics of CFA-induced arthritis
Administration of CFA produced a twofold increase in paw diameter in control arthritic animals treated with denatured etanercept (Fig. 1a) as well as a reduction in both the thermal withdrawal latency (Fig. 1b) and the mechanical withdrawal threshold (Fig. 1c). Histological assessment demonstrated a pronounced inflammatory infiltrate with modest bone necrosis and cartilage loss (Fig. 1d).
Effects of TNF antagonism on CFA-induced arthritis
The effects of the TNF antagonist etanercept on inflammation and behavioural pain measurements during CFA arthritis were assessed. Etanercept produced a modest decrease in paw swelling when given either before or after the onset of inflammation (Fig. 1a). No gross histological differences were seen between etanercept-treated and control-treated animals (Fig. 1d).
Treatment with etanercept before the onset of inflammation had no effect on thermal withdrawal latencies in naive animals or those tested 3 days after CFA injection, but abolished thermal hyperalgesia at 7 days (Fig. 1b). It was notable that treatment with etanercept after the onset of inflammation also abolished thermal hyperalgesia at the 7-day time point.
Mechanical hyperalgesia in naive animals was unaffected by etanercept. When administered before the onset of inflammation, etanercept significantly attenuated mechanical hyperalgesia such that the withdrawal threshold increased from 12 g to 23 g at 3 days and to 39 g at 7 days (P < 0.05). In contrast, little effect was observed on mechanical hyperalgesia when etanercept treatment was commenced 3 days after injection of CFA (Fig. 1c).
TNF receptor mRNA expression in rat dorsal root ganglia during CFA arthritis
In order to investigate changes in TNF receptor expression in response to inflammation, we performed real-time RT-PCR and in situ mRNA hybridisation using gene-specific primers on DRG tissues taken from normal animals and at 3- and 7-day time points after induction of CFA arthritis. Real-time RT-PCR showed a modest expression of TNFR1 mRNA in naive tissues, with a threefold increase in TNFR1 mRNA in both the ipsilateral and contralateral DRG at 3 and 7 days after induction of inflammation (Fig. 2a). In situ hybridisation failed to detect TNFR1 mRNA in the naive rat (Fig. 2b) but identified TNFR1 mRNA in DRG cells at 3 and 7 days (Fig. 2c).
Expression of TNFR2 mRNA measured by real-time RT-PCR was increased after the induction of inflammation (Fig. 3a), rising to four times the initial level by 7 days after induction of inflammation. In contrast to TNFR1, TNFR2 mRNA was predominantly increased in the ipsilateral DRG. In situ hybridisation did not detect TNFR2 mRNA in the DRGs of naive animals (Fig. 3b) but identified ipsilateral expression of TNFR2 mRNA in small, non-neuronal cells 3 and 7 days after induction of inflammation (Fig. 3c,d).
Development of antibody against TNFR2 by subtractive phage display
The cellular distribution of TNFR2 in DRG tissues remains controversial, and in the absence of a specific antibody to TNFR2 for immunohistochemical use, studies to date have mostly relied on identifying TNFR2 mRNA in DRG sections or ganglia extracts. In order to overcome this problem, we produced a specific anti-TNFR2 ScFv for use in immunohistochemical studies in the rat.
TNFR1 and TNFR2 were cloned with a histidine tag and were then used to transfect 293T cells. Subsequently, phage display human antibody library was used to produce a ScFv against TNFR2 employing 293T cells transfected with TNFR2. Specificity of the ScFv was confirmed through ELISA (9.3-fold increased binding to TNFR2 relative to TNFR1). The ScFv was shown to immunoprecipitate TNFR2 but not TNFR1 (Fig. 4a). Specificity was also tested using immunohistochemistry against TNFR2-transfected 293T cells. An anti-His-FITC antibody was used to detect cells expressing the TNFR2 (Fig. 4b), and colocalisation was shown by double immunohistochemistry using the ScFv and an anti myc-CY-3 antibody (Fig. 4c).
Characterisation of TNFR2-expressing DRG cells
The ScFv TNFR2 antibody described in the previous section was used in triple immunohistochemical studies to characterise TNFR2-expressing DRG cells in naive and arthritic rats. A monoclonal antibody against rat ED1 antigen was used to recognise cells of macrophage–monocyte lineage, and an antibody against glial fibrillary acidic protein (GFAP) was used to identify glial cells.
Seven days after CFA injection, TNFR2 protein was detected in small DRG cells (Fig. 4d). At all time points studied, TNFR2 colocalisation was observed exclusively with ED1-positive cells (Fig. 4e), and no colocalisation was observed with GFAP-positive cells (Fig. 4f).
Modest numbers of ED1-positive cells were observed in the naive rat DRG. Significantly, there were 7- and 10-fold increases in the ratio of ED1-positive cells to neuronal cells at, respectively, 3 and 7 days after CFA injection in the ipsilateral DRG (Fig. 5a). A less marked increase in ED1-positive cell numbers was observed in the contralateral DRG, reaching statistical significance only at 7 days.
In order to assess for possible neuronal injury in the CFA model, we used double immunohistochemistry with ED1 (Fig. 5b) and activating transcription factor-3 (ATF-3), a marker of nerve damage (Fig. 5c.) At all assessed time points, little or no ATF-3 expression was observed, with a maximum of two ATF-3-positive cells per DRG section.
Effects of etanercept on ED1-positive cell numbers in rat DRG
Administration of etanercept before onset of CFA-induced inflammation produced a 75% reduction in ED1-positive cell numbers in the ipsilateral DRG in comparison with controls at 7 days (Fig. 5d). In contrast, treatment with etanercept after the onset of inflammation had no effect on ED1-positive cell numbers at the same time point (Fig. 5d). A highly significant correlation between numbers of ED1-positive cells in the DRG and development of mechanical hyperalgesia was observed 7 days after inflammation (R2 = 0.677, P < 0.01.)
Discussion
This study has provided several novel findings. One is that etanercept exerts a differential effect on the thermal and mechanical hyperalgesia associated with rat CFA-induced arthritis. Another is that experimental joint inflammation is associated with increased TNF-receptor expression on DRG cells that is maintained through the time course of the model. We also found that DRG TNFR2 expression occurs exclusively on non-neuronal cells of the macrophage–monocyte lineage, with cell numbers increasing in a TNF-dependent fashion during CFA arthritis. And, finally, we found a strong correlation between numbers of macrophage–monocyte cells and development of mechanical hyperalgesia in CFA arthritis.
CFA-induced arthritis has been used extensively in studies of behavioural pain responses [29,30]. Three hours after induction of inflammation, TNF is significantly up-regulated in local tissues, with the rise being maintained for a minimum of 5 days [3]. Consistent with a nociceptive role for TNF in this and other experimental models, antisera to TNF have been reported to transiently reduce thermal and mechanical hyperalgesia early in the course of CFA arthritis (within 24 hours) as well as after intraplantar injection of carageenan and also lipopolysaccharide, [31]. The effect of anti-TNF treatment at later time points has not previously been reported; however, prior inflammation (of rat knee) has been shown to be associated with an increased sensitivity to TNF [32], suggesting the presence of a dynamic, time-dependent process.
Etanercept had little or no effect on pain-related behaviour in naive animals. This observation accords with the observation that TNF administration into noninflamed tissues induces only modest hyperalgesia [33]. It is also consistent with the present finding of relatively low-grade TNF receptor expression in both neuronal and non-neuronal DRG cells taken from animals without inflammation.
In this study, the hyperalgesic response to thermal stimuli was reduced by etanercept, irrespective of whether it was administered before or after the onset of inflammation. In contrast, etanercept reduced mechanical hyperalgesia only when administered before the onset of inflammation. These results imply that although TNF plays a role in both the thermal and the mechanical hyperalgesia that accompanies inflammation, different mechanisms may be operating.
After induction of CFA arthritis, expression of both TNFR1 and TNFR2 within the DRG tissues was markedly increased. The observation that neuronal expression of neuronal TNFR1 increased with time confirms and extends previous observations made in neuropathic models [15,13]. It also accords with the time-dependent effects of etanercept, which was effective at reducing thermal hyperalgesia at 7 days but not at earlier time points of the experimental model, when TNFR1 expression was less apparent. This evolving pattern mirrors that observed after spinal ligation, which is associated with increasing sensitivity of both injured and noninjured neurons to TNF [34].
In contrast to TNFR1, we found no evidence for neuronal expression of TNFR2. Previously, TNFR2 has variously been reported either as being expressed on neuronal cells [11,12] or as not being expressed [13]. In our study, we used a highly specific antibody to TNFR2 protein, and although low-grade neuronal TNFR2 expression may have been below the detection threshold for the immunocytochemical technique used, it seems unlikely that functionally important expression was present.
TNFR2 was, however, shown to be expressed by non-neuronal cells of macrophage–monocyte lineage. The expression of TNFR2 increased significantly during the time course of CFA arthritis as a result of increased numbers of ED1-positive cells in the DRG following inflammation. To our knowledge, macrophage invasion of the DRG has previously been reported only in association with nerve injury [28]. In neuropathic models, ATF-3, a member of the ATF/CREB family, is up-regulated in damaged neurons [35]. Although its role after nerve damage is not known, it is regarded as a unique neurochemical marker of nerve injury. Little or no ATF-3 DRG expression was observed at any time point during CFA-induced inflammation; this observation provides strong evidence that there is not significant nerve injury in the CFA model.
A unique finding in the present study is that there was a highly significant correlation between macrophage numbers and the mechanical pain threshold 7 days after the onset of inflammation (Fig. 5e). Pretreatment with etanercept virtually abolished CFA-related macrophage infiltration, while substantially reducing the development of mechanical hyperalgesia. Paradoxically, treatment after the onset of inflammation had no effect on macrophage numbers and did not ameliorate mechanical hyperalgesia.
The antihyperalgesic properties of etanercept observed in the CFA model show striking similarities to those observed in chronic constriction injury of peripheral nerves [16]. Etanercept treatment that is started before the nerve is injured reduces subsequent mechanical hyperalgesia dramatically, while treatment started after injury has little effect. In addition, both prophylactic and postconstriction treatment with etanercept in the chronic-constriction-injury model reduces thermal hyperalgesia only at later time points. The magnitude of the infiltration of macrophages into the DRG after induction of CFA arthritis was of a similar degree to that observed following axotomy [28], leading us to question both the mechanism leading to the infiltration and the resulting functional role that DRG macrophages might play in hyperalgesic responses in models of inflammatory disease.
Traditionally, macrophage infiltration into the DRG has been thought to play a key role in degeneration and repair of the damaged nervous system [36]. However, the apparent absence of nerve damage in the CFA arthritis model leads to speculation that DRG macrophage infiltration may play additional roles. Depletion of macrophages during neuropathic conditions reduces hyperalgesia [37]. A study comparing macrophage infiltration into the DRG in neuropathic models [38] found low-level macrophage infiltration in models with extensive nerve damage, whereas macrophage numbers were highest in models with the greatest mechanical hyperalgesia. The conclusion was that DRG macrophages played a more important role in the genesis of hyperalgesia than in repair of neuronal damage. It therefore seems likely that DRG macrophages play a role in hyperalgesic states associated both with neuronal injury and with conditions associated with chronic inflammation.
Conclusion
This study has shown a differential effect of anti-TNF therapy on thermal and mechanical inflammatory hyperalgesia. Inflammation is associated with increased TNFR1 expression by neuronal cells, potentially leading to direct activation of these cells by TNF under inflammatory conditions. Inflammation was associated with a TNFR2-dependent accumulation of macrophages into the DRG. Macrophage infiltration into the DRG has previously only been reported following nerve injury and has been linked to a role in tissue damage and repair. The correlation between DRG macrophage numbers and mechanical hyperalgesia suggests a much broader role for these cells in the maintenance of chronic pain states. The analgesic efficacy of anti-TNF therapies in human arthropathies attests to the importance of TNF in the pathogenesis of chronic arthritic pain. These results indicate that TNF makes a more direct contribution to chronic inflammatory pain than has hitherto been assumed.
In summary, present evidence suggests that under noninflammatory conditions, TNF-α acts on peripheral cells to induce a proinflammatory cascade resulting in the release of additional mediators to sensitise and activate nociceptive neurons. The present study has shown that chronic inflammation is associated with up-regulation of TNFR1 on DRG neurons, thereby providing an opportunity for direct interaction between TNF and sensory neurons. This study has also shown the presence of TNFR2-expressing infiltrating macrophages, providing a second pathway by which TNF can modulate neuronal function in inflammatory as well as neuropathic conditions (see Fig. 6).
Abbreviations
ATF-3 = activating transcription factor-3; CFA = complete Freund's adjuvant; DRG = dorsal root ganglion; ELISA = enzyme-linked immunosorbent assay; FITC, fluorescein isothiocyanate; GFAP = glial fibrillary acidic protein; L4/L5 = fourth/fifth lumbar vertebrae; PBS = phosphate-buffered saline; ScFv = single-chain variable fragment; TNF = tumour necrosis factor; TNFR(1/2) = tumour necrosis factor receptor (type 1/2).
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
JJI carried out the study and analysed the data. JJI, BLK, YC, and SPH designed the study. AN contributed to the phage display experimental design. YC contributed to the cloning and expression of the TNFRs and riboprobes. DML assisted with the real-time RT-PCR experiments. All authors read and approved the final manuscript.
Acknowledgements
This work was supported by Barts and the London Joint Research Board, the Arthritis Research Campaign, and the William Harvey Research Foundation.
Figures and Tables
Figure 1 Etanercept treatment attenuates mechanical and thermal hyperalgesia, with little effect on swelling and histological damage. (a) Prearthritic etanercept treatment in rats with arthritis induced by complete Freund's adjuvant (CFA) reduced paw swelling significantly, by 10%, until 4 days after inflammation induction (n = 4). Thermal (b) and mechanical (c) hyperalgesia were assessed throughout the study. (b)Both prearthritic and postarthritic etanercept therapy abolished thermal hyperalgesia 7 days after inflammation (n = 4). (c) Prearthritic etanercept treatment reduced mechanical hyperalgesia significantly throughout the period studied, and postarthritic treatment reduced hyperalgesia to a lesser extent. (d) Joints were stained with haematoxylin and eosin and were assessed on a 4-point scale (where 0 = absent, 1 = mild, 2 = moderate, 3 = severe) for severity of inflammatory infiltrate, bone necrosis, and cartilage damage. CFA induced inflammatory infiltration 3 days (not shown) and 7 days (d) after CFA, while bone and cartilage were largely preserved. We found no reduction in histological score with either etanercept treatment regime. *P < 0.05; ***P < 0.005 in comparison with controls.
Figure 2 TNFR1 mRNA is increased following inflammation in the dorsal root ganglion (DRG). (a) Real-time RT-PCR showed a threefold increase in mRNA of tumour necrosis receptor type 1 (TNFR1) (expressed as the ratio of TNFR1 mRNA to 18S mRNA) in the DRG following the induction of inflammation by complete Freund's adjuvant (CFA) (n = 4). The increase was bilateral, being observed on both the ipsilateral (black bars) and the contralateral (white bars) DRGs. (b) In situ mRNA hybridisation showed no TNFR1 detection in the naive DRG. (c) In situ hybridisation for TNFR1 in the ipsilateral DRG 7 days after inflammation showed receptor expression in neuronal cells. *P < 0.05, **P < 0.01.
Figure 3 TNFR2 mRNA is increased following inflammation in the dorsal root ganglion (DRG). (a) Real-time RT-PCR showed an increase in mRNA of tumour necrosis factor receptor type 2 (TNFR2) (expressed as the ratio of TNFR2 mRNA to 18S mRNA) in the DRG ipsilateral to inflammation (n = 4), reaching five times its original level by 7 days. Contralateral TNFR2 levels increased significantly 7 days after inflammation. (b) In situ mRNA hybridisation in the naive DRG detected no TNFR2 mRNA. (c) In situ hybridisation detected TNFR2 mRNA in small cells of the DRG surrounding neuronal somata in the ipsilateral DRG 7 days after inflammation. (d) High-power magnification of TNFR2-labelled cells in the DRG, showing a perisomal distribution. *P < 0.05, **P < 0.01. CFA, complete Freund's adjuvant.
Figure 4 Monitoring expression of TNFR2 with anti-TNFR2 single-chain variable fragment (ScFv). (a) Immunoprecipitation of His-tagged tumour necrosis factor types 1 and 2 (TNFR1 and TNFR2) with anti-His antibody (left panel), and anti-TNFR2 single-chain variable fragment (ScFv) (right panel). The ScFv precipitated TNFR2, at ~75 KDa, but not TNFR1, at ~55 KDa, while the anti-His antibody precipitated both receptors, indicating specificity of the ScFv. Double immunohistochemistry against TNFR2-transfected 293T cells with anti-His antibody (b), and ScFv plus anti-myc-CY-3 (c) indicates that the selected ScFv specifically binds toTNFR2. (d) Immunohistochemistry with anti-TNFR2 ScFv in the dorsal root ganglion 7 days after inflammation ipsilateral to injection. TNFR2 colocalisation was observed with the macrophage marker, ED1 (e) but not with glial fibrillary acidic protein (GFAP) (f), indicating expression of TNFR2 by macrophages following inflammation.
Figure 5 Prearthritic treatment with etanercept reduces postarthritic macrophage accumulation in the dorsal root ganglion (DRG). (a) Inflammation due to complete Freund's adjuvant (CFA) induced a 10-fold increase in perineuronal macrophages (n = 4) by 7 days after inflammation. Infiltration of macrophages was restricted to the ipsilateral side, except for 7 days after inflammation, when a threefold increase in perineuronal macrophages was observed in the contralateral DRG. Nerve damage was assessed through double immunohistochemistry using the macrophage marker, ED1 (b) and a neuropathic marker, ATF-3 (activating transcription factor-3) (c). A maximum of two ATF-3-positive cells was detected per DRG, indicating that little nerve damage occurs in this model. (d) Prearthritic treatment with etanercept reduced macrophage numbers in the ipsilateral DRG by 75% 7 days after inflammation in comparison with controls treated with denatured etanercept (n = 4). No differences in macrophage numbers were detected with etanercept treatment commencing 3 days after inflammation induction. (e) A significant correlation between macrophage numbers in the DRG and mechanical hyperalgesia was observed 7 days after inflammation (n = 12). All treatment groups and controls were plotted. R2 = 0.667, P < 0.01. *P < 0.05,**P < 0.01, ***P < 0.005.
Figure 6 Actions of tumour necrosis factor (TNF) on hyperalgesia in health and in chronic inflammation. In the naive dorsal root ganglion (DRG), TNFα acts on peripheral cells to induce a proinflammatory cascade resulting in the release of mediators, such as prostaglandins, that activate nociceptive neurons, resulting in pain. After chronic inflammation, tumour necrosis factor receptor type 1(TNFR1) is up-regulated on DRG neurons, while TNFR2 is expressed by infiltrating macrophages. TNF can directly modulate neuronal function and act on peripheral cells and DRG macrophages to induce inflammatory mediators that can modulate neuronal function. This results in exaggerated pain.
==== Refs
Millan MJ The induction of pain: an integrative review Prog Neurobiol 1999 57 1 164 9987804 10.1016/S0301-0082(98)00048-3
Feldmann M Development of anti-TNF therapy for rheumatoid arthritis Nat Rev Immunol 2002 2 364 371 12033742 10.1038/nri802
Woolf CJ Allchorne A Safieh-Garabedian B Poole S Cytokines, nerve growth factor and inflammatory hyperalgesia: the contribution of tumour necrosis factor alpha Br J Pharmacol 1997 121 417 424 9179382
Ferreira SH Lorenzetti BB Cunha FQ Poole S Bradykinin release of TNF-alpha plays a key role in the development of inflammatory hyperalgesia Agents Actions 1993 38 C7 C9 8317326
Cunha FQ Lorenzetti BB Poole S Ferreira SH Interleukin-8 as a mediator of sympathetic pain Br J Pharmacol 1991 104 765 767 1797337
Sorkin LS Xiao WH Wagner R Myers RR Tumour necrosis factor-alpha induces ectopic activity in nociceptive primary afferent fibres Neuroscience 1997 81 255 262 9300418 10.1016/S0306-4522(97)00147-4
Junger H Sorkin LS Nociceptive and inflammatory effects of subcutaneous TNFalpha Pain 2000 85 145 151 10692613 10.1016/S0304-3959(99)00262-6
Opree A Kress M Involvement of the proinflammatory cytokines tumor necrosis factor-alpha, IL-1 beta, and IL-6 but not IL-8 in the development of heat hyperalgesia: effects on heat-evoked calcitonin gene-related peptide release from rat skin J Neurosci 2000 20 6289 6293 10934280
Nicol GD Lopshire JC Pafford CM Tumor necrosis factor enhances the capsaicin sensitivity of rat sensory neurons J Neurosci 1997 17 975 982 8994052
Morioka N Inoue A Nakata Y [Neural-immune interactions in dorsal root ganglia] [in Japanese] Nippon Yakurigaku Zasshi 2000 115 219 227 10876806
Pollock J McFarlane SM Connell MC Zehavi U Vandenabeele P MacEwan DJ Scott RH TNF-alpha receptors simultaneously activate Ca2+ mobilisation and stress kinases in cultured sensory neurones Neuropharmacology 2002 42 93 106 11750919 10.1016/S0028-3908(01)00163-0
Schafers M Sorkin LS Geis C Shubayev VI Spinal nerve ligation induces transient upregulation of tumor necrosis factor receptors 1 and 2 in injured and adjacent uninjured dorsal root ganglia in the rat Neurosci Lett 2003 347 179 182 12875915 10.1016/S0304-3940(03)00695-5
Li Y Ji A Weihe E Schafer MK Cell-specific expression and lipopolysaccharide-induced regulation of tumor necrosis factor alpha (TNFalpha) and TNF receptors in rat dorsal root ganglion J Neurosci 2004 24 9623 9631 15509749 10.1523/JNEUROSCI.2392-04.2004
Shubayev VI Myers RR Axonal transport of TNF-alpha in painful neuropathy: distribution of ligand tracer and TNF receptors J Neuroimmunol 2001 114 48 56 11240015 10.1016/S0165-5728(00)00453-7
Parada CA Yeh JJ Joseph EK Levine JD Tumor necrosis factor receptor type-1 in sensory neurons contributes to induction of chronic enhancement of inflammatory hyperalgesia in rat Eur J Neurosci 2003 17 1847 1852 12752784 10.1046/j.1460-9568.2003.02626.x
Sommer C Schafers M Marziniak M Toyka KV Etanercept reduces hyperalgesia in experimental painful neuropathy J Peripher Nerv Syst 2001 6 67 72 11446385 10.1046/j.1529-8027.2001.01010.x
Kidd BL Inglis JJ Vetsika K Hood VC De Felipe C Bester H Hunt SP Cruwys SC Inhibition of inflammation and hyperalgesia in NK-1 receptor knock-out mice Neuroreport 2003 14 2189 2192 14625445 10.1097/00001756-200312020-00011
Tal M Bennett GJ Extra-territorial pain in rats with a peripheral mononeuropathy: mechano-hyperalgesia and mechano-allodynia in the territory of an uninjured nerve Pain 1994 57 375 382 7936715 10.1016/0304-3959(94)90013-2
Hargreaves K Dubner R Brown F Flores C Joris J A new and sensitive method for measuring thermal nociception in cutaneous hyperalgesia Pain 1988 32 77 88 3340425 10.1016/0304-3959(88)90026-7
Himmler A Maurer-Fogy I Kronke M Scheurich P Pfizenmaier K Lantz M Olsson I Hauptmann R Stratowa C Adolf GR Molecular cloning and expression of human and rat tumor necrosis factor receptor chain (P60) and its soluble derivative, tumor necrosis factor-binding protein DNA Cell Biol 1990 9 705 715 1702293
Bader T Nettesheim P Tumor necrosis factor-alpha modulates the expression of its P60 receptor and several cytokines in rat tracheal epithelial cells J Immunol 1996 157 3089 3096 8816419
Meltzer JC Sanders V Grimm PC Stern E Rivier C Lee S Rennie SL Gietz RD Hole AK Watson PH Production of digoxigenin-labelled RNA probes and the detection of cytokine mRNA in rat spleen and brain by in situ hybridization Brain Res Brain Res Protoc 1998 2 339 351 9630715 10.1016/S1385-299X(98)00010-5
Lees DM Khan NQ Barker S Corder R Quantitative measurement of mRNA levels by RT-PCR. studies of ECE-1 isoforms Methods Mol Biol 2002 206 125 145 12152224
Lewis M Tartaglia LA Lee A Bennett GL Rice GC Wong GH Chen EY Goeddel DV Cloning and expression of cDNAs for two distinct murine tumor necrosis factor receptors demonstrate one receptor is species specific Proc Natl Acad Sci USA 1991 88 2830 2834 1849278
DuBridge RB Tang P Hsia HC Leong PM Miller JH Calos MP Analysis of mutation in human cells by using an Epstein-Barr virus shuttle system Mol Cell Biol 1987 7 379 387 3031469
White SJ Nicklin SA Sawamura T Baker AH Identification of peptides that target the endothelial cell-specific Lox-1 receptor Hypertension 2001 37 449 455 11230317
de Wildt RM Mundy CR Gorick BD Tomlinson IM Antibody arrays for high-throughput screening of antibody-antigen interactions Nat Biotechnol 2000 18 989 994 10973222 10.1038/79494
Lu X Richardson PM Responses of macrophages in rat dorsal root ganglia following peripheral nerve injury J Neurocytol 1993 22 334 341 8315414 10.1007/BF01195557
Pearson CM Development of arthritis, periarthritis and periostitis in rats given adjuvant Proc Soc Exp Biol Med 1956 91 95 101 13297719
Walker K Fox AJ Urban LA Animal models for pain research Mol Med Today 1999 5 319 321 10498437 10.1016/S1357-4310(99)01493-8
Ferreira SH Lorenzetti BB Poole S Bradykinin initiates cytokine-mediated inflammatory hyperalgesia Br J Pharmacol 1993 110 1227 1231 8298813
Tonussi CR Ferreira SH Tumour necrosis factor-alpha mediates carrageenin-induced knee-joint incapacitation and also triggers overt nociception in previously inflamed rat knee-joints Pain 1999 82 81 87 10422663 10.1016/S0304-3959(99)00035-4
Cunha FQ Poole S Lorenzetti BB Ferreira SH The pivotal role of tumour necrosis factor alpha in the development of inflammatory hyperalgesia Br J Pharmacol 1992 107 660 664 1472964
Schafers M Lee DH Brors D Yaksh TL Sorkin LS Increased sensitivity of injured and adjacent uninjured rat primary sensory neurons to exogenous tumor necrosis factor-alpha after spinal nerve ligation J Neurosci 2003 23 3028 3038 12684490
Tsujino H Kondo E Fukuoka T Dai Y Tokunaga A Miki K Yonenobu K Ochi T Noguchi K Activating transcription factor 3 (ATF3) induction by axotomy in sensory and motoneurons: a novel neuronal marker of nerve injury Mol Cell Neurosci 2000 15 170 182 10673325 10.1006/mcne.1999.0814
Zimmermann M Pathobiology of neuropathic pain Eur J Pharmacol 2001 429 23 37 11698024 10.1016/S0014-2999(01)01303-6
Liu T van Rooijen N Tracey DJ Depletion of macrophages reduces axonal degeneration and hyperalgesia following nerve injury Pain 2000 86 25 32 10779657 10.1016/S0304-3959(99)00306-1
Cui JG Holmin S Mathiesen T Meyerson BA Linderoth B Possible role of inflammatory mediators in tactile hypersensitivity in rat models of mononeuropathy Pain 2000 88 239 248 11068111 10.1016/S0304-3959(00)00331-6
| 15987482 | PMC1175031 | CC BY | 2021-01-04 16:02:38 | no | Arthritis Res Ther. 2005 Apr 12; 7(4):R807-R816 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1743 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17491598748310.1186/ar1749Research ArticleDecreased levels of soluble receptor for advanced glycation end products in patients with rheumatoid arthritis indicating deficient inflammatory control Pullerits Rille [email protected] Maria [email protected] Leif [email protected] Andrej [email protected] Department of Rheumatology and Inflammation Research, University of Göteborg, Göteborg, Sweden2 Joint and Soft Tissue Unit, Department of Clinical Sciences, Lund University, Department of Orthopaedics, Malmö University Hospital, Malmö, Sweden2005 25 4 2005 7 4 R817 R824 8 12 2004 6 1 2005 4 3 2005 16 3 2005 Copyright © 2005 Pullerits et al, licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is cited.
The receptor for advanced glycation end products (RAGE) is a member of the immunoglobulin superfamily being expressed as a cell surface molecule and binding a variety of ligands. One of these ligands is high-mobility group box chromosomal protein 1, a potent proinflammatory cytokine, expression of which is increased in synovial tissue and in synovial fluid of rheumatoid arthritis (RA) patients. The interaction of high-mobility group box chromosomal protein 1 with cell-surface RAGE leads to an inflammatory response. In contrast, the presence of soluble RAGE (sRAGE) may abrogate cellular activation since the ligand is bound prior to interaction with the surface receptor.
Our aim was to analyse to what extent sRAGE is present in patients with chronic joint inflammation (RA) as compared with patients with non-inflammatory joint disease and with healthy subjects, and to assess whether there is an association between sRAGE levels and disease characteristics.
Matching samples of blood and synovial fluid were collected from 62 patients with RA with acute joint effusion. Blood from 45 healthy individuals, synovial fluid samples from 33 patients with non-inflammatory joint diseases and blood from six patients with non-inflammatory joint diseases were used for comparison. sRAGE levels were analysed using an ELISA.
RA patients displayed significantly decreased blood levels of sRAGE (871 ± 66 pg/ml, P < 0.0001) as compared with healthy controls (1290 ± 78 pg/ml) and with patients with non-inflammatory joint disease (1569 ± 168 pg/ml). Importantly, sRAGE levels in the synovial fluid of RA patients (379 ± 36 pg/ml) were lower than in corresponding blood samples and correlated significantly with blood sRAGE. Interestingly, a significantly higher sRAGE level was found in synovial fluid of RA patients treated with methotrexate as compared with patients without disease-modifying anti-rheumatic treatment.
We conclude that a decreased level of sRAGE in patients with RA might increase the propensity towards inflammation, whereas treatment with methotrexate counteracts this feature.
==== Body
Introduction
Rheumatoid arthritis (RA) is a chronic inflammatory synovitis that is dominated by the presence of macrophages, lymphocytes and synovial fibroblasts, which leads to the destruction of bone and cartilage. The pathogenesis of the disease is complex, involving a wide range of molecules.
The receptor for advanced glycation end products (RAGE) is a multi-ligand member of the immunoglobulin superfamily being expressed as a cell surface molecule and interacting with a diverse class of ligands [1,2]. RAGE is expressed by many of the cells that participate in the development of RA, including macrophages, neutrophils and T cells. RAGE is expressed on macrophages and T cells within synovial tissues of RA patients as well as on synovial fluid macrophages [3]. Moreover, synovial fibroblasts that account for about 50% of the cellular constituents of the synovial lining layer constitutively express RAGE [4].
The RAGE protein is composed of three immunoglobulin-like regions, a transmembrane domain and a highly charged short cytosolic tail that is essential for post-RAGE signalling. One of the features of the receptor is its recognition of families of ligands, rather than a single protein. The RAGE repertoire of ligands includes products of non-enzymatic glycoxidation (advanced glycation end products), the amyloid-β protein, the S100/calgranulin family of proinflammatory cytokine-like mediators, β2-integrin Mac-1 on leukocytes and the high mobility group box chromosomal protein 1 (HMGB1), all of which are associated with inflammation [2]. Studies have shown that engagement of RAGE by a ligand results in a rapid and sustained cellular activation and gene transcription [1]. Sustained receptor expression leads to a positive feedback loop in which the ligand–receptor interaction increases expression of the receptor itself on the cell surface, leading to further amplification of inflammatory response.
Soluble RAGE (sRAGE), a truncated form of the receptor, is composed of only the extracellular ligand-binding domain lacking the cytosolic and transmembrane domains (i.e. the part that transfers a signal into the cell). This soluble form of the receptor has the same ligand binding specificity and therefore competes with cell-bound RAGE for ligand binding, therefore functioning as a 'decoy' abrogating cellular activation, since the cell surface receptor remains unoccupied. Indeed, it has been demonstrated in a number of experimental animal models that treatment of animals with sRAGE prevents cell-bound RAGE signalling. For example, in a mouse model of collagen-induced arthritis, treatment of mice with sRAGE significantly reduced synovial inflammation, as well as cartilage and bone destruction [5].
In humans, sRAGE is produced by alternative splicing of RAGE mRNA [6-8]. In addition, it has also been shown that pericytes and endothelial cells produce and release sRAGE extracellularly, suggesting the presence of a negative feedback mechanism in RAGE signalling [7]. The proportion and production of the soluble form of the endogenous receptor may therefore influence the regulation of RAGE-mediated functions in various tissues and inflammatory conditions, including RA.
Since sRAGE acts as a competitive receptor for cellular RAGE, the balance between these two types of receptors might be of importance in the pathogenesis of RA. Our aim was to evaluate the levels of sRAGE in patients with RA and to assess whether there is an association between sRAGE levels and disease characteristics. As a comparison, we analysed sRAGE levels in patients with non-inflammatory joint diseases (NIDs) and in healthy subjects.
Materials and methods
Patients and controls
Blood and synovial fluid samples were collected from 62 RA patients (mean age 62 ± 13 years, mean disease duration 10 ± 8 years) who met the American College of Rheumatology criteria for RA [9]. Synovial fluids from 33 patients (mean age 43 ± 18 years) with NID were used as controls. In addition, paired blood samples from six NID patients (mean age 58 ± 12 years) were available for analysis. Patients in the NID group were diagnosed to have the following diseases: osteoarthritis, six patients (two blood samples); anterior cruciate ligament rupture, 21 patients; rupture of meniscus, four patients (three blood samples); and knee joint contusion, two patients (one blood sample). All NID patients were examined by an orthopaedic surgeon and a rheumatologist, and chronic inflammatory joint diseases were excluded.
Blood samples from 45 healthy adults with no history of diabetes mellitus or renal disease (mean age 54 ± 9 years) who underwent routine blood testing at the Sahlgrenska University Hospital as blood donors or volunteered in our laboratories were collected to determine serum sRAGE levels in a healthy population. Thirty-six out of 62 RA patients received disease-modifying anti-rheumatic drugs (DMARDs). Methotrexate predominated and was used by 27 patients, either as a monotherapy (19 patients) or in combination with biological agents (six patients [anti-tumour necrosis factor alpha targeted therapy, five patients; anti-IL-1 therapy, one patient]) or sulphasalazin (two patients). One patient was receiving anti-tumour necrosis factor alpha targeted agent in combination with azathioprin and cyclosporin A, while eight patients received monotherapy with other DMARDs (parenteral or oral gold salt compounds, three patients; cyclosporin A, one patient; sulphasalazin, three patients; leflunomide, one patient). The remaining 26 patients, receiving non-steroidal anti-inflammatory drugs or on monotherapy with corticosteroids, were considered as having no DMARD treatment.
The clinical investigation was approved by the Ethical Committee of Göteborg University, and informed consent was obtained from all patients.
Clinical and laboratory assessment
Clinical examinations were performed by the rheumatologist in all RA patients, and disease activity variables were recorded. The serum concentration of C-reactive protein was measured with a standard nephelometric assay, with normal range 0–5 mg/l. White blood cell counts in the blood were assessed using a microcell counter (F300; Sysmex, Norderstedt, Germany). The white blood cell count in the synovial fluid was also assessed in 24 RA patients.
A murine hybridoma cell line (B13.29, subclone B9), which is dependent on IL-6 for its growth, was employed for the measurement of IL-6 in synovial fluid as previously described in detail [10].
Radiographs of the hands and feet were obtained from all RA patients. Criteria for the erosive disease were the presence of one or more bone erosions, defined as loss of cortical definition of the joint and recorded in proximal interphalangeal joints, metacarpophalangeal joints, carpal joints, wrist joints and metatarsophalangeal joints. Thirty-nine patients out of 62 had erosive disease. The presence of rheumatoid factor of any of the immunoglobulin isotypes was considered positive. Thirty-eight patients had seropositive RA.
Data of patients and healthy controls are summarized in Table 1.
Collections and preparation of patient samples
Synovial fluid was collected from RA patients who attended the Department of Rheumatology at Sahlgrenska University Hospital in Göteborg with acute knee joint effusion. Synovial fluids were aseptically aspirated and immediately transferred into sodium citrate solution (0.129 mol/l, pH 7.4). Blood samples from the same patients were simultaneously obtained from the cubital vein into the sodium citrate containing tubes. Synovial fluid from NID patients who attended the Department of Orthopaedics at Malmö University Hospital in Malmö was obtained by arthrocentesis.
The collected blood and synovial fluid samples were centrifuged at 2000 × g for 10 min, aliquoted, and stored at -70°C until use.
Reagents
The levels of sRAGE in sera and synovial fluid were determined using a specific sandwich ELISA kit (R&D Systems, Minneapolis, MN, USA) according to the manufacturer's protocol. ELISA plates coated with mouse monoclonal antibody against RAGE were used for quantitative detection of sRAGE. After incubation with blood or synovial fluids, polyclonal capture antibody against the extracellular portion of RAGE was used. The minimum detectable dose of sRAGE was 4 pg/ml. According to the manufacturer, no significant cross-reactivities to EN-RAGE, HMGB1, S100A10 or S100Baa were observed.
Recombinant human HMGB1 was purchased from Sigma (St Louis, MO, USA).
Statistical analysis
Non-parametric methods were used for statistical comparisons since data showed a non-normal distribution. Statistical differences with respect to sRAGE levels between independent groups were calculated using the Kruskall–Wallis test followed by the Mann–Whitney U test. The Wilcoxon signed rank test for paired samples was used to compare differences between variables in matched samples. Correlations between different variables in patients were assessed with the Spearman rank correlation test. Fisher's exact probability test was used to assess differences between groups with regard to disease characteristics. All sRAGE values are expressed showing the median and the mean ± standard error of the mean. Patients' age and disease duration are reported as the mean ± standard deviation. P < 0.05 is considered significant.
Results
The levels of sRAGE in blood and synovial fluid
We investigated sRAGE levels in the synovial fluid and in the bloodstream of 62 patients who had RA. Blood samples from 45 healthy controls, synovial fluid from 33 patients with NID and paired blood samples from six patients with NID were assessed as controls.
RA patients displayed significantly decreased (P < 0.0001) blood levels of sRAGE (872 ± 65 pg/ml) as compared with healthy controls (1290 ± 78 pg/ml) and with NID patients (1569 ± 168 pg/ml). The sRAGE levels in synovial fluid of RA patients (379 ± 36 pg/ml) were two times lower than in corresponding blood samples (P < 0.0001), and were in the same range as in the synovial fluid of patients with NID (364 ± 30 pg/ml) (Fig. 1). There was a significant positive correlation between sRAGE levels in the matching samples of blood and synovial fluid (rs = 0.48, P = 0.0002) (Fig. 2).
Patients who had RA were significantly older than healthy controls and patients with NID (mean age 61.8 ± 13.9 years versus 54.4 ± 9.0 years and 43.0 ± 18.0 years, respectively). However, no correlation with age was found in any of the groups with respect to synovial fluid and blood sRAGE levels. Indeed, when RA patients were stratified into younger (≤ 65 years) and older (>65 years) subgroups, no statistically significant difference was found between these groups with respect to sRAGE levels. Our results indicate, however, that within the age-matched groups (mean age 52.7 ± 10.2 years for RA versus 54.4 ± 9.1 years for controls) up to 65 years of age there was still a major statistical significance regarding circulating sRAGE levels (873 ± 72 pg/ml versus 1290 ± 78 pg/ml, P = 0.0001) (Fig. 3). Synovial sRAGE level was in the same range in both younger RA patients (≤ 65 years, 345 ± 36 pg/ml) and in older RA patients (>65 years old, 430 ± 73 pg/ml), and in patients with NID (364 ± 30 pg/ml).
Correlation between sRAGE levels and clinical features of RA
We investigated further the association between sRAGE levels with main characteristics of the disease. Stratification of patient data by radiological imaging showed that 39 patients fulfilled the criteria for erosive disease, and 23 patients had no erosions on recent radiographs. There was no difference in patients' age between these two radiographic groups (61.3 ± 12.6 years versus 62.6 ± 16.1 years, respectively). No statistically significant differences in synovial fluid and blood sRAGE levels were found between these two groups (Table 2). However, patients with seropositive RA had a tendency towards lower serum sRAGE levels than patients with seronegative disease (Fig. 4). Blood and synovial levels of sRAGE were not associated with disease duration or acute-phase reactant C-reactive protein. In contrast, the synovial sRAGE levels in RA patients with erosive disease correlated significantly with synovial white blood cell counts (rs = 0.53, P < 0.04), whereas no association was found between synovial fluid sRAGE and synovial IL-6 levels in RA patients.
The effect of the treatment on sRAGE levels in RA patients
At the time of sampling all patients were receiving anti-inflammatory treatment. Since methotrexate is the most used DMARD in RA treatment and was predominant in our patient population, we decided to investigate whether this treatment had an effect of sRAGE levels in RA patients. A subgroup of patients (n = 19) receiving monotherapy with methotrexate was analysed and compared with patients without DMARD treatment (n = 26). The patients' data are presented in Table 3.
The baseline characteristics of patients in both groups were similar with respect to age and sex of patients and the presence of rheumatoid factor. However, as expected, patients receiving DMARD treatment had significantly longer disease duration than patients who did not take disease-modifying drugs (13.1 ± 9.6 years versus 8.2 ± 8.3 years, P < 0.04). Also, erosive disease was more common in this group (15/19 [79%] versus 10/26 [39%], P < 0.02).
Importantly, significantly higher sRAGE levels were found in the synovial fluid of RA patients treated with methotrexate (Fig. 5) as compared with non-treated patients. Even in this case, the synovial fluid sRAGE displayed significant correlation (rs = 0.47, P < 0.05) with blood levels.
HMGB1 expression does not influence sRAGE detection by ELISA
One of the high-affinity binding ligands for RAGE is HMGB1. Previous studies have shown that high (microgram) levels of HMGB1 are found in the synovial fluid and sera of RA patients [11,12]. In addition, we demonstrated (results not shown) that blood sRAGE in RA patients may be found on Western blot examination at 60–80 kDa, indicating in vivo or in vitro complex formation or dimerization. The complex formation between these two proteins could possibly affect the measurement of sRAGE by ELISA.
This prompted us to test whether HMGB1 binding to sRAGE influenced the detection of the latter in our experimental settings. If it were the case, the decreased sRAGE levels found in our RA patient population would be explained by in vivo or ex vivo HMGB1 interaction. Recombinant human RAGE in concentrations of 500 pg/ml and 2000 pg/ml was incubated with different concentrations (0, 0.1, 1 and 10 μg/ml) of recombinant human HMGB1, and a standard ELISA analysis was performed. Our results showed that HMGB1 did not affect the sRAGE detection by ELISA (data not shown), indicating that lower sRAGE levels measured in RA patients are not due to soluble receptor engagement with HMGB1.
Discussion
This is the first study examining sRAGE levels in patients with RA. Cell surface RAGE expression is largely dictated by the interaction with its ligands. The expression of cellular RAGE is rather low in mature animals and in human adults. Accumulation of RAGE ligands results in increased expression of the cell surface receptor itself [13]. Furthermore, the receptor–ligand interaction leads to increased RAGE-mediated signalling, resulting in an activation of several intracellular pathways including NF-κB [14].
sRAGE, a truncated form of the receptor, binds ligands with affinity equal to that of cellular RAGE. It therefore has the ability to prevent RAGE signalling acting as a decoy by binding ligands and preventing them from reaching cell surface RAGE. sRAGE has successfully been used in variety of animal disease models to antagonize RAGE-mediated pathologic processes [5,14-16]. Experiments to date have shown that pericytes and endothelial cells produce and release RAGE extracellularly, suggesting the presence of a negative feedback mechanism and immune surveillance mechanisms in RAGE signalling [7].
In our study, we found that RA patients have significantly decreased blood levels of sRAGE as compared with the healthy population and patients with NID. Why do RA patients display low levels of sRAGE? In the case of RA, there is a wide diversity of RAGE ligands present in the inflamed joints, as well as in the circulation, that could lead to the binding and consumption of sRAGE during the inflammatory process. One of the high-affinity ligands for RAGE/sRAGE is HMGB1, a potent cytokine playing an important role in the pathogenesis of chronic inflammation. HMGB1 is a potent trigger of arthritis and its expression is increased in synovial tissue of RA patients as well as in experimental arthritis [12,17]. HMGB1 levels in the synovial fluid and sera of RA patients are significantly elevated as compared with levels in osteoarthritis patients [11,18]. It is thus probable that sRAGE may form in vivo complexes with HMGB1 in the sera/synovial fluid of RA patients, leading to inaccurately low levels of sRAGE. Upon co-incubation of these two proteins, however, HMGB1 binding to sRAGE did not affect the detection of the latter, indicating that lower sRAGE levels measured in RA patients are not due to neutralization by HMGB1.
An alternative explanation for the decreased sRAGE levels in RA might be a true consumption of this molecule. In the inflammatory milieu, such as in the rheumatoid joint, other sRAGE ligands also exist. Foell and colleagues have recently reported that extracellular newly identified RAGE-binding protein (EN-RAGE), a member of the S100/calgranulin family, was strongly expressed in inflamed synovial tissue. Furthermore, highly increased serum and synovial fluid levels of EN-RAGE were found in arthritic patients in comparison with control subjects [19]. Finally, raised advanced glycation end product levels have been found in serum and synovial fluid of patients with RA [20]. The presence of high levels of these soluble ligands in RA patients provides a basis for increased consumption of the sRAGE by interaction, followed by elimination of such sRAGE–ligand complexes via the reticuloendothelial system [21].
In addition, cell-bound RAGE functions as a counter-receptor for leukocyte integrins, thereby being directly involved in leukocyte recruitment, especially in inflammatory conditions when the receptor expression increases [22]. Also, in this context, sRAGE has been suggested to function as a potential inhibitor of leukocyte recruitment [22]. In RA patients with erosive disease, we observed a positive correlation between the white blood cell count and synovial sRAGE levels, indicating that endothelial cells in the synovial blood secrete sRAGE extracellularly as a negative feedback mechanism to limit the inflammation. Alternatively, MMP-9 has been found to shed cell-bound RAGE into the culture medium in mice [23]. It is possible that in the rheumatoid joint, where expression of MMP-8 and MMP-9 is increased [24], sRAGE levels are regulated by matrix metalloproteinases in a similar manner.
Taken together, we suggest that soluble RAGE may block the ligand–RAGE interaction on the cell surface by directly binding leukocyte β2-integrin Mac-1 and thereby decreasing influx of inflammatory cells into the joint cavity, functioning as an immune surveillance mechanism. Lower levels of sRAGE detected in RA patients might thus increase the propensity towards inflammation since RAGE ligands have better access to cell membrane-bound receptor, the binding of which leads to the activation of inflammatory pathways.
Consistent with this concept, RA patients treated with methotrexate, one of the most efficient DMARDs, displayed increased sRAGE as compared with RA patients with no immunosuppressive treatment. It is known that methotrexate induces an increase of extracellular adenosine, which further downregulates the expression of adhesion molecules including β2-integrin Mac-1, a ligand for RAGE/sRAGE [25,26]. Methotrexate is also known to downregulate EN-RAGE expression in the synovium of arthritis patients [19] and to suppress activity of tumour necrosis factor alpha [25,27], the cytokine that has been shown to upregulate cellular RAGE [28]. Hypothetically, as the level of membrane-bound receptor and its ligands declines with treatment, less sRAGE is consumed and the balance is restored.
We found that sRAGE levels in RA patients' synovial fluid and sera displayed strong correlation on an individual level. Diverse splicing variants of RAGE have been found in many tissues and the proportion seems to differ between individuals [6-8]. The proportion and production of the soluble form of the endogenous receptor may therefore influence the regulation of RAGE-mediated functions in various tissues and inflammatory conditions, including RA. Whether low sRAGE levels in RA patients are the consequence of the disease or a potential contributing factor to the disease needs to be elucidated.
Conclusion
We conclude that a decreased level of sRAGE in patients with RA might increase the propensity towards inflammation, whereas treatment with methotrexate counteracts this feature.
Abbreviations
DMARD = disease-modifying anti-rheumatic drug; ELISA = enzyme-linked immunosorbent assay; EN-RAGE = extracellular newly identified RAGE-binding protein; HMGB1 = high-mobility group box chromosomal protein 1; IL = interleukin; NID = non-inflammatory joint disease; RA = rheumatoid arthritis; RAGE = receptor for advanced glycation end products; sRAGE = soluble receptor for advanced glycation end products.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
RP carried out all the experiments, performed the statistical analyses and wrote the manuscript. MB and LD participated in patients' examinations, provided samples from synovial fluid/blood as well as collected clinical data about patient groups. AT conceived of the study, participated in its design and helped in the writing of the manuscript.
Acknowledgements
This work was supported by grants from the Göteborg Medical Society, the Swedish Association against Rheumatism, the Göteborg Association against Rheumatism, the King Gustaf V foundation, the Swedish Medical Research Council, the Nanna Svartz Foundation, Stiftelsen Goljes Minne, the Lundberg Foundation, the Swedish Center for Research in Sports, Medical Faculty of Lund University and the University of Göteborg.
Figures and Tables
Figure 1 Levels of soluble receptor for advanced glycation end products (soluble RAGE) in blood and synovial fluid (SF) of rheumatoid arthritis (RA) patients and in patients with degenerative/traumatic joint diseases (non-inflammatory joint disease [NID]). In addition, blood levels of soluble RAGE were assessed in healthy controls. Box plots show the 25th and 75th percentiles. Horizontal lines in bold within boxes indicate medians, and dashed lines indicate means. Vertical bars indicate the 5th and 95th percentiles. Statistical differences with respect to soluble RAGE levels between groups were calculated using the Mann–Whitney U test, and differences between paired samples were calculated by the Wilcoxon signed rank test. Mean ± standard error of the mean (median) values are shown. NS, not significant.
Figure 2 Scattergram showing an association between blood and synovial soluble receptor for advanced glycation end products (sRAGE) levels in rheumatoid arthritis patients. The Spearman rank correlation coefficient (rs) and P value are given.
Figure 3 Blood soluble receptor for advanced glycation end products (sRAGE) levels in age-matched groups of rheumatoid arthritis (RA) patients and healthy controls. Box plots show the 25th and 75th percentiles. Horizontal lines within boxes in bold indicate medians, and dashed lines indicate means. Vertical bars indicate the 5th and 95th percentiles. Statistical differences with respect to sRAGE levels between groups were calculated using the Mann–Whitney U test.
Figure 4 Blood soluble receptor for advanced glycation end products (sRAGE) levels of rheumatoid arthritis patients stratified with respect to seropositivity and erosivity in comparison with healthy controls. Box plots show the 25th and 75th percentiles. Horizontal lines in bold within boxes indicate medians, and dashed lines indicate means. Vertical bars indicate the 5th and 95th percentiles. Statistical differences with respect to sRAGE levels between groups were calculated using the Mann–Whitney U test. The mean ± standard deviation (median) values are shown. * P < 0.01 as compared with healthy controls. RF, rheumatoid factor; no eros, no erosion.
Figure 5 Levels of soluble receptor for advanced glycation end products (soluble RAGE) in blood and synovial fluids of rheumatoid arthritis (RA) patients who received methotrexate treatment or were not treated with disease-modifying antirheumatic drugs (DMARDs) at all. Box plots show the 25th and 75th percentiles. Horizontal lines in bold within boxes indicate medians, and dashed lines indicate means. Vertical bars indicate the 5th and 95th percentiles. Statistical differences with respect to soluble RAGE levels between groups were calculated using the Mann–Whitney U test. Mean ± standard error of the mean (median) values are shown. NS, not significant.
Table 1 Clinical and demographic characteristics of patients and healthy controls
Rheumatoid arthritis patients Non-inflammatory joint disease patients Healthy controls
Patients 62 33 45
Age (years ± standard deviation) 61.8 ± 13.9 43.0 ± 18.0 54.4 ± 9.0
Sex (male/female) 18/44 20/13 2/43
Disease duration (years ± standard deviation) 10.1 ± 8.5
Rheumatoid factor (+/-) 38/24
Radiographic changes (erosive/non-erosive) 39/23
Treatment (DMARD/no DMARD) 36/26
DMARD, disease-modifying anti-rheumatic drug.
Table 2 Levels of soluble receptor for advanced glycation end products (sRAGE) in sera and in synovial fluid of rheumatoid arthritis patients according to different disease characteristics
Disease characteristic n Blood sRAGE Synovial sRAGE
Erosive rheumatoid arthritis 39
Rheumatoid factor-positive 33 832 ± 87 (771) 345 ± 39 (323)
Rheumatoid factor-negative 6 1105 ± 209 (935) 447 ± 171 (273)
Non-erosive rheumatoid arthritis 23
Rheumatoid factor-positive 5 582 ± 141 (602) 463 ± 153 (498)
Rheumatoid factor-negative 18 945 ± 128 (772) 397 ± 79 (281)
Data presented as the mean ± standard error of the mean (median).
Table 3 Clinical and demographic characteristics of patients receiving disease-modifying anti-rheumatic treatment with methotrexate or having no disease-modifying anti-rheumatic drug (DMARD) treatment
Characteristic Methotrexate treated No DMARD
Patients (n) 19 26
Age (years ± standard deviation) 63.8 ± 14.3 60.8 ± 14.0
Sex (male/female) 5/14 8/18
Disease duration (years ± standard deviation) 13.1 ± 9.6* 8.2 ± 8.3
Rheumatoid factor (+/-) 12/7 14/12
Radiographic data (erosive/non-erosive) 15/4* 10/16
* P < 0.05 as compared with patients without DMARD treatment.
==== Refs
Schmidt AM Yan SD Yan SF Stern DM The multiligand receptor RAGE as a progression factor amplifying immune and inflammatory responses J Clin Invest 2001 108 949 955 11581294 10.1172/JCI200114002
Schmidt AM Yan SD Yan SF Stern DM The biology of the receptor for advanced glycation end products and its ligands Biochim Biophys Acta 2000 1498 99 111 11108954 10.1016/S0167-4889(00)00087-2
Drinda S Franke S Ruster M Petrow P Pullig O Stein G Hein G Identification of the receptor for advanced glycation end products in synovial tissue of patients with rheumatoid arthritis Rheumatol Int 2004. Mar 26
Hou FF Jiang JP Guo JQ Wang GB Zhang X Stern DM Schmidt AM Owen WF Jr Receptor for advanced glycation end products on human synovial fibroblasts: role in the pathogenesis of dialysis-related amyloidosis J Am Soc Nephrol 2002 13 1296 1306 11961018 10.1097/01.ASN.0000013702.73570.3B
Hofmann MA Drury S Hudson BI Gleason MR Qu W Lu Y Lalla E Chitnis S Monteiro J Stickland MH RAGE and arthritis: the G82S polymorphism amplifies the inflammatory response Genes Immun 2002 3 123 135 12070776 10.1038/sj.gene.6363861
Malherbe P Richards JG Gaillard H Thompson A Diener C Schuler A Huber G cDNA cloning of a novel secreted isoform of the human receptor for advanced glycation end products and characterization of cells co-expressing cell-surface scavenger receptors and Swedish mutant amyloid precursor protein Brain Res Mol Brain Res 1999 71 159 170 10521570 10.1016/S0169-328X(99)00174-6
Yonekura H Yamamoto Y Sakurai S Petrova RG Abedin MJ Li H Yasui K Takeuchi M Makita Z Takasawa S Novel splice variants of the receptor for advanced glycation end-products expressed in human vascular endothelial cells and pericytes, and their putative roles in diabetes-induced vascular injury Biochem J 2003 370 1097 1109 12495433 10.1042/BJ20021371
Park IH Yeon SI Youn JH Choi JE Sasaki N Choi IH Shin JS Expression of a novel secreted splice variant of the receptor for advanced glycation end products (RAGE) in human brain astrocytes and peripheral blood mononuclear cells Mol Immunol 2004 40 1203 1211 15104125 10.1016/j.molimm.2003.11.027
Arnett FC Edworthy SM Bloch DA McShane DJ Fries JF Cooper NS Healey LA Kaplan SR Liang MH Luthra HS The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis Arthritis Rheum 1988 31 315 324 3358796
Pullerits R Bokarewa M Jonsson IM Verdrengh M Tarkowski A Extracellular cytochrome c, a mitochondrial apoptosis-related protein, induces arthritis Rheumatology (Oxford) 2005 44 32 9 15367748
Taniguchi N Kawahara K Yone K Hashiguchi T Yamakuchi M Goto M Inoue K Yamada S Ijiri K Matsunaga S High mobility group box chromosomal protein 1 plays a role in the pathogenesis of rheumatoid arthritis as a novel cytokine Arthritis Rheum 2003 48 971 981 12687539 10.1002/art.10859
Andersson U Erlandsson-Harris H HMGB1 is a potent trigger of arthritis J Intern Med 2004 255 344 350 14871458 10.1111/j.1365-2796.2003.01303.x
Stern D Du Yan S Fang Yan S Marie Schmidt A Receptor for advanced glycation endproducts: a multiligand receptor magnifying cell stress in diverse pathologic settings Adv Drug Deliv Rev 2002 54 1615 1625 12453678 10.1016/S0169-409X(02)00160-6
Schmidt AM Hofmann M Taguchi A Yan SD Stern DM RAGE: a multiligand receptor contributing to the cellular response in diabetic vasculopathy and inflammation Semin Thromb Hemost 2000 26 485 493 11129404 10.1055/s-2000-13204
Hofmann MA Drury S Fu C Qu W Taguchi A Lu Y Avila C Kambham N Bierhaus A Nawroth P RAGE mediates a novel proinflammatory axis: a central cell surface receptor for S100/calgranulin polypeptides Cell 1999 97 889 901 10399917 10.1016/S0092-8674(00)80801-6
Taguchi A Blood DC del Toro G Canet A Lee DC Qu W Tanji N Lu Y Lalla E Fu C Blockade of RAGE-amphoterin signalling suppresses tumour growth and metastases Nature 2000 405 354 360 10830965 10.1038/35012626
Pullerits R Jonsson IM Verdrengh M Bokarewa M Andersson U Erlandsson-Harris H Tarkowski A High mobility group box chromosomal protein 1, a DNA binding cytokine, induces arthritis Arthritis Rheum 2003 48 1693 1700 12794838 10.1002/art.11028
Ulloa L Batliwalla FM Andersson U Gregersen PK Tracey KJ High mobility group box chromosomal protein 1 as a nuclear protein, cytokine, and potential therapeutic target in arthritis Arthritis Rheum 2003 48 876 881 12687528 10.1002/art.10854
Foell D Kane D Bresnihan B Vogl T Nacken W Sorg C Fitzgerald O Roth J Expression of the pro-inflammatory protein S100A12 (EN-RAGE) in rheumatoid and psoriatic arthritis Rheumatology (Oxford) 2003 42 1383 1389 12832707
Drinda S Franke S Canet CC Petrow P Brauer R Huttich C Stein G Hein G Identification of the advanced glycation end products N(epsilon)-carboxymethyllysine in the synovial tissue of patients with rheumatoid arthritis Ann Rheum Dis 2002 61 488 492 12006318 10.1136/ard.61.6.488
Renard C Chappey O Wautier MP Nagashima M Lundh E Morser J Zhao L Schmidt AM Scherrmann JM Wautier JL Recombinant advanced glycation end product receptor pharmacokinetics in normal and diabetic rats Mol Pharmacol 1997 52 54 62 9224812
Chavakis T Bierhaus A Al-Fakhri N Schneider D Witte S Linn T Nagashima M Morser J Arnold B Preissner KT Nawroth PP The pattern recognition receptor (RAGE) is a counterreceptor for leukocyte integrins: a novel pathway for inflammatory cell recruitment J Exp Med 2003 198 1507 1515 14623906 10.1084/jem.20030800
Devaux Y Senior RM Ray P RAGE: a new target for MMP-9 in the regulation of inflammatory response in the lung during oxidative stress Am J Respir Crit Care Med 2004 169 A456 [Abstract]
Tchetverikov I Ronday HK Van El B Kiers GH Verzijl N TeKoppele JM Huizinga TW DeGroot J Hanemaaijer R MMP profile in paired serum and synovial fluid samples of patients with rheumatoid arthritis Ann Rheum Dis 2004 63 881 883 15194590 10.1136/ard.2003.013243
Chan ES Cronstein BN Molecular action of methotrexate in inflammatory diseases Arthritis Res 2002 4 266 273 12106498 10.1186/ar419
Wollner A Wollner S Smith JB Acting via A2 receptors, adenosine inhibits the upregulation of Mac-1 (Cd11b/CD18) expression on FMLP-stimulated neutrophils Am J Respir Cell Mol Biol 1993 9 179 185 8393328
Sajjadi FG Takabayashi K Foster AC Domingo RC Firestein GS Inhibition of TNF-alpha expression by adenosine: role of A3 adenosine receptors J Immunol 1996 156 3435 3442 8617970
Tanaka N Yonekura H Yamagishi S Fujimori H Yamamoto Y Yamamoto H The receptor for advanced glycation end products is induced by the glycation products themselves and tumor necrosis factor-alpha through nuclear factor-kappa B, and by 17beta-estradiol through Sp-1 in human vascular endothelial cells J Biol Chem 2000 275 25781 25790 10829018 10.1074/jbc.M001235200
| 15987483 | PMC1175032 | CC BY | 2021-01-04 16:02:38 | no | Arthritis Res Ther. 2005 Apr 25; 7(4):R817-R824 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1749 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17511598748410.1186/ar1751Research ArticleCharacterization of histopathology and gene-expression profiles of synovitis in early rheumatoid arthritis using targeted biopsy specimens Tsubaki Takahito [email protected] Norimasa 1Kawakami Takuma 2Shiratsuchi Takayuki 2Yamamoto Haruyasu 1Takubo Nobuo 3Yamada Kazuhito 3Nakata Sanpei 3Yamamoto Sumiki 3Nose Masato [email protected] Ehime University School of Medicine, Ehime, Japan2 Otsuka Pharmaceutical Co Ltd, Tokushima, Japan3 Center for Rheumatic Diseases, Matsuyama Red Cross Hospital, Ehime, Japan2005 25 4 2005 7 4 R825 R836 30 9 2004 27 10 2004 17 3 2005 29 3 2005 Copyright © 2005 Tsubaki et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The disease category of early rheumatoid arthritis (RA) has been limited with respect to clinical criteria. Pathological manifestations of synovitis in patients whose disease is clinically classified as early RA seem to be heterogeneous, with regular variations. To clarify the relation between the molecular and histopathological features of the synovitis, we analyzed gene-expression profiles in the synovial lining tissues to correlate them with histopathological features. Synovial tissues were obtained from knee joints of 12 patients with early RA by targeted biopsy under arthroscopy. Surgical specimens of long-standing RA (from four patients) were examined as positive controls. Each histopathological parameter characteristic of rheumatoid synovitis in synovial tissues was scored under light microscopy. Total RNAs from synovial lining tissues were obtained from the specimens selected by laser capture microdissection and the mRNAs were amplified by bacteriophage T7 RNA polymerase. Their cDNAs were analyzed in a cDNA microarray with 23,040 cDNAs, and the levels of gene expression in multilayered lining tissues, compared with those of normal-like lining tissues in specimens from the same person, were determined to estimate gene-expression profiles characteristic of the synovial proliferative lesions in each case. Based on cluster analysis of all cases, gene-expression profiles in the lesions in early RA fell into two groups. The groups had different expression levels of genes critical for proliferative inflammation, including those encoding cytokines, adhesion molecules, and extracellular matrices. One group resembled synovitis in long-standing RA and had high scores for some histopathological features – involving accumulations of lymphocytes and plasma cells – but not for other features. Possible differences in the histopathogenesis and prognosis of synovitis between the two groups are discussed in relation to the candidate genes and histopathology.
==== Body
Introduction
Synovial lesions in rheumatoid arthritis (RA) show complex histopathological manifestations, involving several diagnostic hallmarks such as multilayered synovial lining tissues associated with a palisading structure of the intimal lining cells and the presence of non-foreign-body-type giant cells, formation of lymphoid follicles, and massive accumulation of plasma cells and macrophages [1]. Mesenchymoid transformation and fibrinoid degeneration are definite histopathological features of RA [2]. These lesions are specific to the synovium in the progression stage of RA and their developmental processes remain unclear.
'Early RA' is a clinical term referring to the early stage of RA used to predict the eventual progression stage of RA. The American College of Rheumatology (ACR) 1987 classification criteria for RA [3] have often been used as a diagnostic tool in patients with recent-onset arthritis. However, these criteria were developed in a population of patients selected according to their disease status to classify rather than to diagnose RA. Thus, the diagnostic usefullness of these criteria in early arthritis is probably not optimal. Likewise, previous histopathological studies have been inconclusive with respect to elucidating histological features typical of early RA [4-6]. Therefore, studies of potential molecular changes in the synovium of patients with early RA may improve our understanding of this disease entity and aid diagnosis in the future.
Biopsy targeting of articular lesions in synovial tissues should be a powerful tool for clarifying the initial events of synovitis in RA. Immunohistochemical analyses of synovitis in RA using targeted biopsy specimens have shown that the histopathological features of synovium in early RA are representative of those in long-standing RA [7,8], suggesting quantitative rather than qualitative differences between various forms of synovitis in RA [9,10]. Laser capture microdissection (LCM) and extraction of total RNA followed by a cDNA microarray are techniques that have been developed mainly in molecular oncology and are used for clarifying molecular markers that have the potential to predict metastasis, sensitivity to drugs, and prognosis [11,12]. The use of these techniques to study the histopathogenesis of the initial step of synovitis in RA and its progression should improve our understanding at the molecular level.
In this study, we focused on the analysis of gene-expression profiles characteristic of proliferative lesions in the synovial lining tissues, which are one of the initial histopathological events of synovitis in early RA. That is, we prepared synovial specimens from early RA by targeted biopsy under arthroscopy, and analyzed gene-expression profiles in the synovial lining tissues selected by LCM in a cDNA microarray by comparing those in multilayered lining tissues with those in normal-like lining tissues in each case. On the basis of a cluster analysis, we propose that the synovial proliferative lesions in early RA can be classified into at least two groups. We discuss the histopathological manifestations characteristic of rheumatoid synovitis in these two groups and also the possible differences in pathogenesis and prognosis of synovitis between them.
Materials and methods
Patients and tissue samples
We studied 12 patients with early RA (duration of less than 1 year before the diagnosis), and 4 with long-standing RA (duration of more than 3 years before the diagnosis). Not all patients with early RA could be accurately diagnosed at the time of targeted biopsy, although diagnosis was possible with follow-up assessments. All patients had arthritis of the knee and fulfilled the ACR criteria for RA [3] except E-09 (early RA case no. 9) (see Table 1). Written, informed consent was obtained from each patient before they were entered into the study.
Synovial specimens in early RA were obtained from knee joints by targeted biopsy under arthroscopy, and specimens from long-standing RA were obtained by total knee arthroplasty at the Center for Rheumatic Disease, Matsuyama Red Cross Hospital. The number of specimens obtained from each patient and the macroscopic signs of synovitis with the maximum inflammatory activity at biopsy sites are shown in Table 1. For intraindividual comparison, normal-like synovial specimens that were macroscopically thin and translucent and contained only a few vessels were also obtained from each patient [13].
Histopathology
One-half of each synovial specimen was used for histopathological analysis. The tissue specimens were fixed with 10% formalin in 0.01 mol/l phosphate buffer, pH 7.2, and embedded in paraffin wax. They were stained with hematoxylin and eosin for examination by light microscopy. Histopathological parameters of synovitis were evaluated in accordance with established criteria [14], with modifications involving the degree of proliferation of synovial cells, typical palisading of synovial cells in the intimal lining layers, non-foreign-body-type giant cells in the lining regions, lymphoid and plasma cell infiltration, neovascularization, mesenchymoid transformation, and fibrinoid necrosis in synovium. Of these features, the degree of proliferation of synovial cells was scored as follows: fewer than three layers (0), three to four layers (1), five to six layers (2), or more than six layers (3). Lymphoid cell infiltration was scored as follows: none to diffuse infiltration (0), lymphoid cell aggregates (1), lymphoid follicles (2), or lymphoid follicles with germinal center formation (3). The other features were evaluated using a quantitative grading system consisting of a 4-point scale: none (0), mild (1), moderate (2), or severe (3). The maximum score with this system was 24. The results of scoring of each histopathological feature are presented as the highest score among all the specimens for the patient. The remaining half of the synovial specimen showing the highest score in the feature 'proliferation of synovial cells' was used as multilayered lining tissue for LCM. Nearly normal synovial tissues from the same patient that had no inflammatory lesions and received a score of 0 for all of the histopathological features were used as 'normal-like lining tissue' for LCM.
Laser capture microdissection
The tissue samples were placed in embedding medium (Tissue-Tek OCT Compound, Sakura Finetechnical, Tokyo, Japan) and immediately snap frozen in acetone/dry ice in the operating room before transport to the laboratory. All cryoblocks were stored at -80°C until 7-μm-thick cryosections were prepared and mounted on a 1.35-μm-thick polyethylene membrane (PALM, Wolfratshausen, Germany). The sections were immediately fixed for 3 min with acetone and for 1 min with 70% ethanol and then stained rapidly for 1 min with HistoGene™ staining solution (Arctrus, BM Equipment Co Ltd, Tokyo, Japan). They were washed with distilled water and were then dehydrated with 100% ethanol and air-dried with a fan for 3 min.
LCM was done to collect small regions from a specimen using a Robot-Microbeam (PALM) and an inverted microscope (Carl Zeiss, Oberkochem, Germany) [15]. In brief, the specimen was set on a computer-controlled microscope stage and observed from the upper side with a charged-coupling device (CCD) camera. The image was displayed, and the multilayered lining tissue and the normal-like lining tissue of the same case were selected using the computer mouse (Fig. 1a,d). We traced around the lining and then dissected it to the bottom of the specimen together with the thin membrane, using a laser microbeam through the objective lens (Fig. 1b,e). The selected tissue was then catapulted with a single laser shot into a microcentrifuge cap (0.6 ml), which was held by the micromanipulator (Fig. 1c,f). More than 5,000 cells in each specimen were dissected and pooled for RNA extraction.
RNA extraction and T7-based RNA amplification
Total RNA was extracted from the samples collected by LCM using an RNeasy spin column purification kit (Qiagen, Hilden, Germany) in accordance with the manufacturer's procedure. To remove possible genomic DNA contamination, RNase-free DNase (Qiagen) was used during the RNA purification steps. Messenger RNA was then amplified by bacteriophage T7 RNA polymerase using a RiboAmp™RNA amplification kit (Arctrus). Two or three rounds of in vitro amplification were done with the samples. The amplified RNAs from each multilayered lining tissue and normal-like lining tissue of each case were reverse-transcribed using the SuperScript preamplification system (Life Technologies, Rockville, MD, USA) with random hexamers in the presence of Cy5-dCTP and Cy3-dCTP (Amersham Biosciences Co, Piscataway, NJ, USA), respectively.
cDNA microarray
A cDNA microarray was fabricated with 23,040 cDNAs selected from the UniGene database of the National Center for Biotechnology . The cDNAs were amplified by RT-PCR using poly(A) + RNAs isolated from various human organs as templates. The PCR products were spotted in duplicate on type VII glass slides (Amersham Biosciences) with a Microarray Spotter Generation III (Amersham Biosciences).
Labeled probes were mixed with Microarray Hybridization Solution Version 2 (Amersham Biosciences) and formamide (Sigma Chemical Co, St Louis, MO, USA) to a final concentration of 50%. After hybridization for 14 to 16 hours at 42°C, the slides were washed for 10 min at 55°C in 2 X saline sodium citrate (SSC) and 1% SDS, for 10 min at 55°C in 0.2 X SSC and 0.1% SDS, and for 1 min at room temperature in 0.1 X SSC. They were then scanned using an Array Scanner Generation III (Amersham Biosciences). The fluorescence intensities of Cy5 and Cy3 for each target spot were evaluated photometrically by the ArrayVision computer program (Amersham Biosciences). Since data derived from low signal intensities are less reliable, a cutoff value for signal intensities of 10,000 was used.
Cluster analysis
To obtain reproducible clusters for classifying the 16 samples, we selected 1,035 genes for which valid expression data were obtained in all the experiments, and which included an up-regulated (Cy5/Cy3 >2) or down-regulated gene (Cy5/Cy3 <0.5) in at least two of all samples. The analysis was performed using Cluster 3.0 and TreeView software written by M Eisen and updated by Michiel de Hoon, and available on the World Wide Web . Before the clustering algorithm was applied, the fluorescence ratio for each spot was log-transformed (base 2). Then the data were median-centered and normalized for each sample, to remove experimental biases.
Statistical analysis
Euclidean distance was used to determine the differences between expression levels of individual genes. Statistical analysis on microarray data was performed using the significance analysis of microarrays (SAM) method, available on the World Wide Web . The fold change in expression was calculated for each gene between groups, and significance levels were indicated by the Q value. A Q value less than 5% was considered significant. A t-test was used to confirm the results by SAM. A P value less than 0.05 was considered significant. The Mann–Whitney U test was used to test for differences in histological scores and disease duration between groups.
Results
Histopathological features of synovitis with variations
The histopathology of the early RA specimens showed regular variations. The histological score for each lesion is summarized in Table 2. For example, as shown in Fig. 2, in E-02 the proliferation of synovial lining cells resulted in fewer than four layers (score 1), and a typical palisading structure of the lining cells was not clear (score 1); there was diffuse infiltration of lymphocytes in the sublining regions (score 0). In E-07, the proliferative lining contained fewer than four layers (score 1) but showed a typical palisading structure (score 2).
Some cases of early RA manifested synovitis, in which the histopathological features were similar to those of long-standing RA such as L-01. In E-12, the specimen showed proliferation of synovial lining cells, forming 5 to 6 layers (score 2), associated with a typical palisading structure (score 2), and there were foci of lymphocyte aggregates in the sublining regions, resembling lymphoid follicles but lacking germinal centers (score 1). Many plasma cells were involved in these lesions (score 3) (Fig. 2). Partial fibrinoid necrosis was also present (score 1).
Gene-expression profiles and clustering
As shown in Fig. 3, 18 samples from 16 cases were clustered into two major groups based on their gene-expression profiles. The dendrogram shown at the top of Fig. 3 represents similarities in expression patterns among individual cases, with shorter branches indicating greater similarities. Two cases (E-07 and E-08), which were examined with two and three rounds of amplification, were clustered most closely, supporting the reliability of our RNA amplification procedures. Of the 16 cases, ten (L-01, L-04, L-02, E-01, E-10, E-04, L-03, E-06, E-12, and E-09) clustered into one group (I) and the other six (E-03, E-02, E-08, E-07, E-05, and E-11) clustered into another group (II). The clustering analysis of only the cases with early RA, not including those with long-standing RA, gave results similar to those shown in Fig. 3. (The result is attached as Additional file 1). Moreover, there was no significant difference in disease duration of the cases with early RA in groups I and II (P = 0.34 on the Mann–Whitney test). Each group appeared to have a specific gene-expression profile that should explain the molecular nature of their etiological differences.
Candidate gene profiles in each group
Using the SAM software, we examined 1,035 genes to find which were expressed significantly differently in groups I and II. We found that the expression of 180 genes was significantly increased and that of 235 was significantly decreased in group II versus group I (Q value <5%). From these genes, we selected ones that were of interest on the basis of the data previously reported regarding the mechanisms of rheumatoid synovitis and on the positional candidate genes obtained from our genome data from arthritis models as described in the Discussion. As shown in Table 3A, the genes encoding caspase 9 (CASP9), p53 induced gene 11 (TP53I11, also called PIG11), cathepsin G (CTSG), colony-stimulating factor 2 receptor, β (CSF2RB), tumor necrosis factor receptor superfamily member 1A (TNFRSF1A), and interleukin-10 receptor, β (IL10RB) were expressed more abundantly in group II than in group I (Q < 5%, P <0.05). On the other hand, the genes encoding fibronectin 1 (FN1), β2-microglobulin (B2M), syndecan 2 (SDC2), cathepsin B (CTSB), signal transducer and activator of transcription 1 (STAT1), integrin, β2 (ITGB2), and interferon γ receptor 2 (IFNGR2) were expressed more abundantly in group I than in group II (Q < 5%, P <0.05) (Table 3B).
Comparative study of histopathological features
There were significant differences in the histological scores of groups I and II (Table 2). The mean total score for group I (13.80) was significantly higher than that for group II (6.67). The mean group I scores for 'typical palisading', 'lymphoid cell infiltration', and 'plasma cell infiltration' were all significantly higher than those for group II. Moreover, in the comparative study of only the cases with early RA, the mean total score and the mean scores for 'lymphoid cell infiltration' and 'plasma cell infiltration' in were significantly higher in group I than in group II. There were no differences between groups I and II in other histopathological features.
Discussion
There are several reports about gene-expression profiles in rheumatoid synovitis. The analysis by Zanders and colleagues [16] showed an overall increased expression of inflammation-related genes in synovial tissues in RA compared with normal synovium. However, those authors performed the analysis on pooled RA synovial tissues and pooled tissues from healthy controls. Their approach did not consider disease heterogeneity, which may have obscured differences between tissues. Van der Pouw Kraan and colleagues [17] reported that RA synovial tissues could be separated into two patterns of gene expression. The first one had a gene-expression profile consistent with inflammation and active immunity, and the second, which was histopathologically similar to that in osteoarthritis (OA) tissues, exhibited a low level of expression of inflammatory and immune system genes and instead expressed genes related to tissue remodeling. However, their study was performed with whole synovial tissues obtained at synovectomy from long-standing RA and OA patients. Therefore, it may be difficult to use these results to elucidate the developmental process of rheumatoid synovitis.
In this study, we analyzed gene-expression profiles in proliferative lesions of the synovial lining tissues in early RA using targeted biopsy of synovial tissues and LCM, followed by a cDNA microarray. We showed that synovitis in early RA could be divided into at least two different groups based on the gene-expression profiles, although their histopathologies were complex. Group I included the cases with long-standing RA, and some of its synovitis histopathological features were significantly different from those of group II, including lymphoid cell and plasma cell infiltration. Features that seemed to be characteristic of RA, such as synovial cell proliferation in the lining layers, palisading structure of the intimal lining layers, non-foreign-body-type giant cells in the lining regions, neovascularization, and fibrinoid necrosis, were not significantly different in the two groups. On the basis of these findings, we speculate that the two groups may reflect differences in the pathogenesis of synovitis. The different expression profiles of several candidate genes for RA reported previously may support this idea.
Cytokine networks
Synovial macrophages and fibroblasts in the lining tissue produce factors that activate adjacent cells and enhance synovial inflammation in both paracrine and autocrine fashion [18]. Synovial macrophages activated by tumor necrosis factor α (TNF-α) can increase the production of IL-10. This interleukin has anti-inflammatory effects through its receptor, IL-10R, which is up-regulated on synovial macrophages by TNF-α. IL-10R signaling suppresses the production of IL-1β and TNF-α. The presence of IL-10 may suppress the production of IFNγ by T cells in the synovial tissue [19]. Our study suggests that a negative feedback mechanism by anti-inflammatory cytokines such as IL-10 is predominant in group II, in light of the higher expression of TNFRSF1A and IL10RB (Table 3A). Thus, IL-10 may play regulatory roles in the progression of synovitis in the early stage of RA.
Synovial macrophages and fibroblasts are strongly activated to express high amounts of IFNγ-inducible genes, despite a low concentration of extracellular IFNγ [20,21]. STAT1 is one of the IFNγ-inducible genes. Recently, it was reported that STAT1 protein expression was elevated in rheumatoid synovitis, especially in the lining layer containing highly activated macrophages [17,22]. IFNγ, even in a low concentration, can induce sustained expression of STAT1 through its heterodimeric receptor complex consisting of IFNγ receptors 1 and 2 (IFNGR1 and IFNGR2) [23]. In our study, the signal intensity of IFNG itself was very low in all samples (data not shown), while IFNγ-inducible genes such as STAT1 and B2M were more abundantly expressed in group I (Table 3B). Thus, the effect of IFNγ in rheumatoid synovitis may be evaluated indirectly by the expression profiles of these IFNγ-inducible genes. Considering that infiltrating T cells in the rheumatoid synovium in the early stage of RA are predominantly T helper type 1 cells [8], our findings that the degree of lymphoid cell infiltration was significantly different in the two groups (Table 2) may support this idea.
Adhesion molecules
There are several histological studies showing the expression of extracellular matrices and integrins in rheumatoid synovitis [24-27]. These adhesion molecules may contribute to a positive feedback mechanism in the cytokine networks [27-29]. In our study, fibronectin 1 was more abundantly expressed in group I than in group II (Table 3B). In the whole genome analysis of rheumatic-disease-susceptibility loci in MRL/lpr mice, Sdc2 (encoding syndecan 2) was a candidate gene for progressive arthritis [30]. This was highly expressed in group I in this study. Itgb2 was a candidate gene for enthesopathy [31] and coincidentally for sialoadenitis [32], and was also highly expressed in group I.
Cathepsins
CTSB, the gene for cathepsin B, one of the cysteine proteases, was more abundantly expressed in group I than in group II (Table 3B). This protease, which can cleave collagens and proteoglycans, is thought to have a prominent role in destructive arthropathies [33]. It is spontaneously expressed in cultured synovial fibroblasts and can be increased by TNF-α, IL-1, and IFNγ [34,35]. Immunolocalization studies showed cathepsin B to be expressed predominantly in synovial cells attached to the cartilage and bone at sites of rheumatoid joint erosion [33,36]. Taken together, these observations suggest the development of cartilage degeneration and bone resorption in group I, possibly in the progression stage.
On the other hand, CTSG, the gene for cathepsin G, one of the serine proteases, was more abundantly expressed in group II than in group I (Table 3A). This protease is normally associated with myeloid cells such as neutrophils and macrophages and can be induced by TNF-α [37]. It has been shown that cathepsin G proteolytically activates caspase 7 [38], an intracellular cysteine proteinase, and, more recently, that it has a role in apoptosis through cleavage of substrates regulating chromatin conformation [39]. This suggests that apoptosis may be impaired in group I.
p53 tumor suppressor gene
Although RA has many features of autoimmunity, nonimmunologic factors also play a significant role, especially in the progression stage [40-42]. Rheumatoid synovial tissues and synovial fibroblasts exhibit some features of transformation, including autonomous invasion into cartilage, expression of oncogenes, loss of contact inhibition, and insufficient apoptosis [41-44]. p53 protein is induced by many genotoxic stresses, which leads to cell cycle arrest and apoptosis of the injured cells [45]. In our study, CASP9 [46] and PIG11 [47], which encode proteins involved in apoptosis as downstream targets of p53, were abundantly expressed in group II, but not in group I (Table 3A).
Reactive oxygen and nitrogen species produced at chronic inflammatory sites may damage DNA. If the p53 gene itself gets damaged, apoptosis may be impaired. The p53 mutations are dominant negative and can interfere with endogenous wild-type p53 function [48]. Significantly higher expression of p53 is detected in rheumatoid synovial tissues than in those tissues in patients with OA or reactive arthritis [49]. Of interest, p53 was found in early RA and also in clinically uninvolved joints in RA patients [50]. Yamanishi and colleagues [51] showed that abundant p53 transition mutations, which are characteristic of the DNA damage caused by oxidative stress, were located mainly in the lining tissues, in studies using microdissected rheumatoid synovial tissues. Considering these findings, mutant p53 may be over expressed in the multilayered lining in group I, which fails to induce CASP9 and PIG11, while wild-type p53 in group II may induce these genes in group II.
The results of the study suggest that a combination of histopathology and gene-expression profiling is a useful tool for diagnostic and prognostic studies of early RA. For example, patients E-01 and E-06 had a few histopathological features specific for RA and showed lower total scores in histopathological features (Table 2), despite the fact that the villous synovial tissues were targeted and examined. However, these patients belonged to group I with respect to their gene-expression profiles. Their disease might advance to the progression stage, the same as the cases of long-standing RA, but different from those in group II. Patient E-05 was a 77-year-old woman who had polyarthralgia associated with marked pitting edema of the dorsum of the hands. The serological tests gave negative results except for mild elevation of erythrocyte sedimentation rate and C-reactive protein. These clinical manifestations could not rule out the possibility of remitting seronegative symmetrical synovitis with pitting edema syndrome (RS3PE) originally described by McCarty and colleagues [52]. This case had a few histopathological features specific for RA except for the proliferation of synovial lining cells associated with a typical palisading structure and it had lower total scores and belonged to group II.
Additional studies will be needed to compare gene-expression profiles of such a case in group II with those of other synovitis diseases such as reactive arthritis and OA, especially with respect to the candidate genes described above. Follow-up studies will be conducted to investigate potential differences in the clinical course of cases in groups I and II.
Conclusion
In this study, we analyzed gene-expression profiles in the synovial lining tissues in situ in early RA using synovial specimens obtained by targeted biopsy, followed by LCM and cDNA microarray analyses. Based on cluster analysis, we found at least two groups in synovitis in early RA, one of which resembled that in long-standing RA. This grouping may reflect differences in the histopathogensis of synovitis in early RA. Different expression profiles of the several candidate genes may provide useful information for future studies of the diagnosis and prognosis of early RA.
Abbreviations
ACR = American College of Rheumatology; IFN = interferon; IL = interleukin; LCM = laser capture microdissection; OA = osteoarthritis; RA = rheumatoid arthritis; SAM = significance analysis of microarrays; SSC = saline sodium citrate; TNF = tumor necrosis factor.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
TT carried out critical examinations in this study, especially synovial targeted biopsy, histopathological analyses, laser capture microdissection, and cluster analysis, and drafted the manuscript as a part of his doctoral thesis, with the assistance of the coauthors. NA prepared histological specimens and carried out laser capture microdissection. TK and TS carried out RNA extraction, the amplification, and a cDNA microarray. HY gave critical suggestions concerning orthopedics. NT, KY, SN, and SY carried out the clinical studies of each case and performed targeted biopsy of synovial tissues with the informed consent of the patients. MN conceived of the study, participated in its design and coordination, and is the corresponding author. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
A PDF showing a dendrogram of two-dimensional hierarchical clustering analysis of 1,035 genes among 12 patients with early rheumatoid arthritis (RA), not including the cases with long-standing RA. On the horizontal axis, 12 samples from early RA are clustered into two major groups. The results were similar to those shown in Fig. 3. This may indicate that there was no influence of the cases with long-standing RA in the cluster analysis.
Click here for file
Additional File 2
A PDF file showing the results of RT-PCR of multilayered lining tissues. (A) Signal intensity of candidate genes in microarray of the four cases of early RA; (B) their RT-PCR results. The expression levels of these genes themselves seemed to be well correlated in the two assays.
Click here for file
Additional File 3
A PDF file showing dendrograms of two-dimensional hierarchical clustering analysis with two different similarity measures and with two kinds of cutoff value for signal intensities among 18 samples from the 16 cases of rheumatoid synovitis. (Similarity measures: Euclidean distance and Pearson correlation coefficient. Cutoff value for signal intensities: 10,000 and 20,000.) There was no major difference between them regarding the cases belonging to each group.
Click here for file
Acknowledgements
We wish to thank Dr Herbert M Schulman for critically reviewing the manuscript.
Figures and Tables
Figure 1 Laser capture microdissection of synovial lining regions with normal-like lining or multilayered lining. (a,d)before microdissection; (b,e) after tracing around the lining regions together with the intimal lining layer, using a laser microbeam; (c,f) catapulted into a microcentrifuge tube by the micromanipulator with a single, precisely aimed laser shot.
Figure 2 Histopathological features of synovium in patients with early (E) or long-standing (L) rheumatoid arthritis. (E-02) The proliferation of synovial lining cells resulted in fewer than four layers. There is diffuse infiltration of macrophages in the sublining regions. (E-07) The proliferative lining layer shows a typical palisading structure of the intimal lining layer. (E-12) The specimen shows proliferation of synovial lining cells, in places to more than five layers, associated with a typical palisading structure and several non-foreign-body-type giant cells. The lesions manifest underlying proliferation of blood vessels at the arteriole level, associated with many cell infiltrates composed of lymphocytes and plasma cells in the sublining regions. There are foci of lymphocyte aggregates, close to postcapillary venules, resembling lymphoid follicles, but lacking germinal centers. (L-01) In contrast to E-12, there are lymphoid follicles with germinal centers.
Figure 3 Dendrogram of two-dimensional hierarchical clustering analysis of 1,035 genes from patients with rheumatoid synovitis. Red represents relative expression greater than the median expression level among all samples, and green represents relative expression lower than the median expression level. The color intensity represents the magnitude of the deviation from the median. Black indicates unchanged expression. On the horizontal axis, 18 samples from rheumatoid synovitis were clustered into two major groups. On the vertical axis, the 1,035 genes were clustered in different branches according to similarities in their relative expression ratios.
Table 1 Characteristics of studied patients with early (E) and long-standing (L) rheumatoid arthritis (RA)
Patient Age Sex Disease duration ACR criterion nos. fulfilleda Number of samples Macroscopic signs of synovitis
With early RA
E-01 51 F 11 months 1, 2, 3, 4 13 Vi, Ve
E-02 50 F 2 months 1, 2, 3, 4, 6 8 Vi, Ve
E-03 34 F 4 months 1, 2, 3, 4, 7 8 Vi, Ve
E-04 34 F 3 months 1, 2, 3, 4, 6, 7 13 Vi, Ve
E-05 77 F 2 months 1, 2, 3, 4, 7 11 Vi
E-06 50 M 4 months 1, 2, 3, 4 11 Vi, Ve
E-07 37 F 7 months 1, 2, 3, 4, 6 6 Ve
E-08 61 F 2 months 1, 2, 3, 4 7 Vi
E-09 75 F 4 months 1, 4, 6 12 Vi, Ve, Gr
E-10 25 F 12 months 1, 2, 3, 4 12 Vi, Ve, Gr
E-11 54 M 12 months 1, 2, 3, 4, 6 11 Ve
E-12 60 F 4 months 1, 2, 3, 4, 6 13 Vi, Ve, Gr
With long-standing RA
L-01 54 M 9 years 1, 2, 3, 4, 6, 7 6 Vi, Ve, Gr
L-02 77 M 5 years 1, 2, 3, 4, 5, 6, 7 8 Vi, Ve, Gr
L-03 54 F 7 years 1, 2, 3, 4, 6, 7 6 Vi, Ve
L-04 55 F 3 years 1, 2, 3, 4, 6, 7 11 Vi, Ve, Gr
aACR (American College of Rheumatology) criteria: 1, morning stiffness; 2, arthritis of three or more joint areas; 3, arthritis of hand joints; 4, symmetric arthritis; 5, rheumatoid nodules; 6, serum rheumatoid factor; 7, radiographic changes. F, female; Gr, granulation; M, male; Ve, increased number of vessels; Vi, villi.
Table 2 Histological scores in patients with early (E) and long-standing (L) rheumatoid arthritis (RA)
Group I Group II
Histological feature L-01 L-04 L-02 E-01 E-10 E-04 L-03 E-06 E-12 E-09 E-03 E-02 E-08 E-07 E-05 E-11
Proliferation of synovial cells 3 2 1 1 1 2 2 2 2 2 1 1 2 1 2 1
1.80 ± 0.63 (1.67 ± 0.52) 1.33 ± 0.52
Typical palisading 3 3 3 2 2 2 2 1 2 3 1 1 2 2 2 0
2.30 ± 0.68*(2.00 ± 0.63) 1.33 ± 0.82
Non-foreign-body giant cells 2 3 3 1 2 1 1 1 2 1 1 3 1 2 0 0
1.70 ± 0.82 (1.33 ± 0.52) 1.17 ± 0.48
Lymphoid cell infiltration 3 1 3 0 2 1 2 1 1 2 0 0 0 0 0 0
1.60 ± 0.97† (1.17 ± 0.75*) 0.00 ± 0.00
Plasma cell infiltration 3 3 3 0 3 2 3 1 3 3 0 0 1 0 0 0
2.40 ± 1.08† (2.00 ± 1.27*) 0.17 ± 0.41
Neovascularization 2 2 2 2 2 2 3 2 2 3 3 3 2 2 1 3
2.20 ± 0.42 (2.17 ± 0.41) 2.33 ± 0.82
Mesenchymoid transformation 1 1 2 0 0 0 1 0 0 3 0 0 0 0 0 0
0.80 ± 1.03 (0.50 ± 1.23) 0.00 ± 0.00
Fibrinoid necrosis 1 3 2 0 0 0 1 0 1 2 0 0 1 0 1 0
1.00 ± 1.05 (0.50 ± 0.84) 0.33 ± 0.52
Total 18 18 19 6 12 10 15 8 13 19 6 8 9 7 6 4
13.80 ± 4.76† (11.33 ± 4.37*) 6.67 ± 1.75
The value in the upper row is the histological score of each case. More than 6 samples were taken from each patient for the feature studied. The value in the lower row is the mean ± standard deviation for the group. Values in parentheses (group I) are those for only the patients with early RA. †P <0.01, *P <0.05 versus group II on the Mann–Whitney test. ACR, American College of Rheumatology.
Table 3 Comparison of the expression of selected genes in two groups of patients with rheumatoid arthritisa
Candidate gene Group Ib Group IIb Q (%) P
c
A – Expressed at higher levels in group II than in group I
CASP9 -0.029 ± 0.018 0.020 ± 0.014 0.25 <0.001
PIG11 -0.025 ± 0.021 0.024 ± 0.017 0.25 <0.001
CTSG -0.018 ± 0.024 0.031 ± 0.021 0.25 0.001
CSF2RB -0.020 ± 0.024 0.019 ± 0.021 0.61 0.006
TNFRSF1A -0.023 ± 0.029 0.015 ± 0.016 0.77 0.010
IL10RB -0.021 ± 0.027 0.017 ± 0.021 0.90 0.012
B – Expressed at higher levels in group I than in group II
FN1 0.023 ± 0.018 -0.030 ± 0.019 0.25 <0.001
B2M 0.020 ± 0.023 -0.028 ± 0.022 0.25 0.001
SDC2 0.023 ± 0.024 -0.024 ± 0.021 0.25 0.001
CTSB 0.006 ± 0.019 -0.036 ± 0.026 0.84 0.002
STAT1 0.021 ± 0.031 -0.020 ± 0.014 1.19 0.008
ITGB2 0.021 ± 0.025 -0.018 ± 0.023 1.51 0.008
IFNGR2 0.019 ± 0.026 -0.016 ± 0.026 3.28 0.022
aStatistical analysis on microarray data was performed using the significance analysis of microarrays (SAM) method (see Materials and methods). Q, which is the lowest false discovery rate, was considered significant at less than 5%. It is similar to the familiar P value, but adapted to the analysis of a large number of genes. bMeans ± standard deviations for the group, using log-transformed and median-centered microarray data. cValues calculated by t-test. P <0.05 was considered significant.
==== Refs
Zvaifler NJ The immunopathology of joint inflammation in rheumatoid arthritis Adv Immunol 1973 16 265 336 4599390
Fassbender HG Histomorphological basis of articular cartilage destruction in rheumatoid arthritis Coll Relat Res 1983 3 141 155 6851475
Arnett FC Edworthy SM Bloch DA McShane DJ Fries JF Cooper NS Healey LA Kaplan SR Liang MH Luthra HS The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis Arthritis Rheum 1988 31 315 324 3358796
Schumacher HR Kitridou RC Synovitis of recent onset: a clinicopathologic study during the first month of disease Arthritis Rheum 1972 15 465 485 4117789
Konttinen YT Bergroth V Nordstrom D Koota K Skrifvars B Hagman G Friman C Hamalainen M Slatis P Cellular immunohistopathology of acute, subacute, and chronic synovitis in rheumatoid arthritis Ann Rheum Dis 1985 44 549 555 3875323
Kraan MC Versendaal H Jonker M Bresnihan B Post WJ Hart BA Breedveld FC Tak PP Asymptomatic synovitis precedes clinically manifest arthritis Arthritis Rheum 1998 41 1481 1488 9704648 10.1002/1529-0131(199808)41:8<1481::AID-ART19>3.0.CO;2-O
Goldenberg DL Cohen AS Synovial membrane histopathology in the differential diagnosis of rheumatoid arthritis, gout, pseudogout, systemic lupus erythematosus, infectious arthritis and degenerative joint disease Medicine (Baltimore) 1978 57 239 252 642792
Smeets TJM Dolhain RJ Breedveld FC Tak PP Analysis of the cellular infiltrates and expression of cytokines in synovial tissue from patients with rheumatoid arthritis and reactive arthritis J Pathol 1998 186 75 81 9875143 10.1002/(SICI)1096-9896(199809)186:1<75::AID-PATH142>3.0.CO;2-B
Hitchon CA El-Gabalawy HS Immune features of seronegative and seropositive arthritis in early synovitis studies Curr Opin Rheumatol 2002 14 348 353 12118166 10.1097/00002281-200207000-00003
Tak PP Smeets TJM Daha MR Kluin PM Meijers KA Brand R Meinders AE Breedveld FC Analysis of the synovial cell infiltrate in early rheumatoid synovial tissue in relation to local disease activity Arthritis Rheum 1997 40 217 225 9041933
Kikuchi T Daigo Y Katagiri T Tsunoda T Okada K Kakiuchi S Zembutsu H Furukawa Y Kawamura M Kobayashi K Expression profiles of non-small cell lung cancers on cDNA microarrays: Identification of genes for prediction of lymph-node metastasis and snsitivity to anti-cancer drugs Oncogene 2003 22 2192 2205 12687021 10.1038/sj.onc.1206288
He P Varticovski L Bowman ED Fukuoka J Welsh JA Miura K Jen J Gabrielson E Brambilla E Travis WD Identification of carboxypeptidase E and gamma-glutamyl hydrolase as biomarkers for pulmonary neuroendocrine tumors by cDNA microarray Hum Pathol 2004 35 1196 1209 15492986 10.1016/j.humpath.2004.06.014
Lindblad S Hedfors E Intraarticular variation in synovitis; Local macroscopic and microscopic signs of inflammatory activity are significantly correlated Arthritis Rheum 1985 28 977 986 2931087
Koizumi F Matsuno H Wakaki K Ishii Y Kurashige Y Nakamura H Synovitis in rheumatoid arthritis: Scoring of characteristic histopathological features Pathol Int 1999 49 298 304 10365848 10.1046/j.1440-1827.1999.00863.x
Nakatani K Fujii H Hasegawa H Terada M Arita N Ito MR Ono M Takahashi S Saiga K Yoshimoto S Endothelial adhesion molecules in glomerular lesions: association with their severity and diversity in lupus models Kidney Int 2004 65 1290 1300 15086468 10.1111/j.1523-1755.2004.00537.x
Zanders ED Goulden MG Kennedy TC Kempsell KE Analysis of immune system gene expression in small rheumatoid arthritis biopsies using a combination of subtractive hybridization and high-density cDNA arrays J Immunol Methods 2000 233 131 140 10648863 10.1016/S0022-1759(99)00126-X
Van der Pouw Kraan TCTM van Gaalen FA Kasperkovitz PV Verbeet NL Smeets TJM Kraan MC Fero M Tak PP Huizinga TWJ Pieterman E Rheumatoid arthritis is a heterogeneous disease; evidence for differences in the activation of the STAT-1 pathway between rheumatoid tissues Arthritis Rheum 2003 48 2132 2145 12905466 10.1002/art.11096
Firestein GS Evolving concepts of rheumatoid arthritis Nature 2003 423 356 361 12748655 10.1038/nature01661
Katsikis PD Chu CQ Brennan FM Maini RN Feldmann M Immunoregulatory role of interleukin 10 in rheumatoid arthritis J Exp Med 1994 179 1517 1527 8163935 10.1084/jem.179.5.1517
Firestein GS Zvaifler NJ Peripheral blood and synovial fluid monocyte activation in inflammatory arthritis; II. Low levels of synovial fluid and synovial tissue interferon suggest that gamma-interferon is not the primary macrophage activating factor Arthritis Rheum 1987 30 864 871 3115274
Firestein GS Zvaifler NJ How important are T cells in chronic rheumatoid synovitis?; II. T cell-independent mechanisms from beginning to end Arthritis Rheum 2002 46 298 308 11840432 10.1002/art.502
Hu X Herrero C Li WP Antoniv TT Falck-Pedersen E Koch AE Woods JM Haines GK IIIIvashkiv LB Sensitization of IFN-gamma Jak-STAT signaling during macrophage activation Nat Immunol 2002 3 859 866 12172544 10.1038/ni828
Larkin J 3rdJohnson HM Subramaniam PS Differential nuclear localization of the IFNGR-1 and IFNGR-2 subunits of the IFN-gamma receptor complex following activation by IFN-gamma J Interferon Cytokine Res 2000 20 565 576 10888113 10.1089/10799900050044769
Clemmensen I Holund B Andersen RB Fibrin and fibronectin in rheumatoid synovial membrane and rheumatoid synovial fluid Arthritis Rheum 1983 26 479 485 6340698
Scott DL Wainwright AC Walton KW Williamson N Significance of fibronectin in rheumatoid arthritis and osteoarthrosis Ann Rheum Dis 1981 40 142 153 7013719
Nikkari L Aho H Yli-Jama T Larjava H Jalkanen M Heino J Expression of integrin family of cell adhesion receptors in rheumatoid synovium Am J Pathol 1993 142 1019 1027 7682757
Rinaldi N Barth T Henne C Mechterscheimer G Moller P Synoviocytes in chronic synovitis in situ and cytokine stimulated synovial cells in vitro neo-express alpha 1, alpha 3 and alpha 5 chains of beta 1 integrins Virchows Arch 1994 425 171 180 7524977 10.1007/BF00230354
Ruoslahti E Integrins J Clin Invest 1991 87 1 5 1985087
Lin TH Rosales C Mondal K Bolen JB Haskill S Juliano RL Integrin-mediated tyrosine phosphorylation and cytokine message induction in monocytic cells J Biol Chem 1995 270 16189 16197 7541794 10.1074/jbc.270.27.16189
Kamogawa J Terada M Mizuki S Nishihara M Yamamoto H Mori S Abe Y Morimoto K Nakatsuru S Nakamura Y Arthritis in MRL/lpr mice is under the control of multiple gene loci with an allelic combination derived from the original inbred strains Arthritis Rheum 2002 46 1067 1074 11953986 10.1002/art.10193
Oishi H Miyazaki T Mizuki S Kamogawa J Lu LM Tsubaki T Arita N Ono M Yamamoto H Nose M Accelerating effect of an MRL gene locus on the severity and onset of arthropathy in DBA/1 mice Arthritis Rheum 2005 52 959 966 15751084 10.1002/art.20956
Nishihara M Terada M Kamogawa J Ohashi Y Mori S Nakatsuru S Nakamura Y Nose M Genetic basis of autoimmune sialadenitis in MRL/lpr lupus-prone mice: additive and hierarchical properties of polygenic inheritance Arthritis Rheum 1999 42 2616 2623 10616009 10.1002/1529-0131(199912)42:12<2616::AID-ANR16>3.0.CO;2-O
Trabandt A Gay RE Fassbender HG Gay S Cathepsin B in synovial cells at the site of joint destruction in rheumatoid arthritis Arthritis Rheum 1991 34 1444 1451 1953821
Huet G Flipo RM Colin C Janin A Hemon B Collyn-d'Hooghe M Lafyatis R Duquesnoy B Degand P Stimulation of the secretion of latent cysteine proteinase activity by tumor necrosis factor alpha and interleukin-1 Arthritis Rheum 1993 36 772 780 8507218
Lemaire R Huet G Zerimech F Grard G Fontaine C Duquesnoy B Flipo RM Selective induction of the secretion of cathepsins B and L by cytokines in synovial fibroblast-like cells Br J Rheumatol 1997 36 735 743 9255106 10.1093/rheumatology/36.7.735
Cunnane G FitzGerald O Hummel KM Youssef PP Gay RE Gay S Bresnihan B Synovial tissue protease gene expression and joint erosions in early rheumatoid arthritis Arthritis Rheum 2001 44 1744 1753 11508424 10.1002/1529-0131(200108)44:8<1744::AID-ART309>3.0.CO;2-K
McGettrick AF Barnes RC Worrall DM SCCA2 inhibits TNF-mediated apoptosis in transfected Hela cells; the reactive centre loop sequence is essential for this function and TNF-induced cathepsin G is a candidate target Eur J Biochem 2001 268 5868 5875 11722574 10.1046/j.0014-2956.2001.02535.x
Zhou Q Salvesen GS Activation of pro-caspase-7 by serine proteases includes a non-canonical specificity Biochem J 1997 324 361 364 9182691
Biggs JR Yang J Gullberg U Muchardt C Yaniv M Kraft AS The human brm protein is cleaved during apoptosis: the role of cathepsin G Proc Natl Acad Sci USA 2001 98 3814 3819 11259672 10.1073/pnas.071057398
Gay S Gay RE Koopman WJ Molecular and cellular mechanisms of joint destruction in rheumatoid arthritis: two cellular mechanisms explain joint destruction? Ann Rheum Dis 1993 52 S39 47 8481058
Muller-Ladner U Kriegsmann J Franklin BN Matsumoto S Geiler T Gay RE Gay S Synovial fibroblasts of patients with rheumatoid arthritis attach to and invade normal human cartilage when engrafted into SCID mice Am J Pathol 1996 149 1607 1615 8909250
Firestein GS Invasive fibroblast-like synoviocytes in rheumatoid arthritis; passive responders or transformed aggressors? Arthritis Rheum 1996 39 1781 1790 8912499
Lafyatis R Remmers EF Roberts AB Yocum DE Sporn MB Wilder RL Anchorage-independent growth of synoviocytes from arthritic and normal joints; stimulation by exogenous platelet-derived growth factor and inhibition by transforming growth factor-beta and retinoids J Clin Invest 1989 83 1267 1276 2784799
Yamanishi Y Firestein GS Pathogenesis of rheumatoid arthritis: the role of synoviocytes Rheum Dis Clin North Am 2001 27 355 371 11396097
Yonish-Rouach E The p53 tumour suppressor gene: a mediator of a G1 growth arrest and of apoptosis Experientia 1996 52 1001 1007 8917731
Shen Y White E p53-dependent apoptosis pathways Adv Cancer Res 2001 82 55 84 11447765
Zhu J Jiang J Zhou W Zhu K Chen X Differential regulation of cellular target genes by p53 devoid of the PXXP motifs with impaired apoptotic activity Oncogene 1999 18 2149 2155 10321740 10.1038/sj.onc.1202533
Tak PP Zvaifler NJ Green DR Firestein GS Rheumatoid arthritis and p53: how oxidative stress might alter the course of inflammatory diseases Immunol Today 2000 21 78 82 10652465 10.1016/S0167-5699(99)01552-2
Firestein GS Nguyen K Aupperle KR Yeo M Boyle DL Zvaifler NJ Apoptosis in rheumatoid arthritis; p53 overexpression in rheumatoid arthritis synovium Am J Pathol 1996 149 2143 2151 8952546
Tak PP Smeets TJM Boyle DL Kraan MC Shi Y Zhuang S Zvaifler NJ Breedveld FC Firestein GS p53 overexpression in synovial tissue from patients with early and longstanding rheumatoid arthritis compared with patients with reactive arthritis and osteoarthritis Arthritis Rheum 1999 42 948 953 10323450 10.1002/1529-0131(199905)42:5<948::AID-ANR13>3.0.CO;2-L
Yamanishi Y Boyle DL Rosengren S Green DR Zvaifler NJ Firestein GS Regional analysis of p53 mutations in rheumatoid arthritis synovium Proc Natl Acad Sci USA 2002 99 10025 10030 12119414 10.1073/pnas.152333199
McCarty DJ O'Duffy JD Pearson L Hunter JB Remitting seronegative symmetrical synovitis with pitting edema JAMA 1985 254 2763 2767 4057484 10.1001/jama.254.19.2763
| 15987484 | PMC1175033 | CC BY | 2021-01-04 16:02:37 | no | Arthritis Res Ther. 2005 Apr 25; 7(4):R825-R836 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1751 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17521598748610.1186/ar1752Research ArticleSomatic mutations in the mitochondria of rheumatoid arthritis synoviocytes Da Sylva Tanya R [email protected] Alison [email protected] Yvonne 1Keystone Edward [email protected] Gillian E [email protected] Department of Biology, York University, Toronto, Ontario, Canada2 The Wellesley Toronto Arthritis and Immune Disorder Research Centre, University Health Network, Toronto, Ontario, Canada3 Department of Medicine, University of Toronto, Mount Sinai Hospital, Toronto, Ontario, Canada2005 28 4 2005 7 4 R844 R851 19 11 2004 22 12 2004 29 3 2005 31 3 2005 Copyright © 2005 Da Sylva et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Somatic mutations have a role in the pathogenesis of a number of diseases, particularly cancers. Here we present data supporting a role of mitochondrial somatic mutations in an autoimmune disease, rheumatoid arthritis (RA). RA is a complex, multifactorial disease with a number of predisposition traits, including major histocompatibility complex (MHC) type and early bacterial infection in the joint. Somatic mutations in mitochondrial peptides displayed by MHCs may be recognized as non-self, furthering the destructive immune infiltration of the RA joint. Because many bacterial proteins have mitochondrial homologues, the immune system may be primed against these altered peptides if they mimic bacterial homologues. In addition, somatic mutations may be influencing cellular function, aiding in the acquirement of transformed properties of RA synoviocytes. To test the hypothesis that mutations in mitochondrial DNA (mtDNA) are associated with RA, we focused on the MT-ND1 gene for mitochondrially encoded NADH dehydrogenase 1 (subunit one of complex I – NADH dehydrogenase) of synoviocyte mitochondria from RA patients, using tissue from osteoarthritis (OA) patients for controls. We identified the mutational burden and amino acid changes in potential epitope regions in the two patient groups. RA synoviocyte mtDNA had about twice the number of mutations as the OA group. Furthermore, some of these changes had resulted in potential non-self MHC peptide epitopes. These results provide evidence for a new role for somatic mutations in mtDNA in RA and predict a role in other diseases.
==== Body
Introduction
Rheumatoid arthritis (RA) is a chronic inflammatory autoimmune disease. It is multigenic, possibly triggered by exposure to viruses or bacteria, and, it is expected, other environmental stimuli. Consistent with this concept is the strong genetic association with the HLA-DR allele that contains a QK/RAA amino acid motif in its third hypervariable region, namely several alleles of the HLA-DRβ1 gene. The precise role of HLA-DR in pathogenesis is unknown, although its role in antigen presentation is the most obvious [1]. In vitro T-cell proliferation assays using the susceptible major histocompatibility complex (MHC) alleles has led to the discovery of a multiplicity of putative peptide autoantigens including collagen type II, cartilage link protein, heat shock proteins, and aggrecan [1].
There are nonimmune components to RA. RA synovial fibroblasts have many features of transformed cells – including the expression of oncogenes – and they have been shown to invade and destroy cartilage in the absence of T cells [2,3]. The acquisition of these transformed characteristics is thought to be aided by increased somatic mutations caused by reactive oxygen species (ROS) and reactive nitrogen species (RNS) produced endogenously within the inflamed joint [4]. Other studies linking ROS and RNS damage to decreased apoptosis have found ROS-associated damage to p53. The mutated p53 was a dominant negative, suggesting that p53 mutations help protect pathogenic cells from apoptosis [5-7].
Mitochondrial DNA (mtDNA) damage may complement damage to nuclear regulatory genes and have a causative role in the transformation of RA synovial cells. There is limited and sometimes contradictory evidence available concerning the ability of mtDNA mutations to lead to increased or decreased apoptosis [8]. Alterations of mtDNA are now being found in many tumor types and there is evidence that these mutations may contribute to the progression of human cancer [9,10].
There is growing evidence that somatic mutations within protein-coding genes of mtDNA may be recognized by the immune system: damaged mtDNA results in increased expression of MHC class I; and both MHC class I and class II can present mitochondrial peptides [11,12]. Mutated mitochondrial peptides in resident cells may, therefore, be aiding in the recruitment of immunological factors such as cytotoxic T cells to the RA joint.
Complex I – NADH (reduced nicotinamide-adenine dinucleotide) dehydrogenase is exceptionally susceptible to defects due to mtDNA mutations, because it has the most subunits encoded by mtDNA. Cells with complex I defects have also been shown to produce a higher amount of superoxide in vivo [8,13]. Therefore, defects in complex I may help perpetuate a vicious cycle of oxidative damage. The murine homologue of subunit 1 of complex I – NADH dehydrogenase (mtND1) plays a critical role in self recognition. The maternally transmitted antigen of rats and mice is the product of a class I molecule that presents the maternal transplantation factor derived from the amino terminus of mtND1 [14]. These findings provide evidence that antigenic peptides of human mtND1 may be displayed and recognized by the immune system.
To test the hypothesis that mutations in mitochondria play a role in RA, we examined the MT-ND1 gene of RA synoviocytes. As a control we chose synoviocytes from patients with osteoarthritis (OA). This disease was chosen because it is primarily a noninflammatory syndrome that is not thought to be directly dependent on the immune system. RA synoviocyte mtRNA had about twice the number of mutations as the OA group, revealing a greater mutational burden in RA. Furthermore, some of these changes resulted in changes that were potential non-self MHC peptide epitopes.
Materials and methods
RNA extraction from tissue and fibroblast lines
The protocol for the use of human tissues was approved by ethics review committees at the University Health Network and St Michael's Hospital, Toronto, Canada. Synovial tissues were obtained from RA and OA patients at the time of arthroplasty. The patients were not chosen by any criterion other than disease diagnosis. A portion of each sample was added to Trizol (Sigma Aldrich, St. Louis, MO, USA) and stored at -80°C until it was processed according to the manufacturer's instructions. Synovial fibroblast lines derived from the synovial tissue were established as previously described [15]. The fourth passage was used for all RA and OA lines. Cells were maintained in OptiMEM (Invitrogen Life Technologies, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum and 1% antibiotic–antimycotic. They were cultured at 37°C in a humidified chamber containing 95% air, 5% CO2.
RT-PCR and sequencing
Total RNA extracts from the fibroblasts and tissue of RA and OA patient samples were amplified using RT-PCR. This was a two- step protocol using the materials and methods included with the DuraScript RT-PCR Kit (Sigma Aldrich). In brief, first-strand cDNA was generated using 50 ng of total RNA, random nonamers for extension primers, and enhanced avian myeloblastosis virus (AMV) reverse transcriptase. Three PCR reactions were then performed using 5 μl first-strand cDNA in each 50 μl PCR. The primer pairs and amplification conditions are described in Table 1 and have been published previously [16].
Direct sequencing of PCR products does not detect low levels of heteroplasmy; therefore the PCR fragments were cloned into a TA vector (using protocols and materials provided in the TOPO TA Cloning® Kit for Sequencing with One Shot® TOP10 Chemically Competent E. coli; Invitrogen, Carlsbad, CA, USA). Approximately ten colonies from each patient sample were chosen and sequenced using T3 and T7 primers. To rule out sequencing errors, only areas of complete identity (between the T3 and T7 sequence) were aligned with the mitochondrial Anderson Reference Sequence [17]. Nucleotide changes from the reference sequence were recorded and then entered into the online program MitoAnalyzer (National Institute of Standards and Technology, Gaithersburg, MD, USA; ; 2000) which displays the sequence and any amino acid changes resulting and the position number affected.
The same amplification and sequencing procedure as above was followed using PCR primers (Table 1) and conditions previously published for a nuclear gene, that for dihydrolipamide dehydrogenase (DLD) [18,19]. Nucleotide changes from the NCBI (National Center for Biotechnology Information) reference sequence (gi:5016092) were recorded and corresponding amino acid changes determined.
To control for errors induced by PCR and cloning/transformation, three plasmids containing cloned fragments were amplified and sequenced as above (approximately 18,000 bp in both directions). A methodological error frequency was calculated (0.00095 errors/bp for total mutational burden and 0.00063 errors/bp for expressed mutational burden) and subtracted from the final mutational burden data before statistical analyses. Throughout this report, the data presented are corrected for methodological error.
Mutational burden comparisons
The mutational burden of OA and RA patients was defined as the number of mutations identified within that group divided by the total number of base pairs analyzed. This was then further separated into two measurements, total mutational burden (all mutations) and expressed mutational burden (the number of amino acid changes in the MT-ND1 cDNA from each amplified region; see Fig. 1). All patients sequenced with the first set of MT-ND1 primers (1A) had a deletion at nucleotide 3107. The NCBI reference sequence (gi:17981852) also shows a deletion at this position when compared with the Anderson sequence (where there is a C) [20]. Since a C at this position is rarer than the 3107 deletion, the deletion was not included when calculating mutational burden. All patients sequenced also had a nucleotide substitution (T to C) at position 1081 in the DLD gene and, as above, this mutation was also not included in the calculation of mutational burden. Mutational burden was compared between RA and OA for each fragment within the MT-ND1 amplification region and for the amplified DLD region, using a two-tailed Fisher's exact test.
Known polymorphisms analysis
Reported mtDNA polymorphisms were subtracted from the total and expressed mutational burden and the values were reanalyzed, as above. Published polymorphisms were gathered from Mitomap and a table of the known polymorphisms found among the patient data is given in Supplementary Table 1.
Epitope prediction
MHC epitope prediction algorithms were used to search for possible epitope regions within MT-ND1 for RA susceptible HLA alleles . The algorithm, MHCPred, used published IC50 (median inhibitory concentration) values from radioligand competition assays to develop a predictive algorithm [21]. Given an amino acid sequence, the program predicts the peptides likely to bind to the MHC complex (epitopes) and their IC50 values [21]. Peptides with a -logIC50 of more than 6.5 are predicted to be binders [21].
Results
Mutational burden in OA and RA
OA was used as a nonimmunological-based disease control for the study. We examined both synovial tissue (patients OA227, OA315, OA320, and OA324) and synovial fibroblast lines derived from synovial tissue (patients OA302 and OA304). Approximately 37 kbp were sequenced from OA tissue and 18 kbp from OA fibroblasts, with 67 (2.1/kbp) mutations and 38 (1.8/kbp) mutations found respectively (Table 2; Fig. 2).
For the RA analyses, we also examined both synovial tissue (patients RA301C, RA316, RA317, and RA325) and synovial fibroblast lines (patients RA307 and RA313). Approximately 30 kbp were analyzed from fibroblasts and 18 kbp from tissue, with 101 (3.3/kbp) mutations and 60 (3.4/kbp) mutations found respectively (Table 2; Fig. 2). Comparative analyses of the OA and RA patient data demonstrate significantly more changes per base pair in RA patients than OA (Table 2, Fisher's exact P value, ρ < 0.05), whether derived from tissue or fibroblasts. There may be subgroups within the RA or OA set as evidenced by the mutation frequencies between RA and OA patients (Fig. 2). Further studies, with a more detailed patient history, may help correlate mitochondrial mutations to disease factors such as age of onset and response to treatment.
Amino acid (nonsynonymous) changes
The mutations in the gene for MT-ND1 will result in mtND1 protein subunit changes if the mutations created amino acid changes. mtND1 amino acid changes were found in both OA and RA samples. In OA, 7 kbp were analyzed from fibroblast RNA and 16 kbp from tissue RNA. The OA fibroblasts had an expressed mutational burden of 1.7 amino acid changes per kilobase pair (12 changes), and tissue 0.63 amino acid changes per kilobase pair (10 changes) (Table 2; Fig. 2). In RA, 12.4 kbp of the MT-ND1 gene were analyzed from fibroblasts and 6.4 kbp from tissue. The RA fibroblasts had an expressed mutational burden of 2.3 amino acid changes per kilobase pair (28 changes), and tissue, 2.5 amino acid changes per kilobase pair (16 changes) (Table 2; Fig. 2).
Thus, there are more amino-acid-changing mutations in RA patients' MT-ND1 gene in synovial tissue (P < 0.5) (Table 2) than in OA synovial tissue. Although there are more mutations in RA than OA cultured fibroblasts, the expressed mutation frequency is not statistically different (2.5 vs 1.7 amino acid changes per kilobase pair, respectively).
Nuclear DNA mutational burden
A nuclear gene was analyzed to determine whether it, too, had increased mutations in RA, and thus reveal whether the changes in mutational frequency were specific to mitochondria. The gene, DLD, was chosen because its product, dihydrolipoamide dehydrogenase, is a nuclear-encoded mitochondrial subunit peptide, constitutively expressed in all cell types [22]. Mutations were found, as above, in both RA and OA patients. The total mutational burden was high (approximately 2 mutations per kilobase pair); however, there were no significant differences between the RA and OA patient classes (Table 3).
Epitope prediction and somatic mutations
Several findings suggest that the immune system may aid in the destruction of cells containing mtDNA mutations [11]. Peptides altered by somatic mutations would be presented by MHC and may be recognized as non-self. Searches for possible epitopes in mtND1 led to 76 possible epitopes; of these, 15 were altered by somatic mutations in the RA and OA patients' mitochondrial samples (data not shown). We chose to further analyze the 1B amplified fragment of fibroblasts in more detail because it is totally mRNA-derived (see Fig. 1). We searched all six predicted HLA-DRβ1*0101 epitopes and the ten epitopes with highest -logIC50 (pIC50) values for HLA-DRβ*0401 within the 1B fragment for changes. Changes in epitope regions were noted, and the new mutated epitope was submitted to the MHCPred program for prediction of pIC50 values (Table 4).
Although RA fibroblasts did not have a statistically higher expressed mutational burden than OA fibroblasts, out of the 16 epitopes investigated, 5 were changed in RA and only 1 was changed in OA. The new (changed) epitopes were analyzed by the same predictive program and all the new RA epitopes fell above the pIC50 cutoff value of 6.5M while the changed OA epitope fell below this cutoff (Table 4).
Discussion
These studies revealed that mtDNA somatic mutations were present in the synovium of RA patients at a higher frequency than OA controls. We considered the causes of the somatic mutations (ROS plus selection) as well as the effect these mutations may be having on the etiology and pathogenesis of RA.
ROS exposure and survival advantage
Exposure to mutagens, such as ROS, can damage both nuclear and mitochondrial DNA. The mtDNA is in close physical proximity to the free-radical-producing process of oxidative phosphorylation and lacks the protective nucleosome structure found in nuclear DNA [23]. Additionally, there is limited ability within mitochondria to repair DNA damage. Together, these attributes make mtDNA highly prone to damage by ROS produced by both mitochondria and exogenous sources [24].
ROS introduces mutations. If the mutations were in genes regulating cell survival, cells that would otherwise stop dividing and die (from DNA damage) may instead proliferate [4]. Insufficient apoptosis of resident synoviocytes and inflammatory cells has been thought to contribute to the persistence of RA [7]. A higher incidence of lymphoma is also well documented in RA, and somatic mutations may lead to enhancement of the aggressive nature of pathogenic cells [25]. For instance, p53 mutations have been found in RA synovial tissues. Their mutations would be predicted to give a growth advantage to the mutated cells, leading to monoclonal expansion [6]. While these mutations are most likely a consequence of inflammation and not the cause of RA, they would be expected to affect disease progression [1].
The NADPH oxidase system of neutrophils and monocytes produces ROS upon activation [26]. Accumulation of these cells within the inflamed joint and the subsequent increase in ROS may be partially responsible for the increased mtDNA mutational burden of RA patients. There are two situations in which nuclear mutations would be expected to occur at greater frequency in RA patients than in OA controls: first, if random processes (ROS from the NADPH oxidase) were the sole cause of the elevated frequency of mtDNA mutations; and second, if the nuclear mutation is conferring an RA-specific characteristic (survival advantage) on the synoviocyte. Examples of the latter instance are the p53 mutations (noted above) that were not found in peripheral blood from RA patients or joint tissue from OA patients. It is thought that p53 is randomly mutated during chronic inflammation by oxygen radicals. Certain mutations within p53 then confer a survival advantage to the synoviocytes, giving them 'transformed' characteristics and participating in the perpetuation of disease [5]. The high frequency of mutations (approximately 2 mutations per kilobase pair) in both patient classes suggests there may be genotoxic stressors in both RA and OA synovia. However, RA synovial tissue and fibroblasts showed no significant increase in the randomly chosen nuclear gene over OA controls. This suggests that the nuclear gene sequenced was not contributing to the progression of RA and that random mutations through exogenous ROS cannot, alone, explain the increase in RA mutational burden found in the mtDNA.
RA is a member of a large class of inflammatory autoimmune diseases. The presence of exogenous ROS produced by neutrophils and monocytes may also be contributing to the pathology of other inflammatory autoimmune diseases. Such a corollary suggests it may be of interest to investigate mtDNA within other inflammatory autoimmune diseases such as systemic lupus erythematosus.
Altered mitochondrial proteins as non-self
Several findings suggest that the immune system may aid in the destruction of cells containing mtDNA mutations [11]. Peptides altered by somatic mutations would be presented by MHC and may be recognized as non-self. Without a similar analysis of mitochondrial RNA from maternal relatives, it is impossible to say, with certainty, whether all the changes noted were truly somatic mutations. If somatic mutations were changing recognition of mitochondrial peptides from self to non-self, then any inherited changes would be irrelevant. Although we were unable to obtain samples from maternal relatives of patients, there does exist a database of known mitochondrial polymorphisms . When all known polymorphisms were subtracted from the data, the statistical significance of the findings did not change, either for all mutations or just nonsynonymous changes (Table 2).
There are no known data that address the antigenic nature of mitochondrial proteins in RA. However, there is evidence for the involvement of mitochondrial antibodies in another form of arthritis, polymylagia rheumatica. Temporal or giant-cell arteritis is an inflammatory large-vessel disease associated in many patients with polymyalgia rheumatica, and while the etiology of giant-cell arteritis/polymyalgia rheumatica is unclear, there is evidence to support the role of immune mechanisms in its pathogenesis, including the discovery of five autoantigens in patients with the disease [27,28]. Moreover, one of these autoantigens is a mtDNA-encoded subunit of complex IV (cytochrome c oxidase subunit II [28]), implicating mtDNA-derived proteins in autoimmune disease.
Our studies predicted numerous regions within mtND1 that may be possible epitopes for HLA-DRβ1*0101 and HLA-DRβ1*0401 (RA-associated HLAs). Five of these possible epitopes were mutated in RA patients and one was mutated in OA. The new (mutated) peptides were analyzed and found to still be possible epitopes for the RA patients, but the OA patient's mutation caused the mutated epitope to fall below the cutoff value for HLA binding (Table 4). Therefore, peptides from mitochondria have the potential to be presented by MHC II, and somatic mutations may alter the peptide such that it is recognized as non-self. As a result, recognition by the immune system of mitochondrial peptides may be aiding in the recruitment of T cells and inflammatory factors, helping to sustain the synovial inflammation characteristic of RA.
Conclusion
This study demonstrates, for the first time, that mtDNA somatic mutations are present in high frequency in the synovia of RA patients. There are two possible effects of somatic mitochondrial mutations on RA. These somatic mutations may be influencing cellular function, aiding in the acquisition of transformed properties of RA synoviocytes. Second, somatic mutations in peptides displayed by MHC may also be causing an immune reaction, which would further the destructive immune infiltration of the RA joint. The immune system may be primed against these altered peptides because of mimicry with bacterial homologues. Either of these processes would aid in progression of the disease, and earlier immune recognition of mitochondrial peptides may also play a causative role in RA.
Abbreviations
bp = base pairs; IC50 = median inhibitory concentration; MHC = major histocompatibility complex; mtDNA = mitochondrial DNA; NADH = reduced nicotinamide-adenine dinucleotide; NCBI = National Center for Biotechnology Information; OA = osteoarthritis; RA = rheumatoid arthritis; RNS = reactive nitrogen species; ROS = reactive oxygen species.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
TD participated in the design of the study, performed some of the molecular genetic studies/analysis, and wrote the manuscript. AC performed the RNA extraction from synovial tissue and established the fibroblast lines. YM participated in the molecular genetic studies and analysis. EK procured the samples and helped in the analysis of data. GW participated in the design of the study, analysis of data, and writing of the manuscript. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
A Word file containing a table showing previously published polymorphisms (as of Mar 15, 2005) found within patient samples.
Click here for file
Acknowledgements
This study was funded by the Canadian Institutes of Health Research (to GW), The Younger Foundation, and the Lupus Society of Ontario (to GW and EK). We thank Dr E Bogochfor surgical samples, Ms K Griffith Cunningham for co-coordinating the tissue and blood collections, and Ms L Cunningham for expert technical assistance.
Figures and Tables
Figure 1 The three amplified and sequenced regions of mtDNA, corresponding to primers given in Table 1. tRNA-Gln is encoded on the negative (or light) strand of mtDNA. ND1, NADH-dehydrogenase subunit 1; ND2, NADH dehydrogenase subunit 2.
Figure 2 Mitochondrial mutational burden for OA and RA patients. Fibroblast data are given in red, tissue data in blue. kbp, kilobase pairs
Table 1 PCR primers and sequence start position for amplification of MT-ND1 and DLD
Primer name Start position Sequence 5'-3'
mtMT-ND1a
F1A 2995 TTGGATCAGGACATCCCGA
R1A 3645 ACGGCTAGGCTAGAGGTGG
F1B 3536 TTAGCTCTCACCATCGCT
R1B 4239 ATTGTAATGGGTATGGAGACA
F2 4184 TTCCTACCACTCACCCTAG
R2 4869 CATGTGAGAAGAAGCAG
DLDb
DLD – sense 417 ATGATGGAGCAGAAGAGTACTGCA
DLD – antisense 1088 TTTAGTTTGAAATCTGGTATTGAC
aSee Fig. 1 for position of primers. Both forward and reverse primers in addition to the specific nucleotide sequence have a corresponding M13 tag (M13F, 5'-TGTAAAACGACGGCCAGT- 3' ; M13R, 5'-CAGGAAACAGCTATGACC-3'); start position numbering represents location of 5' end corresponding to the Anderson Reference Sequence [17]. bStart position numbering represents location of the 5' end and corresponds to the DLD cDNA numbering system published by Pons and colleagues [19]. mt, mitochondrial.
Table 2 Mitochondrial mutational burden data of OA and RA patient synoviocyte tissue and cultured fibroblasts
Number of mutations Total mutational burden (mutations/kbp) OA vs RAa
Mitochondrial mutational burden Nucleotides sequenced Initial Published polymorphisms removed Initial Published polymorphisms removed Initial Published Polymorphisms removed
Total
Fibroblasts
OA 18489 38 20 2.055 1.082
RA 30503 101 65 3.311 2.131 ρ = 0.01 ρ = 6.9 × 10-3
Tissue
OA 37145 67 40 1.804 1.077
RA 17663 60 42 3.397 2.378 ρ = 4 × 10-4 ρ = 5.1 × 10-4
Expressed
Fibroblasts
OA 6956 12 7 1.725 1.006
RA 12394 28 26 2.259 2.098 ρ = 0.5 ρ = 0.10
Tissue
OA 15805 10 10 0.633 0.633
RA 6397 16 14 2.501 2.189 ρ = 6 × 10-4 ρ = 2.7 × 10-3
aTwo-tailed Fisher's exact test. kbp, kilobase pairs; OA, osteoarthritis; RA, rheumatoid arthritis.
Table 3 Nuclear mutational burden data of OA and RA patient tissue and cultured fibroblasts
Patients Nucleotides sequenced Number of total mutations Total mutational burden (mutations/kbp) Number of amino acid changes Expressed mutational burden (mutations/kbp) OA vs RAa
Fibroblasts
OA 10971 36 3.281 24 2.188
RA 7317 15 2.050 10 1.367 ρ = 0.2
Tissue
OA 18287 32 1.750 15 0.820
RA 24522 41 1.672 21 0.856 ρ = 0.2
aTwo-tailed Fisher's exact test. kbp, kilobase pairs; OA, osteoarthritis; RA, rheumatoid arthritis.
Table 4 Predicted epitopes for HLA DRB1*0101 and HLA DRB1*0401 which were changed by nonsynonymous mutations
Patient Amino acid start position Predicted core epitope (before mutation) Predicted -logIC50 (M) New epitope with amino acid changea New predicted -logIC50 (M)
HLA DRβ*0101
RA313 274 RTAYPRFRY 6.661 RTAHPRFRY 6.896
RA307 99 NLGLLFILA 6.51 SLGLLFILA 6.549
OA302 215 YAAGPFALF 6.681 YAAGPFALS 5.708
HLA DRβ*0401
RA307 259 FVTKTLLLT 7.329 FVAKTLLLT 7.316
RA313 88 PLPMPNPLV 7.093 PLPIPNPLV 6.946
RA307 93 NPLVNLNLG 7.09 NPLVNLSLG 7.148
aBold indicates amino acid changed by mutation; a -logIC50 value above 6.5 is considered to be a binder for prediction purposes. IC50, median inhibitory concentration.
==== Refs
Firestein GS Evolving concepts of rheumatoid arthritis Nature 2003 423 356 361 12748655 10.1038/nature01661
Muller-Ladner U Kriegsmann J Franklin BN Matsumoto S Geiler T Gay RE Gay S Synovial fibroblasts of patients with rheumatoid arthritis attach to and invade normal human cartilage when engrafted into SCID mice Am J Pathol 1996 149 1607 1615 8909250
Lafyatis R Remmers EF Roberts AB Yocum DE Sporn MB Wilder RL Anchorage-independent growth of synoviocytes from arthritic and normal joints. Stimulation by exogenous platelet-derived growth factor and inhibition by transforming growth factor-beta and retinoids J Clin Invest 1989 83 1267 1276 2784799
Tak PP Zvaifler NJ Green DR Firestein GS Rheumatoid arthritis and p53: how oxidative stress might alter the course of inflammatory diseases Immunol Today 2000 21 78 82 10652465 10.1016/S0167-5699(99)01552-2
Firestein GS Echeverri F Yeo M Zvaifler NJ Green DR Somatic mutations in the p53 tumor suppressor gene in rheumatoid arthritis synovium Proc Natl Acad Sci USA 1997 94 10895 10900 9380731 10.1073/pnas.94.20.10895
Yamanishi Y Boyle DL Rosengren S Green DR Zvaifler NJ Firestein GS Regional analysis of p53 mutations in rheumatoid arthritis synovium Proc Natl Acad Sci USA 2002 99 10025 10030 12119414 10.1073/pnas.152333199
Liu H Pope RM The role of apoptosis in rheumatoid arthritis Curr Opin Pharmacol 2003 3 317 322 12810199 10.1016/S1471-4892(03)00037-7
Chomyn A Attardi G MtDNA mutations in aging and apoptosis Biochem Biophys Res Commun 2003 304 519 529 12729587 10.1016/S0006-291X(03)00625-9
Maximo V Soares P Lima J Cameselle-Teijeiro J Sobrinho-Simoes M Mitochondrial DNA somatic mutations (point mutations and large deletions) and mitochondrial DNA variants in human thyroid pathology: a study with emphasis on Hurthle cell tumors Am J Pathol 2002 160 1857 1865 12000737
Lee HC Li SH Lin JC Wu CC Yeh DC Wei YH Somatic mutations in the D-loop and decrease in the copy number of mitochondrial DNA in human hepatocellular carcinoma Mutat Res 2004 547 71 78 15013701
Gu Y Wang C Roifman CM Cohen A Role of MHC class I in immune surveillance of mitochondrial DNA integrity J Immunol 2003 170 3603 3607 12646623
Kita H Lian ZX Van de Water J He XS Matsumura S Kaplan M Luketic V Coppel RL Ansari AA Gershwin ME Identification of HLA-A2-restricted CD8(+) cytotoxic T cell responses in primary biliary cirrhosis: T cell activation is augmented by immune complexes cross-presented by dendritic cells J Exp Med 2002 195 113 123 11781370 10.1084/jem.20010956
Pitkanen S Robinson BH Mitochondrial complex I deficiency leads to increased production of superoxide radicals and induction of superoxide dismutase J Clin Invest 1996 98 345 351 8755643
Loveland B Wang CR Yonekawa H Hermel E Lindahl KF Maternally transmitted histocompatibility antigen of mice: a hydrophobic peptide of a mitochondrially encoded protein Cell 1990 60 971 980 2317868 10.1016/0092-8674(90)90345-F
Scott BB Weisbrot LM Greenwood JD Bogoch ER Paige CJ Keystone EC Rheumatoid arthritis synovial fibroblast and U937 macrophage/monocyte cell line interaction in cartilage degradation Arthritis Rheum 1997 40 490 498 9082937
Taylor RW Taylor GA Durham SE Turnbull DM The determination of complete human mitochondrial DNA sequences in single cells: implications for the study of somatic mitochondrial DNA point mutations Nucleic Acids Res 2001 29 E74 11470889 10.1093/nar/29.15.e74
Anderson S Bankier AT Barrell BG de Bruijn MH Coulson AR Drouin J Eperon IC Nierlich DP Roe BA Sanger F Sequence and organization of the human mitochondrial genome Nature 1981 290 457 465 7219534 10.1038/290457a0
Liu TC Kim H Arizmendi C Kitano A Patel MS Identification of two missense mutations in a dihydrolipoamide dehydrogenase-deficient patient Proc Natl Acad Sci USA 1993 90 5186 5190 8506365
Pons G Raefsky-Estrin C Carothers DJ Pepin RA Javed AA Jesse BW Ganapathi MK Samols D Patel MS Cloning and cDNA sequence of the dihydrolipoamide dehydrogenase component human alpha-ketoacid dehydrogenase complexes Proc Natl Acad Sci USA 1988 85 1422 1426 3278312
Ingman M Kaessmann H Paabo S Gyllensten U Mitochondrial genome variation and the origin of modern humans Nature 2000 408 708 713 11130070 10.1038/35047064
Guan P Doytchinova IA Zygouri C Flower DR MHCPred: A server for quantitative prediction of peptide-MHC binding Nucleic Acids Res 2003 31 3621 3624 12824380 10.1093/nar/gkg510
Harris RA Bowker-Kinley MM Wu P Jeng J Popov KM Dihydrolipoamide dehydrogenase-binding protein of the human pyruvate dehydrogenase complex. DNA-derived amino acid sequence, expression, and reconstitution of the pyruvate dehydrogenase complex J Biol Chem 1997 272 19746 19751 9242632 10.1074/jbc.272.32.19746
Stevnsner T Thorslund T de Souza-Pinto NC Bohr VA Mitochondrial repair of 8-oxoguanine and changes with aging Exp Gerontol 2002 37 1189 1196 12470830 10.1016/S0531-5565(02)00142-0
Khrapko K Coller HA Andre PC Li XC Hanekamp JS Thilly WG Mitochondrial mutational spectra in human cells and tissues Proc Natl Acad Sci USA 1997 94 13798 13803 9391107 10.1073/pnas.94.25.13798
Wolfe F Michaud K Lymphoma in rheumatoid arthritis: the effect of methotrexate and anti-tumor necrosis factor therapy in 18,572 patients Arthritis Rheum 2004 50 1740 1751 15188349 10.1002/art.20311
Kroger H Miesel R Dietrich A Ohde M Altrichter S Braun C Ockenfels H Suppression of type II collagen-induced arthritis by N -acetyl-L-cysteine in mice Gen Pharmacol 1997 29 671 674 9352320 10.1016/S0306-3623(96)00570-8
Cid MC Font C Coll-Vinent B Grau JM Large vessel vasculitides Curr Opin Rheumatol 1998 10 18 28 9448986
Schmits R Kubuschok B Schuster S Preuss KD Pfreundschuh M Analysis of the B cell repertoire against autoantigens in patients with giant cell arteritis and polymyalgia rheumatica Clin Exp Immunol 2002 127 379 385 11876765 10.1046/j.1365-2249.2002.01751.x
| 15987486 | PMC1175034 | CC BY | 2021-01-04 16:02:37 | no | Arthritis Res Ther. 2005 Apr 28; 7(4):R844-R851 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1752 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17531598748510.1186/ar1753Research ArticleOsteoporosis in experimental postmenopausal polyarthritis: the relative contributions of estrogen deficiency and inflammation Jochems Caroline [email protected] Ulrika [email protected] Malin [email protected] Margareta 1margareta.verdrengh@ rheuma.gu.semargareta.verdrengh@ rheuma.guOhlsson Claes [email protected] Hans 1hans.carlsten@ rheuma.gu.sehans.carlsten@ rheuma.gu.se1 Department of Rheumatology and Inflammation Research at the Sahlgrenska Academy, Göteborg, Sweden2 Center for Bone Research at the Sahlgrenska Academy (CBS), Göteborg, Sweden2005 27 4 2005 7 4 R837 R843 18 2 2005 18 3 2005 1 4 2005 12 4 2005 Copyright © 2005 Jochems et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Generalized osteoporosis in postmenopausal rheumatoid arthritis (RA) is caused both by estrogen deficiency and by the inflammatory disease. The relative importance of each of these factors is unknown. The aim of this study was to establish a murine model of osteoporosis in postmenopausal RA, and to evaluate the relative importance and mechanisms of menopause and arthritis-related osteoporosis. To mimic postmenopausal RA, DBA/1 mice were ovariectomized, followed by the induction of type II collagen-induced arthritis. After the mice had been killed, paws were collected for histology, one femur for bone mineral density (BMD) and sera for analyses of markers of bone resorption (RatLaps; type I collagen cross-links, bone formation (osteocalcin) and cartilage destruction (cartilage oligomeric matrix protein), and for the evaluation of antigen-specific and innate immune responsiveness. Ovariectomized mice displayed more severe arthritis than sham-operated controls. At termination of the experiment, arthritic control mice and non-arthritic ovariectomized mice displayed trabecular bone losses of 26% and 22%, respectively. Ovariectomized mice with arthritis had as much as 58% decrease in trabecular BMD. Interestingly, cortical BMD was decreased by arthritis but was not affected by hormonal status. In addition, markers of bone resorption and cartilage destruction were increased in arthritic mice, whereas markers of bone formation were increased in ovariectomized mice. This study demonstrates that the loss of endogenous estrogen and inflammation contribute additively and equally to osteoporosis in experimental postmenopausal polyarthritis. Markers of bone remodeling and bone marrow lymphocyte phenotypes indicate different mechanisms for the development of osteoporosis caused by ovariectomy and arthritis in this model.
==== Body
Introduction
Rheumatoid arthritis (RA) is a common inflammatory joint disease with a prevalence of 0.5 to 1% [1]. RA is more common in women than in men, and the peak incidence in women coincides with the time of menopause [2]. There is evidence that the female sex hormone estrogen can influence both the incidence and the progression of RA. Exposure to oral contraceptives has been shown to reduce the risk of developing RA [3], and disease activity often decreases during pregnancy [4], when levels of female sex hormones are elevated. Recently, we reported beneficial effects of hormone replacement therapy in women with postmenopausal RA. Patients treated with hormone replacement therapy displayed increased bone mineral density (BMD), better clinical outcome, decreased erythrocyte sedimentation rate and elevated levels of serum hemoglobin as well as retarded progression of joint erosion [5].
RA is characterized by different skeletal manifestations including periarticular osteoporosis, bone erosions and generalized osteoporosis. The frequency of generalized osteoporosis in postmenopausal RA has been shown to be almost 50% [6,7], and these patients are at high risk for fractures. The bone loss in postmenopausal RA is believed to be caused by the combined effects of estrogen deficiency [8] and the inflammatory disease [9]. The relative importance of each of these two factors is not yet known.
Collagen-induced arthritis (CIA) is a well established murine model for human RA [10]. It has been shown that treatment with physiological doses of estradiol suppresses the disease progression in this model [11], whereas loss of endogenous estrogen by ovariectomy (OVX) leads to a more severe disease. OVX of mice leads to significant bone loss and is used as a model of postmenopausal osteopenia [12]. It has been demonstrated that OVX enhances the severity of arthritis and bone loss in CIA in rats, whereas exposure to estrogen suppresses it [13].
The aim of this study was to establish a murine model for studies of osteoporosis in postmenopausal RA, and to evaluate the relative importance and possible different mechanisms of estrogen deficiency versus joint inflammation for the induction of bone loss.
Materials and methods
Mice
The ethical committee for animal experiments at the University of Göteborg approved this study. Female DBA/1 mice (Taconic M&B A/S, Ry, Denmark) were kept, 5 to 10 animals to a cage, under standard environmental conditions and were fed with standard laboratory chow and tap water ad libitum.
Castration
OVX or sham operation was performed at 10 weeks of age. Ovaries were removed by using a midline incision of the skin, and flank incisions of the peritoneum. The skin incision was closed with metallic clips. Sham-operated animals had their ovaries exposed but not removed. Surgery was performed under Ketalar® (Pfizer AB, Täby, Sweden) and Domitor® (Orion Pharma, Espoo, Finland) anesthesia.
Induction and evaluation of arthritis
Nine days after surgery the mice were immunized with 100 μg of chicken type II collagen (CII; Sigma, St Louis, MO) dissolved in 0.1M acetic acid and emulsified with an equal volume of incomplete Freund's adjuvant (Sigma) supplemented with 0.5 mg/ml Mycobacterium tuberculosis (Sigma). A total volume of 100 μl was injected intradermally at the base of the tail (50 μl on each side). After 21 days mice received a booster injection in the same way using CII emulsified in incomplete Freund's adjuvant.
The animals were observed twice weekly for frequency and severity of arthritis. Severity was graded as described previously [14], scoring 1 to 3 in each paw (maximum of 12 points per mouse) as follows: 1, swelling or erythema in one joint; 2, swelling or erythema in two joints; 3, severe swelling of the entire paw or ankylosis.
Tissue collection and histological examination
At 45 days after immunization mice were anaesthetized with Ketalar®/Domitor®, bled, and killed by cervical dislocation. Sera were individually stored at -20°C until use. Paws and femurs were collected.
Paws were placed in 4% paraformaldehyde dissolved in water, decalcified, and embedded in paraffin. Sections were stained with eosin/hematoxylin and encoded before examination. In each animal the front and back of all four paws were graded separately on a scale 0 to 4 and divided by 2, with a maximum of 16 points per mouse, as follows: 1, synovial hypertrophy; 2, pannus, erosions of cartilage; 3, erosions of bone; 4, complete ankylosis.
Bone mineral density
One femur was subjected to a peripheral quantitative computed tomography (pQCT) scan with a Stratec pQCT XCT Research M, software version 5.4 B (Norland, Fort Atkinson, WI) at a resolution of 70 μm, as described previously [15]. Trabecular BMD was determined with a metaphyseal scan at a point 3% of the length of the femur from the growth plate. The inner 45% of the area was defined as the trabecular bone compartment. Cortical BMD was determined with a mid-diaphyseal scan, which contains only cortical bone.
Serological markers of bone and cartilage remodeling
As a marker of bone resorption, serum levels of fragments of type I collagen were assessed using a RatLaps ELISA kit (Nordic Bioscience Diagnostics A/S, Herlev, Denmark). Serum levels of osteocalcin, a marker of bone formation, were determined with a Mouse Osteocalcin IRMA kit (Immutopics, Inc., San Clemente, CA).
As a marker of cartilage destruction, serum levels of COMP (cartilage oligomeric matrix protein) were determined with an Animal COMP® ELISA kit (provided by AnaMar Medical AB, Uppsala, Sweden).
Quantification of serum IgG and CII-specific antibodies
Serum levels of IgG were measured by single radial immunodiffusion as described previously [16]. By use of a previously described ELISA, serum levels of anti-CII antibodies were determined [17].
Interleukin-6 bioassay
A bioassay [18] with cell line B13.29, subclone B9 (which is dependent on interleukin (IL)-6 for growth), was used to measure levels of IL-6 in serum. B9 cells were seeded with 5,000 cells per well into flat-bottomed 96-well plates (Nunc, Roskilde, Denmark) and cultured in Iscove's medium (Sigma) enriched with 50 μg/ml gentamicin (Sigma), 4 mM L-glutamine (Sigma), 50 μM mercaptoethanol (Sigma) and 10% fetal calf serum (Biological Ind., Beit Haemek, Israel). Sera were diluted 1:50 and added in triplicates. After 68 hours of culture, 1 μCi of 3H-thymidine (Amersham Pharmacia Biotech, Uppsala, Sweden) was added; the cells were harvested 4 hours later. Recombinant mouse IL-6 (National Institute for Biological Standards and Control, Potters Bar, Hertfordshire, UK) was used as a standard.
Analysis of bone marrow cells
One femur was flushed with 2 ml of phosphate-buffered saline through the bone cavity to harvest bone marrow cells. After centrifugation at 515 g for 5 min, the pellet was resuspended in Tris-buffered 0.83% NH4Cl solution, pH 7.29, for 5 min to lyse erythrocytes, and then washed in phosphate-buffered saline. The cells were kept in complete Iscove's medium (described above) until use. Leukocytes were counted with an automated cell counter (Sysmex, Kobe, Japan).
The cells were stained with anti-CD45R/B220 conjugated with fluorescein isothiocyanate (clone RA3-6B2; BD) for B-lymphocytes and anti-CD3-conjugated with phycoerythrin (PE) (clone 145-2C11; BD), anti-CD4-biotin (clone RM4-5; BD), anti-CD8-biotin (clone 53-6.7; BD), anti-CD69-PE (clone H1.2F3; BD) and anti-CD25-PE (clone 7D4; BD) for T-lymphocytes. Cells were then subjected to fluorescence-activated cell sorting (FACS) analysis with FACSCalibur (BD Pharmingen, Franklin Lakes, NJ) and analyzed with Paint-A-Gate software (BD). Results are expressed as the numbers of positively stained cells per femur.
Statistical analysis
For statistical evaluation the non-parametric Kruskal–Wallis test followed by a post hoc test was used between all four groups. A Mann–Whitney test was used when two groups were compared. P < 0.05 was considered statistically significant.
Results
OVX results in more severe arthritis
Nine days after OVX/sham operation, mice were immunized (day 0) with chicken CII, and 3 weeks later (day 21) they received a booster injection. Arthritis developed from day 24, and arthritic score was evaluated twice a week. Ovariectomized mice displayed a more severe disease (Fig. 1) than sham-operated mice.
Arthritis and loss of endogenous estrogen lead to an additive and similar degree of bone loss
After termination of the experiment (day 45), BMD of the right femur was measured by pQCT. Mice subjected to OVX displayed a trabecular bone loss of 22% compared with sham-operated non-arthritic controls. Arthritic sham-operated mice displayed a bone loss of 26% and, finally, ovariectomized mice with arthritis had a 58% decrease in trabecular BMD (Figs 2a and 3). (These values were obtained by dividing the difference between the medians of each group and the sham-operated control group by the median of the sham-operated control group.) The cortical BMD was decreased by arthritis but was unaffected by hormonal status (Fig. 2b).
Arthritis is associated with increased bone resorption, and OVX with increased bone formation
At day 45, serum levels of osteocalcin were increased in ovariectomized mice compared with sham-operated mice (Fig. 4a). Immunization with CII did not affect the levels of osteocalcin. Serum levels of RatLaps (type I collagen cross-links) were greatly enhanced in the CII-immunized mice, in comparison with controls (Fig. 4b). In contrast, OVX did not increase the levels of RatLaps.
Arthritis, but not estrogen deficiency, increases cartilage destruction
Serum levels of COMP were increased in arthritic mice but were not affected by hormonal status (Fig. 4c).
Hormonal status does not affect arthritis-induced increased levels of pro-inflammatory cytokines, IgG and CII antibodies
As shown in Table 1, serum levels of the pro-inflammatory cytokine IL-6 were low in non-arthritic mice in comparison with the higher levels found in arthritic mice. All arthritic mice displayed high serum levels of IgG and anti-CII antibodies, but no significant differences between the ovariectomized and sham-operated mice were demonstrated.
Phenotypes of bone marrow lymphocytes are influenced both by OVX and by arthritis
Flow cytometry analysis was performed to evaluate the effects of OVX and arthritis on phenotypes of bone marrow lymphocytes (Table 2). OVX was associated with an increased number of B lymphocytes per femur, whereas CII immunization led to a decreased number of B cells. The total numbers of T lymphocytes (CD3+) and CD4+ cells per femur were not affected by either OVX or CII immunization. In contrast, the number of CD8+ cells was significantly decreased in both sham-operated and ovariectomized arthritic mice compared with controls. The CD69 expression, a marker of early activation, was increased on CD4+ and CD8+ cells in arthritic mice. In contrast, T cell CD25 expression remained unchanged in all groups (data not shown).
Histological findings
There was no significant difference in the degree of histological destruction score between ovariectomized and sham-operated arthritic mice (Table 1).
Discussion
Osteoporosis is one of the major problems in postmenopausal RA [7,19] and is a factor contributing to increased risk for fractures [20]. The mechanisms and relative importance of estrogen deficiency versus inflammation for the bone loss in postmenopausal RA are not fully understood. Our study is the first to demonstrate equal contributions of estrogen deficiency and polyarthritis to bone loss in a model of human postmenopausal RA. In addition, serum markers of bone and cartilage turnover and FACS analysis of bone marrow leukocyte phenotypes indicate different mechanisms for the development of osteoporosis.
OVX of the DBA/1 mice several weeks before the development of arthritis enabled separate and concurrent analyses of the effects of estrogen deficiency and the inflammatory disease on bone loss. Our results show that the loss of endogenous estrogen and the ongoing arthritic disease cause a similar degree of trabecular bone loss (22% and 26%, respectively) and clearly have an additive effect, because ovariectomized mice with arthritis lost 58% of trabecular BMD. Interestingly, arthritis also induced a significant decrease in cortical BMD, whereas OVX, irrespective of inflammatory status, did not affect this parameter.
It has previously been demonstrated in CIA in rats that OVX enhances the severity of arthritis and bone loss, whereas exposure to estrogen suppresses it [13]. A more detailed comparison between the previous study and ours is not possible because we ovariectomized the mice 2 weeks before initial immunization (that is, 5 weeks before the development of arthritis) to achieve an established postmenopausal state, whereas Yamasaki and colleagues ovariectomized the rats 1 week after sensitization.
Systemic inflammation, impaired physical activity, low body mass and treatment with corticosteroids are some important factors associated with the development of osteoporosis in RA. The pathophysiological mechanisms of bone loss in arthritis have been shown to be mediated through the activation of osteoclasts by the macrophage-derived proinflammatory cytokines tumor necrosis factor-α and IL-1, and by the production of RANKL by activated T-lymphocytes and fibroblasts. Garnero and colleagues [21] found increased serum levels of markers of bone resorption in patients with erosive RA, and decreased markers of bone formation. The discrepancy between bone formation and bone resorption results in the enhanced bone loss in arthritis.
We showed that there was strongly increased bone resorption measured by RatLaps in the arthritic mice but not in ovariectomized mice. This was expected, as we sought to study changes in established menopause, and not the rapid phase of bone loss that follows OVX. In contrast to what Garnero and colleagues found in RA patients, we failed to demonstrate decreased serum levels of osteocalcin associated with arthritis. In accord with our results, Nishida and colleagues [22] have previously suggested that reduced bone formation might not be a substantial contributor to bone loss in DBA/1 mice, so this difference might be species dependent.
The exact mechanism whereby OVX induces bone loss in mice is not yet known. Several mechanisms are involved, and recent studies have shown that OVX of mice was associated with an increase in the number of activated, tumor necrosis factor-producing, bone marrow T lymphocytes stimulating monocytes to differentiate into osteoclasts [12,23,24]. We did not show an increase in bone marrow T lymphocytes. The explanation for this discrepancy could be either that we used CD3 as a marker for T cells, whereas others have used anti-CD90 (which is also expressed on natural killer cells, monocytes and dendritic cells), or the very late time point (8 weeks after OVX) that we used for analysis of the bone marrow. Indeed, the finding that RatLaps, a serum marker of bone resorption, was unaltered whereas osteocalcin, a serum marker of bone formation, was increased in ovariectomized mice indicates that the period of OVX-induced increased activation of osteoclasts had already ended at this late time point. As has been shown previously, the number of B lymphocytes in bone marrow was increased after OVX [25] and decreased in the arthritic mice [26]. As increased B lymphopoiesis has been shown to be associated with bone loss [27], our data suggest separate mechanisms for the bone loss found in estrogen deficiency and in arthritis.
COMP is an extracellular matrix protein initially found in cartilage but recently also shown to be secreted by synovial fibroblasts. Serum levels of COMP are used as a marker of cartilage destruction and have previously been evaluated in CIA in rats [28,29]. We found increased serum COMP levels in all arthritic mice, irrespective of the estrogen level, indicating a lack of cartilage protection by endogenous ovarian hormones.
Taken together, although the analyses in this study were all performed on day 45, the differences in serum levels of RatLaps, osteocalcin, COMP and frequencies and phenotypes of bone marrow lymphocytes between mice subjected to OVX and CIA suggest the possibility of different mechanisms for the development of osteoporosis in estrogen deficiency and arthritic disease.
The female sex hormone estradiol not only preserves bone but also has a clear anti-arthritic effect both in human RA [4,5] as well as in rat [13] and murine [11,30] CIA. Clinically, the arthritic ovariectomized mice developed a more severe disease than the sham-operated mice. However, at termination of the experiment all mice, irrespective of hormonal status, had developed severe arthritic disease, with histological destruction score, pro-inflammatory cytokines and CII antibodies at similar levels.
Conclusion
We demonstrate that CIA in ovariectomized DBA/1 mice is a relevant model for studies of osteoporosis in postmenopausal RA. Furthermore, the loss of endogenous estrogen and the inflammation contribute equally to bone loss in this model. Markers of bone and cartilage turnover, as well as bone marrow lymphocyte phenotypes, indicate different mechanisms for bone loss induced by estrogen deficiency and inflammation, respectively. We suggest that this model is well suited for future studies, both on anti-arthritic and anti-osteoporotic properties of new medications and on mechanisms for bone loss in postmenopausal polyarthritis.
Abbreviations
BMD = bone mineral density; CIA = collagen-induced arthritis; CII = type II collagen; COMP = cartilage oligomeric matrix protein; ELISA = enzyme-linked immunosorbent assay; FACS = fluorescence-activated cell sorting; IL = interleukin; OVX = ovariectomy; pQCT = peripheral quantitative computed tomography; RA = rheumatoid arthritis.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
HC and CO participated in study design, interpretation of data and manuscript preparation. UI aided with analysis of data and statistical analysis. ME and MV aided with acquisition of data. The study was performed mainly by CJ. All authors read and approved the final manuscript.
Acknowledgements
We thank Berit Eriksson, Anette Hansevi and Maud Petersson for excellent technical assistance. This study was supported by grants from the Göteborg Medical Society, King Gustav V's 80 years' foundation, the Sahlgrenska Foundation, the Novo Nordic Foundation, the Börje Dahlin foundation, the Association against Rheumatism, Reumaforskningsfond Margareta, the Medical Faculty of Göteborg University (ALF) and the Swedish Research Council.
Figures and Tables
Figure 1 Mice after ovariectomy (OVX) displayed a significantly more severe disease than sham-operated mice. (a) The mice were observed twice weekly for frequency of arthritis. They were considered arthritic when they displayed signs of arthritis in one joint for two consecutive assessments, or arthritis in more than one joint. (b) Severity of arthritis was evaluated twice weekly. Severity was graded 1 to 3 in each paw (maximum 12 points per mouse). Open circles, sham (n = 18); filled circles, ovariectomy (n = 15). *P < 0.05; **P < 0.01; ***P < 0.001. CII, type II collagen.
Figure 2 Ovariectomy decreased trabecular BMD whereas arthritis decreased both trabecular and cortical BMD. Peripheral quantitative computer tomography (pQCT) was performed to measure trabecular and cortical bone mineral density (BMD). (a) Trabecular bone mineral density (BMD) was determined with a metaphyseal scan at a point 3% of the length of the femur from the growth plate and the inner 45% of the area was defined as the trabecular bone compartment. (b) Cortical BMD of the femur was determined with a mid-diaphyseal scan. Results are shown as box plots (values are given as medians (horizontal lines), interquartile ranges (box) and ranges (whiskers); circles represent outliers). For controls, n = 10 for sham (open boxes) and ovariectomy (filled boxes); for immunized mice, n = 18 for sham and n = 14 for ovariectomy. **P < 0.01; ***P < 0.001. CII, type II collagen.
Figure 3 Peripheral quantitative computed tomography (pQCT) scans of one representative mouse in each group. Trabecular bone mineral density (BMD) was determined with a metaphyseal scan at a point 3% of the length of the femur from the growth plate and the inner 45% of the area was defined as the trabecular bone compartment. (a) Sham-operated control; (b) ovariectomy control; (c) sham-operated, arthritic mouse; (d) ovariectomized, arthritic mouse. The bar shows the density of the bone, from 0 (black) to 750 mg/cm3 (white).
Figure 4 Ovariectomy increased bone formation and arthritis increased bone resorption and cartilage destruction. (a) Ovariectomy (OVX) increased bone formation. Serum levels of osteocalcin were analyzed by immunoradiometric assay. (b) Arthritis increased bone resorption. Serum levels of RatLaps were analyzed by ELISA. For controls, n = 10 for sham (open boxes) and ovariectomy (filled boxes); for immunized mice, n = 18 for sham and n = 15 for ovariectomy. (c) Arthritis increased cartilage destruction. Serum levels of cartilage oligomeric matrix protein (COMP) were analyzed by ELISA. For controls, n = 9 for sham (open boxes) and n = 10 for ovariectomy (filled boxes); for immunized mice, n = 17 for sham and n = 14 for ovariectomy. **P < 0.01, ***P < 0.001. Results are shown as box plots (values are given as medians (horizontal lines), interquartile ranges (box) and ranges (whiskers), circles represent outliers).
Table 1 Serological markers of inflammation and histopathological findings were not significantly affected by ovariectomy
Arthritis OVX No. of mice IgG (mg/ml) CII antibody (ng/ml) Interleukin-6 (pg/ml) Frequency of arthritis, day 45 (%) Arthritic score, day 45 Histopathology (score)
- - 10 10 (10–12) n.d 62 (48–67) 0 0 0
+ 10 10 (7–15) n.d 80 (42–111) 0 0 0
+ - 18 18 (15–23) 4.6 (2.6–8.0) 343 (214–755) 100 8 (7–10)*** 9.5 (7.0–11.5)
+ 15 18 (15–18) 3.5 (2.3–4.6) 371 (280–602) 100 11 (10–12) 11.0 (8.9–13.0)
Values are medians and interquartile ranges for each group. The maximum arthritic score was 12 points per mouse. ***P < 0.001 between sham-operated and ovariectomized arthritic mice. CII, type II collagen; n.d., not detectable; OVX, ovariectomy.
Table 2 Characteristics of bone marrow lymphocytes were influenced both by ovariectomy and by arthritis
Arthritis OVX n Bone marrow cellularity (× 106) B cells per femur (× 106) T cells per femur (× 106) CD4+ cells per femur (× 106) CD69+/CD4+ cells (%) CD8+ cells per femur (× 106) CD69+/CD8+ cells (%)
- - 10 5.1 (4.0–6.4) 1.5 (1.0–1.7)*** 0.06 (0.05–0.09) 0.02 (0.02–0.04) 25 (18–30) 0.013 (0.008–0.018) 3 (1–3)
+ 10 6.0 (5.0–8.6) 2.4 (2.0–3.0) 0.05 (0.04–0.07) 0.02 (0.02–0.03) 29 (26–31) 0.008 (0.006–0.012) 4 (2–6)
+ - 18 5.2 (4.7–7.0) 1.0 (0.8–1.3)**† 0.05 (0.04–0.06) 0.02 (0.01–0.02) 50 (43–58)***††† 0.004 (0.002–0.007)††† 9 (4–25)†††
+ 15 6.1 (5.3–8.4) 1.5 (1.0–2.5)† 0.05 (0.03–0.05) 0.02 (0.01–0.02) 35 (29–40) 0.005 (0.002–0.006)††† 10 (5–16)†
Values are medians and interquartile ranges for each group; n is the number of mice. Comparison between sham operation and ovariectomy (OVX): **P < 0.01; ***P < 0.001. Comparison between arthritic mice and their controls: †P < 0.05; †††P < 0.001.
==== Refs
Doran MF Pond GR Crowson CS O'Fallon WM Gabriel SE Trends in incidence and mortality in rheumatoid arthritis in Rochester, Minnesota, over a forty-year period Arthritis Rheum 2002 46 625 631 11920397 10.1002/art.509
Goemaere S Ackerman C Goethals K De Keyser F Van der Straeten C Verbruggen G Mielants H Veys EM Onset of symptoms of rheumatoid arthritis in relation to age, sex and menopausal transition J Rheumatol 1990 17 1620 1622 2084234
Doran MF Crowson CS O'Fallon WM Gabriel SE The effect of oral contraceptives and estrogen replacement therapy on the risk of rheumatoid arthritis: a population based study J Rheumatol 2004 31 207 213 14760786
Ostensen M Aune B Husby G Effect of pregnancy and hormonal changes on the activity of rheumatoid arthritis Scand J Rheumatol 1983 12 69 72 6857175
Forsblad D'Elia H Larsen A Mattsson LA Waltbrand E Kvist G Mellstrom D Saxne T Ohlsson C Nordborg E Carlsten H Influence of hormone replacement therapy on disease progression and bone mineral density in rheumatoid arthritis J Rheumatol 2003 30 1456 1463 12858441
Sinigaglia L Nervetti A Mela Q Bianchi G Del Puente A Di Munno O Frediani B Cantatore F Pellerito R Bartolone S A multicenter cross sectional study on bone mineral density in rheumatoid arthritis. Italian Study Group on Bone Mass in Rheumatoid Arthritis J Rheumatol 2000 27 2582 2589 11093437
Forsblad D'Elia H Larsen A Waltbrand E Kvist G Mellstrom D Saxne T Ohlsson C Nordborg E Carlsten H Radiographic joint destruction in postmenopausal rheumatoid arthritis is strongly associated with generalised osteoporosis Ann Rheum Dis 2003 62 617 623 12810422 10.1136/ard.62.7.617
Riggs BL Khosla S Melton LJ 3rd Sex steroids and the construction and conservation of the adult skeleton Endocr Rev 2002 23 279 302 12050121 10.1210/er.23.3.279
Walsh NC Gravallese EM Bone loss in inflammatory arthritis: mechanisms and treatment strategies Curr Opin Rheumatol 2004 16 419 427 15201606 10.1097/01.bor.0000127824.42507.68
Holmdahl R Bockermann R Backlund J Yamada H The molecular pathogenesis of collagen-induced arthritis in mice – a model for rheumatoid arthritis Ageing Res Rev 2002 1 135 147 12039453 10.1016/S0047-6374(01)00371-2
Holmdahl R Jansson L Andersson M Female sex hormones suppress development of collagen-induced arthritis in mice Arthritis Rheum 1986 29 1501 1509 3801072
Roggia C Gao Y Cenci S Weitzmann MN Toraldo G Isaia G Pacifici R Up-regulation of TNF-producing T cells in the bone marrow: a key mechanism by which estrogen deficiency induces bone loss in vivo Proc Natl Acad Sci U S A 2001 98 13960 13965 11717453 10.1073/pnas.251534698
Yamasaki D Enokida M Okano T Hagino H Teshima R Effects of ovariectomy and estrogen replacement therapy on arthritis and bone mineral density in rats with collagen-induced arthritis Bone 2001 28 634 640 11425652 10.1016/S8756-3282(01)00426-4
Holmdahl R Jansson L Larsson E Rubin K Klareskog L Homologous type II collagen induces chronic and progressive arthritis in mice Arthritis Rheum 1986 29 106 113 3947407
Windahl SH Vidal O Andersson G Gustafsson JA Ohlsson C Increased cortical bone mineral content but unchanged trabecular bone mineral density in female ERβ-/- mice J Clin Invest 1999 104 895 901 10510330
Mancini G Carbonara AO Heremans JF Immunochemical quantitation of antigens by single radial immunodiffusion Immunochemistry 1965 2 235 254 4956917 10.1016/0019-2791(65)90004-2
Verdrengh M Jonsson IM Holmdahl R Tarkowski A Genistein as an anti-inflammatory agent Inflamm Res 2003 52 341 346 14504672 10.1007/s00011-003-1182-8
Helle M Boeije L Aarden LA Functional discrimination between interleukin 6 and interleukin 1 Eur J Immunol 1988 18 1535 1540 3263920
Haugeberg G Uhlig T Falch JA Halse JI Kvien TK Bone mineral density and frequency of osteoporosis in female patients with rheumatoid arthritis: results from 394 patients in the Oslo County Rheumatoid Arthritis register Arthritis Rheum 2000 43 522 530 10728744 10.1002/1529-0131(200003)43:3<522::AID-ANR7>3.0.CO;2-Y
Huusko TM Korpela M Karppi P Avikainen V Kautiainen H Sulkava R Threefold increased risk of hip fractures with rheumatoid arthritis in Central Finland Ann Rheum Dis 2001 60 521 522 11302878 10.1136/ard.60.5.521
Garnero P Jouvenne P Buchs N Delmas PD Miossec P Uncoupling of bone metabolism in rheumatoid arthritis patients with or without joint destruction: assessment with serum type I collagen breakdown products Bone 1999 24 381 385 10221550 10.1016/S8756-3282(98)00193-8
Nishida S Tsurukami H Sakai A Sakata T Ikeda S Tanaka M Ito M Nakamura T Stage-dependent changes in trabecular bone turnover and osteogenic capacity of marrow cells during development of type II collagen-induced arthritis in mice Bone 2002 30 872 879 12052456 10.1016/S8756-3282(02)00725-1
Cenci S Toraldo G Weitzmann MN Roggia C Gao Y Qian WP Sierra O Pacifici R Estrogen deficiency induces bone loss by increasing T cell proliferation and lifespan through IFN-gamma-induced class II transactivator Proc Natl Acad Sci U S A 2003 100 10405 10410 12923292 10.1073/pnas.1533207100
Cenci S Weitzmann MN Roggia C Namba N Novack D Woodring J Pacifici R Estrogen deficiency induces bone loss by enhancing T-cell production of TNF-alpha J Clin Invest 2000 106 1229 1237 11086024
Erlandsson MC Jonsson CA Lindberg MK Ohlsson C Carlsten H Raloxifene- and estradiol-mediated effects on uterus, bone and B lymphocytes in mice J Endocrinol 2002 175 319 327 12429030 10.1677/joe.0.1750319
Inoue M Wakabayashi K Ogihara Y Variation of lymphocytes in peripheral blood and bone marrow in collagen-induced arthritis Chem Pharm Bull (Tokyo) 1994 42 733 735 8004725
Miyaura C Onoe Y Inada M Maki K Ikuta K Ito M Suda T Increased B-lymphopoiesis by interleukin 7 induces bone loss in mice with intact ovarian function: similarity to estrogen deficiency Proc Natl Acad Sci U S A 1997 94 9360 9365 9256487 10.1073/pnas.94.17.9360
Larsson E Erlandsson Harris H Larsson A Mansson B Saxne T Klareskog L Corticosteroid treatment of experimental arthritis retards cartilage destruction as determined by histology and serum COMP Rheumatology (Oxford) 2004 43 428 4234 15024132
Larsson E Erlandsson Harris H Lorentzen JC Larsson A Mansson B Klareskog L Saxne T Serum concentrations of cartilage oligomeric matrix protein, fibrinogen and hyaluronan distinguish inflammation and cartilage destruction in experimental arthritis in rats Rheumatology (Oxford) 2002 41 996 1000 12209032
Jansson L Holmdahl R Oestrogen induced suppression of collagen arthritis. IV: Progesterone alone does not affect the course of arthritis but enhances the oestrogen-mediated therapeutic effect J Reprod Immunol 1989 15 141 150 2769647 10.1016/0165-0378(89)90033-8
| 15987485 | PMC1175035 | CC BY | 2021-01-04 16:02:37 | no | Arthritis Res Ther. 2005 Apr 27; 7(4):R837-R843 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1753 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17541598748910.1186/ar1754Research ArticleGene expression profile and synovial microcirculation at early stages of collagen-induced arthritis Gierer Philip 12Ibrahim Saleh 3Mittlmeier Thomas 1Koczan Dirk 3Moeller Steffen 3Landes Jürgen 4Gradl Georg 12Vollmar Brigitte [email protected] Department of Experimental Surgery, University of Rostock, Rostock, Germany2 Department of Trauma & Reconstructive Surgery, University of Rostock, Rostock, Germany3 Institute of Immunology, University of Rostock, Rostock, Germany4 Department of Surgery, Klinikum Innenstadt, Ludwig Maximilians University, Munich, Germany2005 17 5 2005 7 4 R868 R876 10 3 2005 13 4 2005 Copyright © 2005 Gierer et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
A better understanding of the initial mechanisms that lead to arthritic disease could facilitate development of improved therapeutic strategies. We characterized the synovial microcirculation of knee joints in susceptible mouse strains undergoing intradermal immunization with bovine collagen II in complete Freund's adjuvant to induce arthritis (i.e. collagen-induced arthritis [CIA]). Susceptible DBA1/J and collagen II T-cell receptor transgenic mice were compared with CIA-resistant FVB/NJ mice. Before onset of clinical symptoms of arthritis, in vivo fluorescence microscopy of knee joints revealed marked leucocyte activation and interaction with the endothelial lining of synovial microvessels. This initial inflammatory cell response correlated with the gene expression profile at this disease stage. The majority of the 655 differentially expressed genes belonged to classes of genes that are involved in cell movement and structure, cell cycle and signal transduction, as well as transcription, protein synthesis and metabolism. However, 24 adhesion molecules and chemokine/cytokine genes were identified, some of which are known to contribute to arthritis (e.g. CD44 and neutrophil cytosolic factor 1) and some of which are novel in this respect (e.g. CC chemokine ligand-27 and IL-13 receptor α1). Online in vivo data on synovial tissue microcirculation, together with gene expression profiling, emphasize the potential role played by early inflammatory events in the development of arthritis.
==== Body
Introduction
Murine collagen-induced arthritis (CIA) is a chronic inflammatory disease that bears all the hallmarks of rheumatoid arthritis (RA), namely polyarthritis and synovitis with subsequent cartilage and bone erosions [1]. CIA is induced in susceptible strains of mice (e.g. DBA1/J) by immunization with bovine collagen type II in complete Freund's adjuvant (CFA). The development of CIA is thought to depend on T cells, and disease susceptibility is linked to the major histocompatibility region [2]. Activated lymphocytes migrate to the joint, where an inflammatory cascade involving T cells, macrophages, monocytes, B cells and activated synoviocytes is triggered. This cellular infiltration, together with production of a complex array of cytokines and other soluble mediators, contributes to synovial proliferation, pannus formation, cartilage destruction and subchondral bone erosion [3].
Because the inflammatory process within joint tissues represents a key feature of RA, an understanding of the mechanisms that induce and sustain this aspect of RA pathology would permit development of new and powerful therapeutic strategies. With direct online visualization, the technique of intravital fluorescence microscopy permits dissection of the complex cell inflammatory response, with differentiation between cellular subtypes and their distinct adhesion molecule dependent interactions within the microcirculation.
The approach of in vivo microscopy has successfully been applied in joints of mice with antigen-induced arthritis (AIA) [4]. AIA is established in mice by immunizing them with methylated bovine serum albumin (mBSA) in CFA with or without an arthritogenic infectious agent at days 0 and 7, followed by intra-articular injection of mBSA at day 21 [5]. Although AIA is an established animal model for the study of human RA [6], arthritis is more commonly induced using collagen, and this represents the primary animal model for RA in humans [7-9]. Therefore, we employed in vivo microcirculatory analysis of knee joints in mice with CIA, using different strains that are known to acquire CIA, such as DBA1/J mice and T-cell receptor (TCR) transgenic mice that carry the rearranged Vα11.1 and Vβ8.2 chain genes isolated from a type II collagen-specific T-cell hybridoma (DBA-CII-TCR Tg) [10]. FVB/NJ mice were used as controls, because these mice have been reported to be resistant to arthritis induction, probably because of a genomic deletion of TCR Vβ gene segments [11].
Because we were particularly interested in the disease initiation stage, the synovial microcirculation was assessed before the onset of clinical symptoms of arthritis. We further characterized the global gene expression profile at this early stage in the disease in order to define the initial molecular mechanisms and to determine the onset of joint inflammation.
Materials and methods
Animal model
The experimental protocol was approved by the local animal rights protection authorities (LVL M-V/TSD/7221.3-1.1-037/04) and followed the National Institutes of Health guidelines for the care and use of laboratory animals. DBA1/J and FVB/NJ mice were obtained from the Jackson Laboratory (Bar Harbor, ME, USA). Collagen II specific TCR transgenic mice were a kind gift from Professor Ladiges, University of Washington, USA [10]. All mice were kept under standard conditions at the animal care facility of the University of Rostock.
Mice aged 8 weeks (n = 5–10 per strain and group) were immunized intradermally at the base of the tail with 125 μg bovine collagen II (Chondrex, Redmond, WA, USA) emulsified in CFA (DIFCO, Detroit, MI, USA) or with equivalent volumes of CFA only. Six weeks after immunization and before clinical signs of arthritis manifested, animals were anaesthetized with ketamine (90 mg/kg body weight) and xylacin (6 mg/kg) and placed on a heating pad to maintain their body temperature at 37°C. A catheter was placed in the left jugular vein for application of fluorescent dyes.
For in vivo multifluorescence microscopy of synovial microcirculation, we used the knee joint model initially described by Veihelmann and coworkers [4]. Briefly, the skin was incised distal to the patella tendon. After removal of the overlying soft tissues, the patella tendon was transversally cut and the proximal and distal part carefully mobilized. After exposure, the 'Hoffa's fatty body' was superfused with 37°C warm physiological saline solution to prevent the tissues from drying and finally covered with a glass slide. Following a 15-min stabilization period after surgical preparation, in vivo microscopy of the synovial tissue was performed. At the end of the experiments, animals were killed by exsanguination. The complete knee joint was excised and harvested for subsequent histology. Paws were used for gene expression profile analysis.
Clinical evaluation of arthritis
As described by Nanakumar and coworkers [12], scoring of animals was done blindly using a scoring system based on the number of inflamed joints in each paw, inflammation being defined by swelling and redness. In this scoring system each inflamed toe or knuckle is attributed 1 point, whereas an inflamed wrist or ankle is attributed 5 points, resulting in a score of 0 to 15 (five toes + five knuckles + one wrist/ankle) for each paw and 0–60 points for each mouse [12].
In vivo fluorescence microscopy
After intravenous injection of FITC-labelled dextran (15 mg/kg body weight; Sigma, Deisenhofen, Germany) and rhodamine 6G (0.15 mg/kg body weight; Sigma), in vivo microscopy was performed using a Zeiss microscope (Axiotech vario 100HD; Carl Zeiss, Oberkochen, Germany) equipped with a 100 W mercury lamp and filter sets for blue (excitation 465–495 nm, emission >505 nm) and green (excitation 510–560 nm, emission >575 nm) epi-illumination. Using water-immersion objectives (×20 W/numerical aperture 0.5 and ×40 W/numerical aperture 0.8; Carl Zeiss), final magnifications of 306× and 630× were achieved. Images were recorded by means of a charge-coupled device video camera (FK 6990-IQ-S; Pieper, Schwerte, Germany) and transferred to a S-VHS video system for subsequent offline analysis.
Microcirculatory analysis
For quantitative offline analysis a computer-assisted microcirculation image analysis system was used (CapImage v7.4; Zeintl, Heidelberg, Germany). Functional capillary density was defined as the total length of red blood cell perfused capillaries per observation area, and is given in cm/cm2. For assessment of leucocyte–endothelial cell interaction in postcapillary venules, flow behaviour of leucocytes was analyzed with respect to free floating, rolling and adherent leucocytes. Rolling leucocytes were defined as those cells moving along the vessel wall at a velocity less than 40% that of leucocytes at the centre line, and are expressed as a percentage of the total leucocyte flux. Venular leucocyte adherence was defined as the number of leucocytes not moving or detaching from the endothelial lining of the venular vessel wall during an observation period of 20 s. Assuming cylindrical microvessel geometry, leucocyte adherence was expressed as nonmoving cells per endothelial surface (n/mm2), calculated from the diameter and length of the vessel segment analyzed. In postcapillary venules, centre line red blood cell velocity (VRBC) was determined using the line shift method (CapImage; Zeintl, Heidelberg, Germany). The wall shear rate was calculated based on the Newtonian definition: y = 8 × Vmean/D, where Vmean is the mean velocity (VRBC/1.6) and D is diameter of the individual microvessel.
Laboratory analysis
Arterial blood samples were drawn for analysis of blood cell counts using a Coulter Counter (AcTdiff; Coulter, Hamburg, Germany).
Sample preparation for high-density oligonucleotide microarray hybridization
Paws of DBA/1J mice immunized with CFA or CFA/collagen II and unimmunized mice were dissected and snap frozen in liquid nitrogen, and total RNA was extracted using a commercially available system (Qiagen, Hilden, Germany). RNA probes were labelled in accordance with the manufacturer's instructions (Affymetrix, Santa Clara, CA, USA). Analysis of gene expression was conducted using the U430A array (Affymetrix), which has a capacity of about 20,000 genes. Samples from individual mice were hybridized onto individual arrays. Hybridization and washing of gene chips were done in accordance with the manufacturer's instructions and were as described previously [13]. Microarrays were analyzed by laser scanning and the expression levels were calculated using commercially available software provided by Affymetrix [13]. The files were then analyzed using the affylmGUI package of the Bioconductor software suite (Affymetrix) [14,15]. The expression was determined using the robust multichip average method [16]. A linear model of the expression data for Limma was created within affylmGUI, for which the six arrays of mice immunized with CFA/collagen II, the three arrays for mice administered CFA only, and the two arrays for control mice were separated into three groups. Contrasts were calculated for each group against the other two.
The expression data are available in Additional files 1, 2, 3. Genes considered differentially expressed were selected on the basis of P value (<0.001) and a 1.5-fold change in intensity (abs [M value] = log2 [1.5]). These genes are presented in Additional file 4. Gene Ontology (GO) terms were assigned to the selected genes (Fig. 1 and Additional file 4) via the Bioconductor GO package 1.6.8 and the chip annotation package MOE430a of the same version [17]. The Bioconductor GO package provides lists of reachable subterms for each GO term. We used this function to filter genes associated with adhesion, specifically those assigned to 'adhesion offspring' for any term in the following list: GO:0005125 (cytokine activity), GO:0006955 (immune response), GO:0050776 (regulation of immune response), GO:0004895 (cell adhesion receptor activity), GO:0007155 (cell adhesion), GO:0016337 (cell–cell adhesion), GO:0030155 (regulation of cell adhesion), GO:0050839 (cell adhesion molecule binding), GO:0030155 (regulation of cell adhesion), GO:0019955 (cytokine binding), GO:0005912 (adherens junction), GO:0005925 (focal adhesion), GO:0050900 (immune cell migration), GO:0030595 (immune cell chemotaxis), GO:0006954 (inflammatory response) and GO:0006935 (chemotaxis).
Histology
At the end of each experiment, knee joints were fixed in 4% phosphate-buffered formalin for 2–3 days, decalcified in EDTA for 6 weeks, and then embedded in paraffin. From the paraffin-embedded tissue blocks, 4 μm sections were cut and stained with haematoxylin–eosin for histological analysis. For semiquantitative analysis, the score described by Brackertz and coworkers [6] was used (0 = normal knee joint; 1 = occasional mononuclear cells in normal synovium; 2 = perivascular leucocyte infiltration, two or more synovial cell layers; 3 = dense infiltration of leucocytes, synovial hyperplasia; 4 = synovitis, pannus formation and cartilage erosions). The analysis was done by a blinded and independent observer.
Statistical analysis of microcirculatory data
Results are presented as mean ± standard error of the mean. After proving the assumption of normality, comparisons between the experimental groups were performed by one-way analysis of variance, followed by the appropriate post hoc multiple comparison procedure, including Bonferroni correction (SigmaStat; Jandel, San Rafael, CA, USA). P < 0.05 was considered statistically significant.
Results
Gene expression profile in joints at onset of arthritis
To define the gene expression profile at early stages of CIA, we used the murine Affymetrix oligonucleotide microarray MOE430a, with more than 20,000 gene specificities, to compare the three groups of mice (i.e. CFA/collagen II immunized, CFA immunized and unimmunized). As shown in Additional files 1, 2, 3, 4 and Fig. 1, 655 genes were differentially expressed between groups and taking a P < 0.001 as the threshold for significance.
Interestingly, CFA alone induced the greatest number of differentially expressed genes (i.e. 498). A total of 375 genes overlapped between different groups, and 280 were unique to a certain group of mice (Additional file 4). When grouped according to their probable function (i.e. GO terms), the majority of the differentially expressed genes fell into classes of genes involved in cell movement and structure, cell cycle and signal transduction, as well as transcription, protein synthesis and metabolism. One prominent group was that of adhesion molecules and chemokine/cytokine-related genes. Twenty-four genes belonging to this group were identified and are summarized in Table 1 (Additional file 4). Some genes are well known for their contribution to cell activation, cell–cell communication and chemotaxis, such as CD44, CD36, IL-1 receptor antagonist and neutrophil cytosolic factor (NCF)-1, as well as CC and CXC chemokines (Table 1, Additional file 4).
Systemic parameters
The animals from the CFA/collagen II immunized and the CFA immunized groups did not differ with respect to haemoglobin and haematocrit (Table 2). Moreover, there were no differences in blood cell counts between the two groups (Table 2).
Microvascular perfusion in synovial tissue
Functional capillary density did not differ significantly between the experimental groups (Fig. 2), although there was a tendency toward lower values in the DBA mice, and in particular the TCR transgenic mice, after CFA/collagen II exposure (Fig. 2). Capillary diameters increased and VRBC decreased in all animals exposed to CFA/collagen II exposure compared with those subjected to CFA control treatment (Table 3). Wall shear rates were found to be reduced in the CFA/collagen II treated mice in comparison with those treated with CFA (Table 3).
Inflammatory cell response in synovial tissue
Although the animals exhibited no clinical symptoms of arthritic disease, synovial tissue was characterized by an inflammatory cell response with significant (P < 0.05) increases in leucocytes, both rolling along and firmly attaching to the venular endothelium, in DBA and TCR transgenic mice (Figs 3 and 4). In contrast, CIA-resistant FVB animals did not respond with enhanced leucocyte–endothelial cell interaction on CFA/collagen II exposure, with findings equivalent to those in CFA treated control animals (Figs 3 and 4). Concomitant with the lack of clinical signs, the score in haematoxylin and eosin stained knee joints was found to be less than 1 in all animals, irrespective of genotype and treatment (data not shown).
Discussion
In the present study we found that susceptible mice that were exposed to CFA/collagen II for induction of arthritis exhibited marked signs of inflammation within the microcirculation of the knee joint, although animals were still free from clinical symptoms. Collagen II treated TCR transgenic and DBA/1J mice did not differ in terms of the extent of inflammation, which exceeded that in resistant FVB animals markedly. The inflammatory cell response, as indicated by the enhanced activation and interaction of leucocytes with the microvascular endothelium, was mirrored by the expression of genes that contribute to cell activation, cell–cell communication and chemotaxis.
Despite the considerable work done to elucidate disease pathways, several aspects of RA remain poorly defined. A rigorous understanding of the initial mechanisms involved in the pathogenesis of RA would permit the development of strategies to impede the manifestation of the disease. In numerous organ pathologies, the activation of circulating leucocytes, and their interaction with the endothelial lining followed by subsequent transendothelial migration and infiltration into tissue represent the first and determining step in a complex sequence of processes that mediate tissue injury [18-20]. In contrast to our previous study addressing the expression profile of joints in CIA mice at the peak of the disease [13], we intentionally focused on an early stage in the disease, in which the animals were free of clinical symptoms. Although arthritic disease with establishment of pannus tissue is dominated by genes that are involved in cell division and proliferation, rather than immunologically relevant genes [13], early disease appears to be characterized by distinct upregulation of a group of chemotactic and adhesion molecules, such as CD44 and IL-13 receptor α1, as well as CC chemokine ligand (CCL)-24 and CCL-27, which presumably are responsible for cell attraction within the joint microcirculation. Many of those molecules were induced by both CFA and CFA/collagen II treatment, which is unsurprising because CFA is essential for induction of CIA.
Interestingly, the only upregulated adhesion molecule in the comparison between mice treated with CFA/collagen II and those treated with CFA alone was CD44, supporting a role for CD44 in this early stage of arthritis. Indeed, there is considerable published evidence for CD44 involvement in arthritis, although its exact role remains controversial [21,22].
In accord with the importance of adhesion molecules in development of arthritis, frozen section binding assays in rheumatoid synovitis demonstrated that, apart from E-selectin and counter receptors for β1/β2 integrins, P-selectin is the predominant adhesion molecule, mediating monocyte binding to inflamed synovial venules [23]. Similarly, increased cellular infiltration and increased expression of E-selectin, intercellular adhesion molecule-1, vascular cell adhesion molecule-1, platelet/endothelial cell adhesion molecule-1, very late appearing antigen-4, and Mac-1 were found in immunohistochemistry of synovial tissue from patients with RA [24]. Veihelmann and coworkers [25] demonstrated high numbers of adherent leucocytes upon clinical manifestation of AIA in mice, regardless of phase (acute, intermediate, or chronic) of disease. In accordance with and extending the findings of the latter study, we now show that leucocyte adhesion is apparent even if clinical symptoms are still absent, underscoring leucocyte–endothelial interaction as an integral part not only of the perpetuation and propagation of disease but also of its initiation.
Apart from adhesion molecules, a few chemokines and inflammatory mediators were found among the genes predominantly expressed in CIA mice in the present study. This is in accordance with the common knowledge that the key mechanisms underlying synovitis include inflammatory cell activation and adhesion, as well as production of mediators such as cytokines, chemokines and growth factors [26,27]. In particular, tumour necrosis factor (TNF)-α and IL-1 regulate nuclear factor-κB inducible genes that control – apart from other factors – cell adhesion molecules, proinflammatory mediators and immunomodulatory molecules. These properties established a rationale for anticytokine therapeutics and their evaluation in an extensive series of clinical trials [28]. Anti-TNF-α therapy has been shown to reduce expression of adhesion molecules and to decrease cellularity of rheumatoid synovial tissue [29,30], supporting the hypothesis that the anti-inflammatory effect is due to a downregulation of cytokine-inducible vascular adhesion molecules with a consequent reduction in cell traffic into joints.
A few molecules belonging to the category of chemokines and inflammatory mediators were differentially expressed and deserve further investigation. These are NCF-1, IL-13 receptor α1 and CCL-27. NCF-1 is a member of the NADPH (nicotinamide adenine dinucleotide phosphate, reduced form) oxidase complex, which was recently identified as a susceptibility gene for pristine-induced arthritis. However, its exact role in disease remains unclear [31]. The chemokine ligand CCL-27 was recently shown to bind the P-selectin glycoprotein ligand 1 – a molecule that plays a role in homing of T lymphocytes [32]. The role played by the cytokine receptor IL-13 receptor α1 in arthritis has not been established, but its ligand, IL-13, has been described as a cytokine with anti-inflammatory properties in arthritis and was the target of experimental gene therapy experiments [33].
Of interest, mice from the two strains studied did not differ with respect to functional capillary density, averaging about 320 cm/cm2. Corresponding values were found in Balb/c mice during acute and intermediate phases of AIA [21] but were attributed to inflammation-associated angiogenesis, because control animals had values well below 250 [4,21]. If it were angiogenesis driven, this would not account for the high functional capillary density in CFA treated control animals of the present study. Thus, it is more likely that differences in functional capillary density are simply due to the fact that different strains were used.
Conclusion
Our data suggest that upregulation of proinflammatory mediators and molecules facilitate leucocyte adhesion to the endothelium and migration into tissue, thereby representing an essential and primary step in the development of arthritis. Although studies of early RA are few, because there is an inherent delay before patients receive expert care, it has been recognized that early intervention improves outcome. Thus, the early innate immune response should be an ongoing focus of future research to determine whether leucocyte activation predicts severity of disease and is the earliest change to occur in rheumatoid synovium.
Abbreviations
AIA = antigen-induced arthritis; CCL = CC chemokine ligand; CFA = complete Freund's adjuvant; CIA = collagen-induced arthritis; GO = Gene Ontology; IL = interleukin; mBSA = methylated bovine serum albumin; NCF = neutrophil cytosolic factor; RA = rheumatoid arthritis; TCR = T-cell receptor; TNF = tumour necrosis factor; VRBC = centre line red blood cell velocity.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
PG, JL and GG performed the animal experiments and intravital fluorescence microscopic analysis. SI, DK and SM performed gene expression profiling experiments with bioinformatic analysis. SI, TM and BV conceived the study, and participated in its design and coordination. SI and BV drafted the manuscript. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
Excel file showing the output of the topTable function of the GNU R Limma package, as provided by the R affylmGUI package: complete Freund's adjuvant (CFA)/collagen II versus CFA. The columns first describe the genes using an internal identification number (ID), the Affymetrix probe set ID, the gene's symbol and a small description or full name. The statistics for the genes are summarized in the columns M (logarithmic fold change; the difference in logarithm of expression for each group), A (logarithmic mean expression), t (moderated t statistic), P value (nominal P value) and B (log odds that the gene is differentially expressed). With B positive, the gene is more likely to be differentially expressed than not. At 0 it is uncertain.
Click here for file
Additional File 2
Excel file showing the output of the topTable function of the GNU R limma package as provided by the R affylmGUI package: complete Freund's adjuvant (CFA)/collagen II versus no treatment. The columns first describe the genes using an internal identification number (ID), the Affymetrix probe set ID, the gene's symbol and a small description or full name. The statistics for the genes are summarized in the columns M (logarithmic fold change; the difference in logarithm of expression for each group), A (logarithmic mean expression), t (moderated t statistic), P value (nominal P value) and B (log odds that the gene is differentially expressed). With B positive, the gene is more likely to be differentially expressed than not. At 0 it is uncertain.
Click here for file
Additional File 3
Excel file showing the output of the topTable function of the GNU R limma package as provided by the R affylmGUI package: complete Freund's adjuvant (CFA) vs. no treatment. The columns first describe the genes using an internal identification number (ID), the Affymetrix probe set ID, the gene's symbol and a small description or full name. The statistics for the genes are summarized in the columns M (logarithmic fold change; the difference in logarithm of expression for each group), A (logarithmic mean expression), t (moderated t statistic), P value (nominal P value) and B (log odds that the gene is differentially expressed). With B positive, the gene is more likely to be differentially expressed than not. At 0 it is uncertain.
Click here for file
Additional File 4
Excel file summarizing genes differentially expressed in collagen-induced arthritis (CIA) joints at early stages in the disease. The first column gives the probe set identification number (ID) of the Affymetrix chip Moe430a. Columns 2–4 list whether the gene is significantly upregulated (Up) or downregulated (Dn; minimum 1.5-fold; P < 0.001) in a particular comparison, ordered complete Freund's adjuvant (CFA)/collagen II versus CFA, CFA/collagen II versus no treatment, and CFA versus no treatment. Columns 5 and 6 show the gene symbol and its name. For further details, see Materials and method.
Click here for file
Acknowledgements
The authors thank Hans-Jürgen Thiesen, Institute of Immunology, University of Rostock, for his support in gene expression profiling. This work was supported by the EU FP6 contract MRTN-CT-2004-05693 'EURO-RA'.
Figures and Tables
Figure 1 Summary of differentially expressed genes at the early stage of arthritis. Of the approximately 22,000 genes on the Affymetrix chip MOE430a, 655 genes were significantly differentially expressed (minimum 1.5-fold; P < 0.001) in a particular comparison: complete Freund's adjuvant (CFA)/collagen II versus CFA, CFA/collagen II versus no treatment, or CFA versus no treatment. Genes differentially expressed in more than one comparison are shown.
Figure 2 Functional capillary density. Shown is the functional capillary density of the synovium in complete Freund's adjuvant (CFA)/collagen II exposed FVB/NJ, DBA1/J and T-cell receptor (TCR) transgenic mice (collagen +) in comparison with CFA treated controls (collagen -). Intravital fluorescence microscopy of the knee joints was performed at 6 weeks after collagen exposure for induction of arthritis. Values are expressed as means ± standard error (n = 5–10 animals/group).
Figure 3 Leucocytes rolling along the endothelium of postcapillary synovial venules. Shown are the proportions of leucocytes rolling along the endothelium of postcapillary synovial venules (as % of all passing leucocytes) in complete Freund's adjuvant (CFA)/collagen II exposed FVB/NJ, DBA1/J and T-cell receptor (TCR) transgenic mice (collagen +) in comparison with CFA treated controls (collagen -). Intravital fluorescence microscopy of the knee joints was performed at 6 weeks after collagen exposure for induction of arthritis. Values are expressed as means ± standard error (n = 5–10 animals/group); analysis of variance, unpaired post hoc comparison test: #P < 0.05 versus corresponding CFA-treated control animals (collagen -).
Figure 4 Leucocytes adherent to the endothelium of postcapillary synovial venules. Shown are the numbers of leucocytes adherent to the endothelium of postcapillary synovial venules (cells/mm2 endothelial surface) in complete Freund's adjuvant (CFA)/collagen II exposed FVB/NJ, DBA1/J and T-cell receptor (TCR) transgenic mice (collagen +) in comparison with CFA treated controls (collagen -). Intravital fluorescence microscopy of the knee joints was performed at 6 weeks after collagen exposure for induction of arthritis. Values are expressed as mean ± standard error (n = 5–10 animals/group); analysis of variance, unpaired post hoc comparison test: #P < 0.05 versus corresponding CFA-treated control animals (collagen -).
Table 1 Summary of adhesion molecules and chemokines/cytokines differentially expressed in CIA joints at early stages of disease and their GO terms.
Affymetrix ID Change Gene ID Description Go Term/Function
Solely changed in comparison CFA/collagen II versus CFA
1423760_at Up Cd44 CD44 antigen GO:0007155 cell adhesion
Solely changed in comparison CFA/collagen II versus no treatment
1422103_a_at Dn Stat5b Signal transducer and activator of transcription 5B GO:0030155 cell adhesion
1423166_at - Dn Cd36 CD36 antigen GO:0007155 cell adhesion
1427165_at - Up Il13ra1 IL-13 receptor alpha 1 GO:0004907 IL activity
1434044_at - Dn Repin1 Replication initiator 1 GO:0006954 inflammatory response
1452483_a_at Up Cd44 CD44 antigen GO:0007155 cell adhesion
1452514_a_at Dn Kit Kit oncogene GO:0006935 chemotaxis
Solely changed in comparison CFA versus no treatment
1416156_at Dn Vcl Vinculin GO:0005912 adherens junction
1417705_at Dn Otub1 OTU domain, ubiquitin aldehyde binding 1 GO:0006955 immune response
1422873_at Dn Prg2 Proteoglycan 2, bone marrow GO:0006955 immune response
1430375_a_at Up Ccl27 Chemokine (CC motif) ligand 27 GO:0008009 chemokine activity
1437807_x_at Dn Catna1 Catenin α1 GO:0005912 adherens junction
1452020_a_at Up Siva CD27 binding protein GO:0005175 CD27 receptor binding
1455158_at Dn Itga3 Integrin α3 GO:0007155 cell adhesion
Changed in comparisons CFA/collagen II vs. CFA and CFA vs. no treatment
1450488_at Dn-Up Ccl24 Chemokine (CC motif) ligand 24 GO:0008009 chemokine activity
Changed in comparisons CFA/collagen II vs. no treatment and CFA vs. no treatment
1419329_at Dn Dn Sh3d4 SH3 domain protein 4 GO:0007155 cell adhesion
1420465_s_at Up Up Mup1 Major urinary protein 1 GO:0016068 type I hypersensitivity
1420553_x_at Dn Dn Serpina1a Serine proteinase inhibitor, clade A, member 1a GO:0006953 acute-phase response
1423017_a_at Up Up Il1rn IL-1 receptor antagonist GO:0006955 immune response
1423734_at Up Up Rac1 RAS-related C3 botulinum substrate 1 GO:0007155 cell adhesion
GO:0006954 inflammatory response
1423885_at Dn Dn Lamc1 Laminin γ1 GO:0007155 cell adhesion
1427164_at Up Up Il13ra1 IL-13 receptor α1 GO:0004907 IL activity
1435148_at Dn Dn Atp1b2 ATPase, Na+/K+ transporting, β2 polypeptide GO:0007155 cell adhesion
1448303_at Dn Dn Gpnmb Glycoprotein nmb GO:0007155 cell adhesion
1451767_at Up Up Ncf1 Neutrophil cytosolic factor 1 GO:0006954 inflammatory response
Shown are differentially expressed genes associated with adhesion or inflammation with respect to Gene Ontology (GO) terms. The genes are a subset of those shown in the Venn diagram (Fig. 1). The first column gives the probe set identification number (ID) of the Affymetrix chip Moe430a. The second column lists whether the gene is significantly upregulated (Up) or downregulated (Dn; minimum 1.5-fold) in a particular comparison, ordered complete Freund's adjuvant (CFA)/collagen II versus CFA, CFA/collagen II versus no treatment, and CFA versus no treatment. Columns three to five show the gene symbol, its name and GO terms. For further details, see Materials and method. IL, interleukin.
Table 2 Haemoglobin, haematocrit and blood cell counts following CFA or CFA/collagen II exposure
CFA CFA/collagen II
FVB DBA TCR FVB DBA TCR
Haemoglobin (mmol/l) 6.5 ± 0.4 7.1 ± 0.4 7.1 ± 0.4 7.1 ± 0.2 7.7 ± 0.3 7.9 ± 0.1
Haematocrit (%) 36 ± 2 43 ± 3 41 ± 3 40 ± 1 47 ± 2 46 ± 1
Thrombocytes (109/l) 851 ± 96 447 ± 73 551 ± 124 984 ± 210 729 ± 124 743 ± 111
Leucocytes (109/l) 3.8 ± 0.5 4.4 ± 0.9 2.1 ± 0.8 1.8 ± 0.5 4.9 ± 1.0 3.7 ± 0.7
Lymphocytes (%) 76 ± 5 80 ± 3 70 ± 3 65 ± 2 65 ± 4 65 ± 5
Mixed population (%) 14 ± 1 11 ± 2 26 ± 4 25 ± 3 32 ± 5 30 ± 4
Neutrophils (%) 6 ± 4 4 ± 1 5 ± 1 10 ± 3 3 ± 1 5 ± 2
Values are expressed as mean ± standard error. Blood samples were drawn at 6 weeks after complete Freund's adjuvant (CFA)- and CFA/collagen II exposure. TCR, T-cell receptor.
Table 3 Capillary diameter, venular red blood cell velocity and wall shear rates in synovial tissue following CFA or CFA/collagen II exposure
CFA CFA/collagen II
FVB DBA TCR FVB DBA TCR
Capillary diameter (μm) 4.3 ± 0.1 4.5 ± 0.2 4.4 ± 0.1 4.5 ± 0.1 4.7 ± 0.1 4.6 ± 0.1
VRBC (μm/s) 840 ± 143 1085 ± 350 1760 ± 412 620 ± 278 835 ± 251 487 ± 122*
Wall shear rate (s-1) 0.20 ± 0.03 0.21 ± 0.05 0.40 ± 0.09 0.12 ± 0.03 0.18 ± 0.05 0.10 ± 0.02
Values are expressed as means ± standard error. *P < 0.05 versus corresponding complete Freund's adjuvant (CFA) treated control animals. TCR, T-cell receptor; VRBC, venular red blood cell velocity.
==== Refs
Myers LK Rosloniec EF Cremer MA Kang AH Collagen-induced arthritis, an animal model of autoimmunity Life Sci 1997 61 1861 1878 9364191 10.1016/S0024-3205(97)00480-3
Wooley PH Luthra HS Stuart JM David CS Type II collagen-induced arthritis in mice. I. Major histocompatibility complex (I region) linkage and antibody correlates J Exp Med 1981 154 688 700 6792316 10.1084/jem.154.3.688
Feldmann M Brennan FM Maini RN Role of cytokines in rheumatoid arthritis Annu Rev Immunol 1996 14 397 440 8717520 10.1146/annurev.immunol.14.1.397
Veihelmann A Szczesny G Nolte D Krombach F Refior HJ Messmer K A novel model for the study of synovial microcirculation in the mouse knee joint in vivo Res Exp Med 1998 198 43 54 10.1007/s004330050088
Magilavy DB Animal models of chronic inflammatory arthritis Clin Orthop 1990 259 38 45 2208871
Brackertz D Mitchell GF Mackay IR Antigen-induced arthritis in mice. I. Induction of arthritis in various strains of mice Arthritis Rheum 1977 20 841 850 857805
Trentham DE Townes AS Kang AH Autoimmunity to type II collagen an experimental model of arthritis J Exp Med 1977 146 857 868 894190 10.1084/jem.146.3.857
Courtenay JS Dallman MJ Dayan AD Martin A Mosedale B Immunisation against heterologous type II collagen induces arthritis in mice Nature 1980 283 666 668 6153460 10.1038/283666a0
Wooley PH Immunotherapy in collagen-induced arthritis: past, present, and future Am J Med Sci 2004 327 217 226 15084918 10.1097/00000441-200404000-00008
Osman GE Cheunsuk S Allen SE Chi E Liggitt HD Hood LE Ladiges WC Expression of a type II collagen-specific TCR transgene accelerates the onset of arthritis in mice Int Immunol 1998 10 1613 1622 9846690 10.1093/intimm/10.11.1613
Osman GE Hannibal MC Anderson JP Lasky SR Ladiges WC Hood L FVB/N (H2(q)) mouse is resistant to arthritis induction and exhibits a genomic deletion of T-cell receptor V beta gene segments Immunogenetics 1999 49 851 859 10436178 10.1007/s002510050564
Nandakumar KS Svensson L Holmdahl R Collagen type II-specific monoclonal antibody-induced arthritis in mice: description of the disease and the influence of age, sex, and genes Am J Pathol 2003 163 1827 1837 14578183
Ibrahim SM Koczan D Lorenz P Thiesen HJ Gene-expression profile of collagen-induced arthritis J Autoimmunity 2002 18 159 167 11908948 10.1006/jaut.2001.0580
Wettenhall JM Smyth GK limmaGUI: a graphical user interface for linear modeling of microarray data Bioinformatics 2004 20 3705 3706 15297296
Wu Z Irizarry RA Preprocessing of oligonucleotide array data Nat Biotechnol 2004 22 656 658 15175677 10.1038/nbt0604-656b
Gentleman RC Carey VJ Bates DM Bolstad B Dettling M Dudoit S Ellis B Gautier L Ge Y Gentry J Bioconductor: open software development forcomputational biology and bioinformatics Genome Biol 2004 5 R80 15461798 10.1186/gb-2004-5-10-r80
Harris MA Clark J Ireland A Lomax J Ashburner M Foulger R Eilbeck K Lewis S Marshall B Mungall C The Gene Ontology (GO) database and informatics resource Nucleic Acids Res 2004 32 D258 D261 14681407 10.1093/nar/gkh066
Eipel C Bordel R Nickels RM Menger MD Vollmar B Impact of leukocytes and platelets in mediating hepatocyte apoptosis in a rat model of systemic endotoxemia Am J Physiol Gastrointest Liver Physiol 2004 286 G769 G776 14715524 10.1152/ajpgi.00275.2003
Menger MD Vollmar B Role of microcirculation in transplantation Microcirculation 2000 7 291 306 11079248 10.1038/sj.mn.7300117
Menger MD Richter S Yamauchi J Vollmar B Role of microcirculation in hepatic ischemia/reperfusion injury Hepatogastroenterology 1999 46 1452 1457 10431706
Naor D Nedvetzki S CD44 in rheumatoid arthritis Arthritis Res Ther 2003 5 105 115 12723975 10.1186/ar746
Szanto S Gal I Gonda A Glant TT Mikecz K Expression of L-selectin, but not CD44, is required for early neutrophil extravasation in antigen-induced arthritis J Immunol 2004 172 6723 6734 15153489
Grober JS Bowen BL Ebling H Athey B Thompson CB Fox DA Stoolman LM Monocyte-endothelial adhesion in chronic rheumatoid arthritis. In situ detection of selectin and integrin-dependent interactions J Clin Invest 1993 91 2609 2619 7685772
Tak PP Thurkow EW Daha MR Kluin PM Smeets TJ Meinders AE Breedveld FC Expression of adhesion molecules in early rheumatoid synovial tissue Clin Immunol Immunopathol 1995 77 236 242 7586733 10.1006/clin.1995.1149
Veihelmann A Harris AG Krombach F Schutze E Refior HJ Messmer K In vivo assessment of synovial microcirculation and leukocyte-endothelial cell interaction in mouse antigen-induced arthritis Microcirculation 1999 6 281 290 10654279 10.1038/sj.mn.7300074
Szekanecz Z Koch AE Update on synovitis Curr Rheumatol Rep 2001 3 53 63 11177771
Bingham CO III The pathogenesis of rheumatoid arthritis: pivotal cytokines involved in bone degradation and inflammation J Rheumatol Suppl 2002 65 3 9 12236620
Feldmann M Brennan FM Williams RO Woody JN Maini RN The transfer of a laboratory based hypothesis to a clinically useful therapy: the development of anti-TNF therapy of rheumatoid arthritis Best Pract Res Clin Rheumatol 2004 18 59 80 15123038 10.1016/j.berh.2003.09.010
Tak PP Taylor PC Breedveld FC Smeets TJ Daha MR Kluin PM Meinders AE Maini RN Decrease in cellularity and expression of adhesion molecules by anti-tumor necrosis factor alpha monoclonal antibody treatment in patients with rheumatoid arthritis Arthritis Rheum 1996 39 1077 1081 8670314
Paleolog EM Hunt M Elliott MJ Feldmann M Maini RN Woody JN Deactivation of vascular endothelium by monoclonal anti-tumor necrosis factor alpha antibody in rheumatoid arthritis Arthritis Rheum 1996 39 1082 1091 8670315
Olofsson P Holmberg J Tordsson J Lu S Akerstrom B Holmdahl R Positional identification of Ncf1 as a gene that regulates arthritis severity in rats Nat Genet 2003 33 25 32 12461526 10.1038/ng1058
Hirata T Furukawa Y Yang BG Hieshima K Fukuda M Kannagi R Yoshie O Miyasaka M Human P-selectin glycoprotein ligand-1 (PSGL-1) interacts with the skin-associated chemokine CCL27 via sulfated tyrosines at the PSGL-1 amino terminus J Biol Chem 2004 279 51775 51782 15466853 10.1074/jbc.M409868200
Woods JM Katschke KJ JrTokuhira M Kurata H Arai KI Campbell PL Koch AE Reduction of inflammatory cytokines and prostaglandin E2 by IL-13 gene therapy in rheumatoid arthritis synovium J Immunol 2000 165 2755 2763 10946307
| 15987489 | PMC1175036 | CC BY | 2021-01-04 16:02:37 | no | Arthritis Res Ther. 2005 May 17; 7(4):R868-R876 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1754 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17561598748710.1186/ar1756Research ArticleRegional assessment of articular cartilage gene expression and small proteoglycan metabolism in an animal model of osteoarthritis Young Allan A [email protected] Margaret M 1Smith Susan M 1Cake Martin A 2Ghosh Peter 1Read Richard A 2Melrose James 1Sonnabend David H 1Roughley Peter J 3Little Christopher B 11 Raymond Purves Research Laboratory, Institute of Bone and Joint Research, Royal North Shore Hospital, University of Sydney, St Leonards, New South Wales, Australia2 School of Veterinary and Biomedical Sciences, Murdoch University, Perth, Western Australia, Australia3 Genetics Unit, Shriners Hospital for Children, Montreal, Quebec, Canada2005 12 5 2005 7 4 R852 R861 23 1 2005 16 2 2005 9 4 2005 14 4 2005 Copyright © 2005 Young et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Osteoarthritis (OA), the commonest form of arthritis and a major cause of morbidity, is characterized by progressive degeneration of the articular cartilage. Along with increased production and activation of degradative enzymes, altered synthesis of cartilage matrix molecules and growth factors by resident chondrocytes is believed to play a central role in this pathological process. We used an ovine meniscectomy model of OA to evaluate changes in chondrocyte expression of types I, II and III collagen; aggrecan; the small leucine-rich proteoglycans (SLRPs) biglycan, decorin, lumican and fibromodulin; transforming growth factor-β; and connective tissue growth factor. Changes were evaluated separately in the medial and lateral tibial plateaux, and were confirmed for selected molecules using immunohistochemistry and Western blotting. Significant changes in mRNA levels were confined to the lateral compartment, where active cartilage degeneration was observed. In this region there was significant upregulation in expession of types I, II and III collagen, aggrecan, biglycan and lumican, concomitant with downregulation of decorin and connective tissue growth factor. The increases in type I and III collagen mRNA were accompanied by increased immunostaining for these proteins in cartilage. The upregulated lumican expression in degenerative cartilage was associated with increased lumican core protein deficient in keratan sulphate side-chains. Furthermore, there was evidence of significant fragmentation of SLRPs in both normal and arthritic tissue, with specific catabolites of biglycan and fibromodulin identified only in the cartilage from meniscectomized joints. This study highlights the focal nature of the degenerative changes that occur in OA cartilage and suggests that altered synthesis and proteolysis of SLRPs may play an important role in cartilage destruction in arthritis.
==== Body
Introduction
Articular cartilage exhibits unique hydrodynamic and viscoelastic properties that are largely attributable to its extracellular matrix (ECM), which equips diarthrodial joints with their weight-bearing properties and near frictionless articulation. Cartilage ECM is composed of a collagen network, predominantly type II, in which large chondroitin sulphate and keratan sulphate (KS) substituted proteoglycans (aggrecan) are entrapped. The negatively charged aggrecan glycosaminoglycan side-chains act to create an osmotic swelling pressure in the cartilage matrix that is resisted by tension developed in the collagen network [1]. The generation of a hydrostatic pressure within cartilage allows it to counteract the loads transmitted to it from the long bones during normal joint articulation.
The ECM of cartilage also contains the small leucine-rich proteoglycans (SLRPs) biglycan, decorin, fibromodulin and lumican, which have diverse functions as modulators of tissue organization, cellular proliferation, adhesion and responses to growth factors and cytokines [2,3]. The SLRPs all bind to fibrillar type I and/or II collagens [4-6] and, in the case of decorin, to fibromodulin and lumican; these interactions modulate the rate and ultimate diameter of collagen fibrils formed in vitro [7-9]. Decorin, biglycan and fibromodulin can also form complexes with transforming growth factor (TGF)-β and modulate the action of this growth factor [10,11]. The physical presence of the SLRPs, in addition to the minor type IX and XI collagens, on the surface of type II collagen fibrils has been proposed to restrict sterically the access of collagenases to sites of cleavage on the collagen fibrils [12]. Complexes of matrilin-1 and decorin or biglycan have also been reported to connect type VI collagen to aggrecan and type II collagen, further stabilizing the cartilage ECM [13]. It is evident that there is a complex interplay between the collagenous and proteoglycan components of the cartilage ECM that produces a biocomposite material with unique mechanical properties. Disruption of the normal balance of ECM components through altered synthesis or degradation will have important ramifications for the load-bearing capacity of cartilage.
Chondrocytes, the highly differentiated cells of cartilage, are responsible for maintaining a homeostatic balance between production and degradation of cartilage ECM [14,15]. The metabolic status of the chondrocyte is central to our understanding of the initiation and progression of osteoarthrits (OA) [16]. An initial anabolic response of chondrocytes in OA includes an upregulation of mRNA levels for the major structural components type II collagen and aggrecan, with an associated elevation in synthesis [17,18]. Degradation of the ECM is also elevated in these early stages in OA. Eventually, the biosynthetic machinery of the chondrocyte is unable to keep up with the anabolic demands and a net depletion of ECM occurs during the later stages of OA. Loss of key functional components combined with a disrupted architecture result in compromised tissue function, cell death and, eventually, cartilage loss down to subchondral bone.
Changes in SLRP metabolism in human OA are relatively poorly characterized, with both increased synthesis and degradation of individual molecules reported in arthritic human cartilage [19,20]. Their function within the collagen network means that changes in their tissue content may significantly alter the biomechanical integrity of cartilage. However, because SLRPs are also regulators of growth factor activity, changes in their synthesis and degradation may have significant effects on chondrocyte metabolism. It is unclear whether the changes in SLRP metabolism are restricted to the cartilage undergoing OA degeneration or are more generalized within arthritic joints. An understanding of the changes that occur with the onset and progression of cartilage degeneration in OA may provide important insights into potential regulatory steps in this process.
Animal models of OA have permitted longitudinal evaluation of spatial and temporal changes in joint tissues that occur during the development of joint disease. Total or partial removal of knee joint meniscus in humans commonly results in degeneration of articular cartilage, leading to osteoarthritic changes [21]. In sheep, lateral meniscectomy has been shown to reliably reproduce biochemical, biomechanical and histopathological alterations typical of OA [22,23]. In the present study we used this established model of OA to study the changes in expression of key structural molecules (aggrecan and type II collagen), the collagen-associated SLRPs (biglycan, decorin, lumican and fibromodulin), TGF-β1 and its associated downstream signaling molecule connective tissue growth factor (CTGF), and markers of altered chondrocyte phenotype – types I and III collagen. The expression levels were compared with protein levels in cartilage extracts or by immunohistochemistry in tissues with various histopathological grades of OA in the medial and lateral joint compartments.
Materials and methods
Animal model
Twelve 7-year-old female pure-bred Merino sheep were used in the present study. Six of the sheep underwent open lateral meniscectomy of both stifle joints, as previously described [24], whereas the remaining six served as nonoperated controls. Following recovery from surgery, the animals were maintained in an open paddock for 6 months before sacrifice. The protocol used for the present study was approved by the animal ethics committee of Murdoch University, Western Australia (AEC 832R/00).
Tissue preparation
Full depth articular cartilage from the medial tibial plateau (MTP) and lateral tibial plateau (LTP) was sampled from either the right or left stifle (knee) joint, randomly selected. Care was taken not to sample tissue from the joint margins or osteophytes. Tissue samples were snap frozen in liquid nitrogen before storage at -80°C until they were required. The tibial plateaux from the contralateral joints were isolated by a horizontal cut through the tibia below the epiphyseal growth plate using a band saw. Full thickness coronal osteochondral slabs (5 mm) were subsequently prepared through the mid weight-bearing region of the tibial plateau.
Histology
The coronal tibial osteochondral slices were fixed in 10% (vol/vol) neutral buffered formalin for 48 hours then decalcified in 10% formic acid (vol/vol)/5% formalin (vol/vol) for 5 days. The specimens were then dehydrated in graded alcohols and double-embedded in celloidin–paraffin blocks. Tissue sections (4 μm) were cut using a rotary microtome and attached to microscope slides. They were then deparaffinized in xylene and washed in graded alcohols to 70% (vol/vol) ethanol and then stained for 10 min with 0.04% (weight/vol) toluidine blue in 0.1 mol/l sodium acetate buffer (pH 4.0) to visualize the tissue proteoglycans. This was followed by 2 min counter-staining in an aqueous 0.1% (weight/vol) Food Drug and Cosmetic Green Nos. 3 stain. The slides were subsequently evaluated by bright field microscopy using a Leica MPS-60 (Leica Microsystems, Gladesville, New South Wales, Australia) photomicroscope system by two independent observers using a modified Mankin scoring scheme, previously developed in our laboratory for this ovine model [22]. In each compartment the worst score evident across the width of the tibial plateau was used to calculate the mean score for MTP and LTP of control and meniscectomized joints (n = 6 for each group).
Immunohistochemistry
Immunostaining was performed using monoclonal antibodies against type I collagen (ICN Biomedicals, Aurora, USA; code no. 63170; clone no. I-8H5) and type II collagen (ICN Biomedicals, North Ryde, New South Wales, Australia; code no. 63171; clone no. II-4CII), and a polyclonal antibody against type III collagen (Cedarlane, Hornby, Ontario, Canada; code no. CL50321AP). Endogenous peroxidase activity was initially blocked by incubating the tissue sections in 3% (vol/vol) H2O2 for 5 min and the sections were rinsed in TBS-Tween.
For type I and II collagen localizations, the sections were predigested with proteinase K (Dako, Glostrup, Denmark; code no. S3020) for 6 min at room temperature, followed by bovine testicular hyaluronidase (Sigma, St Louis, MO, USA; code no. H-3506) 1000 U/ml for 1 hour at 37°C in phosphate buffer (pH 5.0). The type III collagen localizations were predigested with hyaluronidase alone. The sections were then incubated in 10% (vol/vol) swine serum for 10 min at room temperature to block any nonspecific binding.
Incubations with the primary antibodies were performed overnight at 4°C with type I (5 μg/ml), type II (10 μg/ml) and type III (1:500 dilution) collagens. Detection of primary antibody was undertaken using a 20 min incubation with a cocktail of biotinylated anti-rabbit and anti-mouse immunoglobulin secondary antibodies (Dako; code no. K1015), followed by a 20 min incubation with streptavidin-conjugated horseradish peroxidase (Dako; code no. K0690). Staining was undertaken using NovaRED substrate (Vector, Burlingame, CA, USA; code no. SK-4800) for 15 min, which gives a red–brown end product. Sections were counter-stained in Mayer's haematoxylin for 1 min, washed in H2O, dehydrated in ethanol, cleared in xylene and mounted. Negative control sections were prepared using irrelevant isotype matched primary antibodies (Dako; code no. X931 or X0936) in place of authentic primary antibody.
RNA extraction
Approximately 100 mg of frozen cartilage samples was fragmented in a Mikro-Dismembrator (Braun Biotech International, Melsungen, Germany), 1 ml of TRIzol Reagent (Invitrogen Life Technologies, Carlsbad, CA, USA) was added, and the mixture was allowed to defrost to room temperature. Total RNA was isolated using the RNeasy Mini Kit from Qiagen (Valencia, CA, USA). Chloroform (300 μl) was subsequently added to the samples and the tubes vortexed vigorously before centrifugation to pellet the tissue residue. The clear supernatant solution (aqueous phase) was recovered and mixed by inversion with an equal volume of 70% ethanol, and then loaded onto spin columns. Following several washing steps and an on-column DNase digestion (Qiagen, Hilden, Germany), RNA was eluted from the column with 32 μl of RNAse free distilled H2O. Total RNA was quantified using a flourimeter (Perkin Elmer, Beaconsfield, UK) using SYBR® Green II colour reagent (Cambrex Bio Science, Rockland, ME, USA), and each sample was assessed for purity to confirm the absence of detectable DNA.
Semiquantitative RT-PCR
RT reactions were undertaken with 1 μg total RNA using the Omniscript RT kit from Qiagen (Germany). Using specific primer sets (Sigma Genosys, Castle Hill, New South Wales, Australia; Table 1), aliquots of cDNA were amplified by PCR, with initial denaturation at 94°C for 5 min, followed by cycles of 30 s of denaturation at 94°C, 30 s annealing at variable primer specific temperatures (Table 1), 30 s for extension at 72°C, and a further 7 min extension at 72°C on completion of the cycles. Reactions generated single PCR products that were identified by sequencing (SUPAMAC, Sydney, Australia) and specificity confirmed by BLAST searches. Cycle optimization was performed for each primer set before PCR, and for all reported experiments amplification levels were compared in the linear range of the PCR reaction. All samples underwent RT and cDNA amplification at the same time to avoid potential variations in experimental conditions.
The amplified products were electrophoresed on 2% (weight/vol) agarose gels, stained with ethidium bromide, imaged using a Fujifilm FLA-3000 fluorescent image analyzer and integrated densities calculated using One-Dscan, 1-D gel analysis software (Scanalytics, Fairfax, VA, USA). Sample loadings were normalized to the housekeeping gene GAPDH (glyceraldehyde-3-phosphate dehydrogenase) to permit semiquantitative comparisons in mRNA levels, as previously described [25,26].
Cartilage extraction, SDS-PAGE and Western blotting of the small leucine-rich proteoglycans
Pooled cartilage samples from all meniscectomized and nonoperated control LTPs were finely diced and extracted with 10 volumes of 4 mol/l GuCl and 50 mmol/l Tris HCl (pH 7.2) in the presence of proteinase inhibitors at 4°C with end over end stirring for 48 hours before dialysis of the extract against ultrapure water, as described previously [27]. Insufficient cartilage was available from MTPs for extraction and Western blot analyses. Dialysed extracts corresponding to equal dry weights of tissue were predigested with either chondroitinase ABC (Seikagaku) 0.1 U/ml alone or in combination with keratanase II (Seikagaku) 0.01 U/ml and endo-β-galactosidase (Seikagaku, Tokyo, Japan) 0.01 U/ml in 0.1 mol/l Tris/0.1 mol/l sodium acetate (pH 7.0) overnight at 37°C before electrophoresis. Electrophoresis was conducted under reducing conditions in 10% NuPAGE Bis-Tris resolving gels (Invitrogen), using MOPS SDS running buffer at 125 V constant voltage for 1 hour. The gels were then electroblotted to nitrocellulose membranes in NuPAGE transfer buffer with 20% (vol/vol) methanol at 200 mA for 2 hours and blocked overnight in 5% (weight/vol) BSA in 50 mmol/l Tris-HCl (pH 7.2) and 0.15 mol/l NaCl 0.02% (weight/vol) NaN3 (TBS-azide). The blots were probed overnight with affinity purified polyclonal antibodies directed against the carboxyl-terminus of decorin, biglycan, fibromodulin and lumican (0.3–1 μg/ml) [12] followed by washing in TBS-azide and detection using alkaline phosphatase conjugated anti-rabbit secondary antibodies and the nitro blue tetrazolium/4-bromo-1-chloro-indolyl phosphate substrate system (BioRad, Hercules, CA, USA). A sample of human OA cartilage harvested from the tibial plateau at the time of joint replacement surgery also underwent identical processing as a positive control.
Statistical analysis
All RT-PCR data were normalized to the housekeeping gene GAPDH (glyceraldehyde-3-phosphate dehydrogenase) to facilitate equal loading of gels for quantitative comparisons of amplified PCR products. Comparison of parametric data from the nonoperated and meniscectomized sample groups were undertaken using the unpaired Student's t-test with Benjamini–Hochberg correction [28] for multiple comparisons. Comparisons of nonparametric data from the modified Mankin histological scoring of the stained tissue sections were assessed using the Mann–Whitney U-test.
Results
Histology
Lateral meniscectomy resulted in macroscopic joint changes characteristic of the early and middle phases of OA with cartilage fibrillation and erosion, in addition to formation of marginal osteophytes, particularly in the lateral compartment (Fig. 1f; arrowheads). The histopathological lesions varied between animals, between medial and lateral joint compartments, and across the width of the tibial plateaux. A significant loss of proteoglycan was evident in the superficial cartilage of both the LTP (Fig. 1d, e) and, to a lesser extent, the MTP of the meniscectomized joints (Fig. 1b) compared with nonoperated controls (Fig. 1a, c). Chondrocyte cloning was also a prominent feature in the LTP specimens after meniscectomy (Fig. 1d; asterisk), which is in keeping with the validity of this model's representation of human OA. The most severe lesions were confined to the weight-bearing region of the LTP, with significant proteoglycan loss and surface fibrillation (Fig. 1e, f).
Histological grading of the meniscectomized and nonoperated control cartilage specimens confirmed and quantitated the histological observations. In control sheep the modified Mankin score (mean ± standard deviation) was significantly higher in the MTP specimens than in the LTP ones (9.3 ± 1.9 versus 3.1 ± 1.1; P < 0.01). Following meniscectomy there was a slight although not statistically significant change in the modified Mankin score for the MTP specimens (10.7 ± 3.3). The same could not be said of the LTP specimens, in which meniscectomy resulted in a significant increase from 3.1 ± 1.1 to 23.3 ± 1.8 (P < 0.01).
Immunolocalization of types I, II and III collagens
An increase in type I collagen matrix immunostaining was evident following meniscectomy in the most superficial cartilage of the LTP specimens (Fig. 2d) and, to a lesser extent, in the MTP specimens (Fig. 2b), corresponding to areas of degenerative change. In nonoperated control sections (Fig. 2a, c), type I collagen was restricted to the uppermost surface lamina, as reported previously [29]. Type III collagen, which is typically seen pericellularly in normal cartilage [30], also exhibited increased matrix staining after meniscectomy (Fig. 2j, l) compared with nonoperated control (Fig. 2i, k). Type II collagen was immunolocalized in the matrix throughout the depth of the cartilage in both MTP and LTP, and there was a generalized decrease in staining following meniscectomy (Fig. 2e–h). As expected [31,32], types I and III collagens were also prominently immunolocalized in the marginal osteophytic fibrocartilaginous regions in the meniscectomized joints (data not shown).
RT-PCR
It was not possible to undertake all procedures with some of the cartilage samples that did not yield at least 1 μg total RNA. This resulted in four samples being excluded, all from MTP cartilage (one from the meniscectomy group and three from the nonoperated control group). Statistical comparisons of mRNA levels following meniscectomy as a percentage of control values were undertaken separately for LTP and MTP cartilages and are presented graphically in Fig. 3. Following lateral meniscectomy, mRNA levels in LTP cartilage were found to be upregulated for the following molecules: aggrecan (1.5 fold; P < 0.01), type I collagen (11.7-fold; P < 0.01), type II collagen (3.9-fold; P < 0.01), type III collagen (2.3-fold; P < 0.05), biglycan (1.8-fold; P < 0.01) and lumican (14.6-fold; P < 0.01). In the same region there were downregulations of decorin (1.6-fold; P < 0.01) and CTGF (2.1-fold; P < 0.05), and unchanged expression of fibromodulin and TGF-β. In the MTP cartilage samples, none of the changes in mRNA levels following meniscectomy relative to nonoperated controls were statistically significant.
Western blotting of the small leucine-rich proteoglycans
Western blot analysis of extracts of an equivalent dry weight of pooled LTP cartilage from control and meniscectomized joints and OA human cartilage are shown in Fig. 4. There was little difference in total staining intensity between nonoperated and meniscectomized cartilage for the 45 kDa intact core protein of decorin that was also evident in human OA cartilage. Additional fragmented forms of decorin core protein (32 and 20 kDa) were evident in the cartilage extracts from both the control and meniscectomy specimens, whereas a 40 kDa fragment was identified only in meniscectomized cartilage extracts (Fig. 4; asterisk). Blotting for biglycan identified intact core protein (43 kDa) and a number of fragments (39, 32, 28 and 26 kDa) in all of the specimens. There was an increase in staining intensity for all biglycan core protein species in meniscectomized cartilage. The predominant fibromodulin core protein species identified in all specimens was about 55 kDa in size, with a slight increase in staining following meniscectomy. This 55 kDa fibromodulin band is consistent with full-length core protein [12]. A 28 kDa fibromodulin fragment was detected only in the extract from meniscectomized joints (Fig. 4; asterisk). Lumican electrophoresed as two predominant species, a 60–64 kDa band with similar staining intensities evident in control and meniscectomy extracts. A smaller, approximately 50 kDa band, which was the predominant species in the human OA sample, exhibited greater staining intensity after meniscectomy compared with cartilage from nonoperated joints. Removal of KS side-chains with keratanase II/endo-β-galactosidase treatment resulted in all of the lumican migrating at 50 kDa, suggesting that the 60–64 kDa band represented KS substituted lumican.
Discussion
Our laboratory previously reported biochemical, biomechanical and histological changes that occur in the articular cartilage in the ovine lateral meniscectomy model of OA [22,23,33]. The present study extends these earlier investigations by examining the expression of a number of important extracellular matrix components at the mRNA level. One of the difficulties we encountered was relatively low average RNA yields (0.85–9.13 μg per 100 mg), which resulted in exclusion of some MTP samples. Studies utilizing other animal models of OA have reported RNA yields from 2.5 to 21 μg/100 mg of normal cartilage [34,35], but the animals used in those studies (rabbit and canine) were of a much younger age than ours. Studies using aged human cartilage report much lower average yields, ranging from 0.669 to 0.839 μg/100 mg of OA and 'normal' cartilage [36]. We attributed the low RNA yield in our study to our use of an aged population of sheep, although other factors such as species differences, RNA degradation and technical factors cannot be excluded. Although we were able to analyze medial and lateral tibial cartilage separately, the low yields of RNA from the older sheep precluded further topographical separation. Future studies using younger animals may permit analysis of affected and unaffected cartilage within one joint area.
Although morphological and histological changes in cartilage were most notable in the lateral compartment, changes in the medial femoro-tibial joint were nevertheless still evident but of a markedly lesser magnitude, as previously reported [23,37]. In the present study the MTP cartilage in control joints had significantly worse histopathological scores than did LTP from the same joints, which is consistent with age-related change in the more heavily loaded compartment of these old animals. The histopathology scores did not increase significantly in the medial compartment following meniscectomy, and this was consistent with the lack of change in mRNA levels. Our inability to detect differences in mRNA expression in the medial compartment might have resulted from the small number of samples evaluated. However, the standard deviation of the MTP samples was similar to that of the LTP, suggesting that the lower number of MTPs studied did not contribute to the lack of statistical significance. Changes in mRNA levels for a number of molecules were significant in the lateral compartment following meniscectomy. Although our findings are limited to a single time point following induction of OA, restriction of significant alterations in gene expression to the LTP indicates that the changes observed were likely associated with active degradation of cartilage primarily due to altered biomechanical forces rather than humoral factors.
In the present study the changes observed in the expression of aggrecan and type II collagen probably reflect an anabolic response by the chondrocytes to the altered mechanical stresses imposed by this surgical procedure, as well as early OA degeneration. The increase in expression is consistent with an attempted 'repair' response in early OA, as described in other animal models [34,35,38]. Levels of mRNA for a particular molecule may not reflect protein synthesis or its accumulation in tissue, with post-transcriptional regulation and post-translational processing playing significant roles. Indeed, we previously demonstrated increased degradation of newly synthesized aggrecan in cartilage after lateral meniscectomy in sheep [24]. Furthermore, the changes in mRNA levels observed in the present study were representative of the entire MTPs or LTPs and therefore probably included cartilage from areas with different stages of OA.
In addition to the increase in mRNA for the major cartilage matrix components aggrecan and type II collagen, significant increases in expression and protein levels of types I and III collagen were observed following meniscectomy. Type III collagen is present pericellularly in small amounts in normal articular cartilage [16,30], and type I collagen is is evident in the most superficial layer [29]. Contrary to early reports [39], evidence now suggests that both types I and III collagens are significantly increased in OA cartilage, both at the expression and protein levels [40,41]. It has been suggested that a major phenotypic shift occurs in OA toward a de-differentiated chondrocyte [40]. Interestingly, in the present study we observed increased amounts of types I and III collagens by immunohistology in both compartments following meniscectomy, despite increased mRNA levels only being evident in the lateral compartment. A probable explanation was that the increased types I and III collagens observed with immunohistochemistry represented the cumulative changes throughout the course of the disease process while expression levels reflected chondrocyte metabolism at a specific point in time (i.e. 6 months following meniscectomy). Changes in collagen subtypes in pathological cartilage may not only influence the biomechanical integrity of the tissue but may sequester and modulate the actions of cytokines, with types I and III collagen shown to bind oncostatin M specifically [42].
Selective modulation of SLRP mRNA levels in OA cartilage was observed in the present study, with increased biglycan and lumican, decreased decorin, and little or no change in fibromodulin. Additionally, we have shown for the first time that these changes in SLRP expression are confined to the cartilage in the compartment undergoing active OA degeneration. The differential regulation contrasts with the reported increase in expression of all four SLRPs in late-stage human OA in one study [19], but it is consistent with another study [43] that reported no change in decorin but increased biglycan message in late stage OA. In the canine anterior cruciate ligament transection model, increased cartilage mRNA for biglycan, decorin and fibromodulin have been described [38,44]. The reported differences in mRNA expression may relate to variable stages of disease, methods of quantitation and species evaluated.
The SLRPs have been shown to influence cartilage metabolism indirectly via actions on growth factors such as TGF-β, which they inactivate through sequestration and thereby potentially mitigate its effects in OA [11,45]. Although we found no change in the expression of TGF-β following meniscectomy, there was a significant decrease in mRNA levels of CTGF. We speculate that sequestration of TGF-β by the SLRPs may have accounted for the decrease in CTGF expression. Our results contrast with human cartilage, in which an increase in CTGF in OA was recently reported [46], and this could be associated with species differences or the stage of disease. CTGF, a secretory protein involved in fibrotic response mechanisms in tissues, is an important downstream effector of TGF-β [47] and is thought to be involved in promoting the proliferation and/or differentiation of chondrocytes [48-51]. Further investigation of the specific relationships between growth factors, collagens and the SLRPs in normal and diseased cartilage is warranted.
Significant proteolysis of the SLRPs was evident in the present study. SLRP degradation was previously reported in both human OA [20] and spontaneous canine OA [52], but not in a canine cruciate ligament transection model of OA [52]. The catabolites that were identified in meniscectomized cartilage in the present study were also generally evident in normal cartilage, indicating similar proteolytic processes in health and disease. However, in the case of decorin and fibromodulin, fragments unique to the meniscectomized cartilage were identified, suggesting the presence of disease-specific proteolytic processes. In this regard, a specific proteolytic fragment of fibromodulin was recently identified from interleukin-1 stimulated but not normal cartilage [53]. The cleavage site(s) and proteinase(s) responsible for extracellular SLRP breakdown in arthritic cartilage have yet to be identified and are the subject of further investigation.
A particularly novel finding in the present study was the increased lumican core protein present in degenerative cartilage following meniscectomy, which is consistent with the significant increase observed in mRNA levels. Furthermore, the increased lumican observed by Western blotting was present in a non-KS substituted form. Limited studies [19,54] have suggested that lumican primarily exists lacking KS in adult cartilage, but cultured chondrocytes have been observed to produce a KS-substituted form that appeared to be the default synthesis preference [55]. The catabolic cytokine interleukin-1β, which may be present in OA joints, stimulates secretion of lumican deficient in KS [55]. It has been shown that OA chondrocytes synthesize SLRPs that are differently glycosylated, and that nonglycosylated biglycan and decorin are more abundant in OA cartilage [20]. Changes in glycosylation of the SLRPs, whether by altered synthesis or subsequent degradation, are likely to influence the functional properties of these molecules in cartilage.
Conclusion
We showed that degradation of cartilage in OA is associated with significant focal changes in expression and content of matrix proteins. Accelerated proteolysis of aggrecan and type II collagen overwhelms the increase in expression of these major structural proteins. Furthermore, there is a shift in chondrocyte phenotype, with increased synthesis of collagens types I and III and a change in the relative levels of the fibril-associated SLRPs. In particular there is decrease in synthesis of decorin and an increase in biglycan and lumican, with the latter lacking KS substitution. It seems likely that the altered pattern of SLRP synthesis, which is localized to the diseased joint compartment, along with an increase in SLRP proteolysis, modifies the biomechanical properties of the matrix and contributes to cartilage breakdown. Changes in SLRP levels could also significantly modulate the action of potential anabolic factors such as TGF-β and its downstream effector CTGF, possibly adding to disease development. An understanding of the relationship between SLRP metabolism and progressive cartilage breakdown in OA may provide both novel diagnostic markers of disease and therapeutic targets for the treatment of this disorder.
Abbreviations
CTGF = connective tissue growth factor; ECM = extracellular matrix; KS = keratan sulphate; LTP = lateral tibial plateau; MTP = medial tibial plateau; OA = osteoarthritis; RT-PCR = reverse transcription polymerase chain reaction; SLRP = small leucine-rich proteoglycan; TGF = transforming growth factor.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
AAY conducted the RT-PCR and Western blotting studies and drafted the manuscript. MMS designed primers for RT-PCR, performed histopathological cartilage scoring and helped to draft the manuscript. SMS performed histological and immunohistological preparations, and helped to draft the manuscript. MAC performed animal surgery and helped to draft the manuscript. RAR performed animal surgery and helped to draft the manuscript. PG made substantial contributions to the conception and design of the study. JM assisted with performing Western blotting studies and helped to draft the manuscript. DHS was involved in the conception and design of the study, and interpretation of the data, and critically revised the manuscript for important intellectual content. PJR assisted with Western blotting studies and critically revised the manuscript for important intellectual content. CBL performed histopathological cartilage scoring, analyzed and interpreted the data, and critically revised the manuscript for important intellectual content. All authors read and approved the final manuscript.
Acknowledgements
This study was funded by a research grant from the Australian Orthopaedic Association Research Foundation Ltd, whose support is gratefully acknowledged. The authors thank Diana Pethick of Murdoch University for her assistance with the animal handling and care.
Figures and Tables
Figure 1 Histology. Representative histology of medial (a, b) and (c–f) lateral tibial plateau cartilage from nonoperated control (panels a and c) and meniscectomized (panels b and d-f) ovine stifle joints. Cell cloning is a prominent feature in the lateral tibial plateau following meniscectomy (asterisk). Osteophyte formation is evident at the lateral joint margin (panel f, arrowheads), and the area of most severe cartilage damage with surface fibrillation (rectangle, panel e) and the adjacent area (circle, panel d) are indicated. Toluidine blue-fast green stain. Scale bar: 250 μm.
Figure 2 Immunolocalisation. Immunolocalization of types I (a–d), II (e–h) and III (I–l) collagens in medial (panels a, b, e, f, I and j) and lateral (panels c, d, g, h, k and l) tibial plateau cartilage. Sections from representative nonoperated control (panels a, c, e, g, I and k) and meniscectomized (b, d, f, h, j and l) joints are shown. Scale bar: 250 μm.
Figure 3 Changes in mRNA levels. Changes in (a) lateral tibial plateau (LTP) and (b) medial tibial plateau (MTP) cartilage mRNA levels of aggrecan, type I, II, and III collagen, decorin, biglycan, fibromodulin, lumican, transforming growth factor (TGF)-β and connective tissue growth factor (CTGF) following lateral meniscectomy (MEN) relative to nonoperated control (NOC) values. Values are expressed as mean ± standard deviation. There were three samples for the NOC MTP, six for the NOC LTP, five for the MEN MTP and six for the MEN LTP groups. *P < 0.05, **P < 0.01.
Figure 4 Western blot. Western blot analysis of decorin, biglycan, fibromodulin and lumican in extracts of human osteoarthritis cartilage (OA), nonoperated control (NOC) and lateral meniscectomized (MEN) ovine cartilage samples. Core protein fragments of decorin and fibromodulin that were only identified in MEN are marked with an asterisk. Equivalent amounts of extract from equal dry weights of tissue were loaded per lane following treatment with chondroitinase ABC (ChABC). Additionally, Western blot analysis of lumican was performed following treatment with ChABC, endo-β-galactosidase (EBG) and keratanase II (KII). The migration positions of prestained protein standards are indicated on the left.
Table 1 Primers used for RT-PCR
Gene Annealing temperature (°C) Product size (base pairs) Sequence (5' to 3') GenBank accession number
Collagen II 65 141 F ACGGTGGACGAGGTCTGACT
R GGCCTGTCTCTCCACGTTCA AF138883
Aggrecan 65 375 F CCGCTATGACGCCATCTGCT
R TGCACGACGAGGTCCTCACT AF019758
Decorin 55 319 F CAAACTCTTTTGCTTGGGCT
R CACTGGACAACTCGCAGATG AF125041
Biglycan 65 204 F CCATGCTGAACGATGAGGAA
R CATTATTCTGCAGGTCCAGC AF034842
Fibromodulin 65 442 F CTGGACCACAACAACCTGAC
R GGATCTTCTGCAGCTGGTTG AF020291
Lumican 65 284 F CAGCCATGTACTGCGATGAG
R CTGCAGGTCCACCAGAGATT NM173934
TGF-β 60 271 F CGGCAGCTGTACATTGACTT
R AGCGCACGATCATGTTGGAC AF000133
CTGF 65 504 F TCTTCTGCGACTTCGGCTCC
R CCTCCAGGTCAGCTTCGCAA NM174030
Collagen I 65 460 F CCACCAGTCACCTGCGTACA
R GGAGACCACGAGGACCAGAA AF129287
Collagen III 55 243 F GCTGGCTAGTTGTCGCTCTG
R GTGGGGAAACTGCACAACAT L47641
GAPDH 55 320 F TCACCATCTTCCAGGAGCGA
R GGCGTGGACAGTGGTCATAA AF035421
Shown are the details of the primers used for RT-PCR, including annealing temperatures, size of the amplified products, forward (F) and reverse (R) sequences, and primer source. CTCG, connective tissue growth factor; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; TGF, transforming growth factor.
==== Refs
Maroudas AI Balance between swelling pressure and collagen tension in normal and degenerate cartilage Nature 1976 260 808 809 1264261
Iozzo RV The biology of the small leucine-rich proteoglycans. Functional network of interactive proteins J Biol Chem 1999 274 18843 18846 10383378 10.1074/jbc.274.27.18843
Svensson L Oldberg A Heinegard D Collagen binding proteins Osteoarthritis Cartilage 2001 9 S23 S28 11680683 10.1053/joca.2001.0440
Schonherr E Hausser H Beavan L Kresse H Decorin-type I collagen interaction. Presence of separate core protein-binding domains J Biol Chem 1995 270 8877 8883 7721795 10.1074/jbc.270.15.8877
Hedbom E Heinegard D Binding of fibromodulin and decorin to separate sites on fibrillar collagens J Biol Chem 1993 268 27307 27312 8262971
Svensson L Narlid I Oldberg A Fibromodulin and lumican bind to the same region on collagen type I fibrils FEBS Lett 2000 470 178 182 10734230 10.1016/S0014-5793(00)01314-4
Vogel KG Paulsson M Heinegard D Specific inhibition of type I and type II collagen fibrillogenesis by the small proteoglycan of tendon Biochem J 1984 223 587 597 6439184
Font B Eichenberger D Goldschmidt D Boutillon MM Hulmes DJ Structural requirements for fibromodulin binding to collagen and the control of type I collagen fibrillogenesis: critical roles for disulphide bonding and the C-terminal region Eur J Biochem 1998 254 580 587 9688269 10.1046/j.1432-1327.1998.2540580.x
Chakravarti S Magnuson T Lass JH Jepsen KJ LaMantia C Carroll H Lumican regulates collagen fibril assembly: skin fragility and corneal opacity in the absence of lumican J Cell Biol 1998 141 1277 1286 9606218 10.1083/jcb.141.5.1277
Burton-Wurster N Liu W Matthews GL Lust G Roughley PJ Glant TT Cs-Szabo G TGF beta 1 and biglycan, decorin, and fibromodulin metabolism in canine cartilage Osteoarthritis Cartilage 2003 11 167 176 12623288 10.1053/S1063-4584(02)00349-7
Hildebrand A Romaris M Rasmussen LM Heinegard D Twardzik DR Border WA Ruoslahti E Interaction of the small interstitial proteoglycans biglycan, decorin and fibromodulin with transforming growth factor beta Biochem J 1994 302 527 534 8093006
Sztrolovics R White RJ Poole AR Mort JS Roughley PJ Resistance of small leucine-rich repeat proteoglycans to proteolytic degradation during interleukin-1-stimulated cartilage catabolism Biochem J 1999 339 571 577 10215595 10.1042/0264-6021:3390571
Wiberg C Klatt AR Wagener R Paulsson M Bateman JF Heinegard D Morgelin M Complexes of matrilin-1 and biglycan or decorin connect collagen VI microfibrils to both collagen II and aggrecan J Biol Chem 2003 278 37698 37704 12840020 10.1074/jbc.M304638200
Muir H The chondrocyte, architect of cartilage: biomechanics, structure, function and molecular-biology of cartilage matrix macromolecules Bioessays 1995 17 1039 1048 8634065 10.1002/bies.950171208
Urban J The chondrocyte: a cell under pressure Br J Rheumatol 1994 33 901 908 7921748
Eyre D Collagen of articular cartilage Arthritis Res 2002 4 30 35 11879535 10.1186/ar380
Rizkalla G Reiner A Bogoch E Poole AR Studies of the articular cartilage proteoglycan aggrecan in health and osteoarthritis. Evidence for molecular heterogeneity and extensive molecular changes in disease J Clin Invest 1992 90 2268 2277 1281828
Matyas JR Adams ME Huang D Sandell LJ Discoordinate gene expression of aggrecan and type II collagen in experimental osteoarthritis Arthritis Rheum 1995 38 420 425 7533495
Cs-Szabo G Melching LI Roughley PJ Glant TT Changes in messenger RNA and protein levels of proteoglycans and link protein in human osteoarthritic cartilage samples Arthritis Rheum 1997 40 1037 1045 9182913
Cs-Szabo G Roughley PJ Plaas AH Glant TT Large and small proteoglycans of osteoarthritic and rheumatoid articular cartilage Arthritis Rheum 1995 38 660 668 7538297
Roos H Lauren M Adalberth T Roos EM Jonsson K Lohmander LS Knee osteoarthritis after meniscectomy: prevalence of radiographic changes after twenty-one years, compared with matched controls Arthritis Rheum 1998 41 687 693 9550478 10.1002/1529-0131(199804)41:4<687::AID-ART16>3.0.CO;2-2
Little C Smith S Ghosh P Bellenger C Histomorphological and immunohistochemical evaluation of joint changes in a model of osteoarthritis induced by lateral meniscectomy in sheep J Rheumatol 1997 24 2199 2209 9375884
Appleyard RC Burkhardt D Ghosh P Read R Cake M Swain MV Murrell GA Topographical analysis of the structural, biochemical and dynamic biomechanical properties of cartilage in an ovine model of osteoarthritis Osteoarthritis Cartilage 2003 11 65 77 12505489 10.1053/joca.2002.0867
Little CBGP Bellenger CR Topographic variation in biglycan and decorin synthesis by articular cartilage in the early stages of osteoarthritis: an experimental study in sheep J Orthop Res 1996 14 433 444 8676257 10.1002/jor.1100140314
Marchuk L Sciore P Reno C Frank CB Hart DA Postmortem stability of total RNA isolated from rabbit ligament, tendon and cartilage Biochim Biophys Acta 1998 1379 171 177 9528651
Boykiw R Sciore P Reno C Marchuk L Frank CB Hart DA Altered levels of extracellular matrix molecule mRNA in healing rabbit ligaments Matrix Biol 1998 17 371 378 9822203 10.1016/S0945-053X(98)90089-0
Little CB Flannery CR Hughes CE Mort JS Roughley PJ Dent C Caterson B Aggrecanase versus matrix metalloproteinases in the catabolism of the interglobular domain of aggrecan in vitro Biochem J 1999 344 61 68 10548534 10.1042/0264-6021:3440061
Benjamini Y Hochberg Y Controlling the false discovery rate: a practical and powerful approach to multiple testing J Roy Stat Soc Ser 1995 57 289 300
Teshima R Ono M Yamashita Y Hirakawa H Nawata K Morio Y Immunohistochemical collagen analysis of the most superficial layer in adult articular cartilage J Orthop Sci 2004 9 270 273 15168182 10.1007/s00776-004-0769-4
Wotton SF Duance VC Type III collagen in normal human articular cartilage Histochem J 1994 26 412 416 8045781 10.1007/BF00160053
Gelse K Soder S Eger W Diemtar T Aigner T Osteophyte development – molecular characterization of differentiation stages Osteoarthritis Cartilage 2003 11 141 148 12554130 10.1053/joca.2002.0873
Aigner T Dietz U Stoss H von der Mark K Differential expression of collagen types I, II, III, and X in human osteophytes Lab Invest 1995 73 236 243 7637324
Little CB Ghosh P Bellenger CR Topographic variation in biglycan and decorin synthesis by articular cartilage in the early stages of osteoarthritis: an experimental study in sheep J Orthop Res 1996 14 433 444 8676257 10.1002/jor.1100140314
Hellio Le Graverand MP Eggerer J Vignon E Otterness IG Barclay L Hart DA Assessment of specific mRNA levels in cartilage regions in a lapine model of osteoarthritis J Orthop Res 2002 20 535 544 12038628 10.1016/S0736-0266(01)00126-7
Matyas JR Ehlers PF Huang D Adams ME The early molecular natural history of experimental osteoarthritis. I. Progressive discoordinate expression of aggrecan and type II procollagen messenger RNA in the articular cartilage of adult animals Arthritis Rheum 1999 42 993 1002 10323456 10.1002/1529-0131(199905)42:5<993::AID-ANR19>3.0.CO;2-U
McKenna LA Gehrsitz A Soder S Eger W Kirchner T Aigner T Effective isolation of high-quality total RNA from human adult articular cartilage Anal Biochem 2000 286 80 85 11038277 10.1006/abio.2000.4779
Hwa SY Burkhardt D Little C Ghosh P The effects of orally administered diacerein on cartilage and subchondral bone in an ovine model of osteoarthritis J Rheumatol 2001 28 825 834 11327258
Adams ME Matyas JR Huang D Dourado GS Expression of proteoglycans and collagen in the hypertrophic phase of experimental osteoarthritis J Rheumatol Suppl 1995 43 94 97 7752150
Aigner T Bertling W Stoss H Weseloh G von der Mark K Independent expression of fibril-forming collagens I, II, and III in chondrocytes of human osteoarthritic cartilage J Clin Invest 1993 91 829 837 7680669
Martin I Jakob M Schafer D Dick W Spagnoli G Heberer M Quantitative analysis of gene expression in human articular cartilage from normal and osteoarthritic joints Osteoarthritis Cartilage 2001 9 112 118 11237658 10.1053/joca.2000.0366
Miosge N Hartmann M Maelicke C Herken R Expression of collagen type I and type II in consecutive stages of human osteoarthritis Histochem Cell Biol 2004 122 229 236 15316793 10.1007/s00418-004-0697-6
Somasundaram R Ruehl M Schaefer B Schmid M Ackermann R Riecken EO Zeitz M Schuppan D Interstitial collagens I, III, and VI sequester and modulate the multifunctional cytokine oncostatin M J Biol Chem 2002 277 3242 3246 11711546 10.1074/jbc.M110011200
Aigner T Zien A Gehrsitz A Gebhard PM McKenna L Anabolic and catabolic gene expression pattern analysis in normal versus osteoarthritic cartilage using complementary DNA-array technology Arthritis Rheum 2001 44 2777 2789 11762938 10.1002/1529-0131(200112)44:12<2777::AID-ART465>3.0.CO;2-H
Dourado GS Adams ME Matyas JR Huang D Expression of biglycan, decorin and fibromodulin in the hypertrophic phase of experimental osteoarthritis Osteoarthritis Cartilage 1996 4 187 196 8895220
Markmann A Hausser H Schonherr E Kresse H Influence of decorin expression on transforming growth factor-beta-mediated collagen gel retraction and biglycan induction Matrix Biol 2000 19 631 636 11102752 10.1016/S0945-053X(00)00097-4
Omoto S Nishida K Yamaai Y Shibahara M Nishida T Doi T Asahara H Nakanishi T Inoue H Takigawa M Expression and localization of connective tissue growth factor (CTGF/Hcs24/CCN2) in osteoarthritic cartilage Osteoarthritis Cartilage 2004 12 771 778 15450526 10.1016/j.joca.2004.06.009
Moussad EE Brigstock DR Connective tissue growth factor: what's in a name? Mol Genet Metab 2000 71 276 292 11001822 10.1006/mgme.2000.3059
Shakunaga T Ozaki T Ohara N Asaumi K Doi T Nishida K Kawai A Nakanishi T Takigawa M Inoue H Expression of connective tissue growth factor in cartilaginous tumors Cancer 2000 89 1466 1473 11013359 10.1002/1097-0142(20001001)89:7<1466::AID-CNCR8>3.0.CO;2-G
Yosimichi G Nakanishi T Nishida T Hattori T Takano-Yamamoto T Takigawa M CTGF/Hcs24 induces chondrocyte differentiation through a p38 mitogen-activated protein kinase (p38MAPK), and proliferation through a p44/42 MAPK/extracellular-signal regulated kinase (ERK) Eur J Biochem 2001 268 6058 6065 11732999 10.1046/j.0014-2956.2001.02553.x
Nishida T Kubota S Fukunaga T Kondo S Yosimichi G Nakanishi T Takano-Yamamoto T Takigawa M CTGF/Hcs24, hypertrophic chondrocyte-specific gene product, interacts with perlecan in regulating the proliferation and differentiation of chondrocytes J Cell Physiol 2003 196 265 275 12811819 10.1002/jcp.10277
Nakanishi T Nishida T Shimo T Kobayashi K Kubo T Tamatani T Tezuka K Takigawa M Effects of CTGF/Hcs24, a product of a hypertrophic chondrocyte-specific gene, on the proliferation and differentiation of chondrocytes in culture Endocrinology 2000 141 264 273 10614647 10.1210/en.141.1.264
Liu W Burton-Wurster N Glant TT Tashman S Sumner DR Kamath RV Lust G Kimura JH Cs-Szabo G Spontaneous and experimental osteoarthritis in dog: similarities and differences in proteoglycan levels J Orthop Res 2003 21 730 737 12798075 10.1016/S0736-0266(03)00002-0
Heathfield TF Onnerfjord P Dahlberg L Heinegard D Cleavage of fibromodulin in cartilage explants involves removal of the N-terminal tyrosine sulfate-rich region by proteolysis at a site that is sensitive to matrix metalloproteinase-13 J Biol Chem 2004 279 6286 6295 14660626 10.1074/jbc.M307765200
Grover J Chen XN Korenberg JR Roughley PJ The human lumican gene. Organization, chromosomal location, and expression in articular cartilage J Biol Chem 1995 270 21942 21949 7665616 10.1074/jbc.270.37.21942
Melching LI Roughley PJ Modulation of keratan sulfate synthesis on lumican by the action of cytokines on human articular chondrocytes Matrix Biol 1999 18 381 390 10517185 10.1016/S0945-053X(99)00033-5
| 15987487 | PMC1175037 | CC BY | 2021-01-04 16:02:38 | no | Arthritis Res Ther. 2005 May 12; 7(4):R852-R861 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1756 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17571598748810.1186/ar1757Research ArticleReliability of computerized image analysis for the evaluation of serial synovial biopsies in randomized controlled trials in rheumatoid arthritis Haringman Jasper J [email protected] Marjolein 1Gerlag Danielle M 1Smeets Tom JM 1Zwinderman Aeilko H 2Tak Paul P [email protected] Division of Clinical Immunology and Rheumatology, Academic Medical Center/University of Amsterdam, The Netherlands2 Department of Clinical Epidemiology and Biostatistics, Academic Medical Center/University of Amsterdam, The Netherlands2005 12 5 2005 7 4 R862 R867 14 9 2004 4 11 2004 13 4 2005 14 4 2005 Copyright © 2005 Haringman et al. licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Analysis of biomarkers in synovial tissue is increasingly used in the evaluation of new targeted therapies for patients with rheumatoid arthritis (RA). This study determined the intrarater and inter-rater reliability of digital image analysis (DIA) of synovial biopsies from RA patients participating in clinical trials. Arthroscopic synovial biopsies were obtained before and after treatment from 19 RA patients participating in a randomized controlled trial with prednisolone. Immunohistochemistry was used to detect CD3+ T cells, CD38+ plasma cells and CD68+ macrophages. The mean change in positive cells per square millimetre for each marker was determined by different operators and at different times using DIA. Nonparametric tests were used to determine differences between observers and assessments, and to determine changes after treatment. The intraclass correlations (ICCs) were calculated to determine the intrarater and inter-rater reliability. Intrarater ICCs showed good reliability for measuring changes in T lymphocytes (R = 0.87), plasma cells (R = 0.62) and macrophages (R = 0.73). Analysis by Bland–Altman plots showed no systemic differences between measurements. The smallest detectable changes were calculated and their discriminatory power revealed good response in the prednisolone group compared with the placebo group. Similarly, inter-rater ICCs also revealed good reliability for measuring T lymphocytes (R = 0.68), plasma cells (R = 0.69) and macrophages (R = 0.72). All measurements identified the same cell types as changing significantly in the treated patients compared with the placebo group. The measurement of change in total positive cell numbers in synovial tissue can be determined reproducibly for various cell types by DIA in RA clinical trials.
==== Body
Introduction
Rheumatoid arthritis (RA) is characterized by chronic and symmetric inflammation of synovial joints [1,2]. Although the aetiology of RA is still unknown, it is thought of as an autoimmune disease with the synovial tissue (ST) being its primary target. The microscopic appearance of RA ST includes marked intimal lining layer hyperplasia due to increased numbers of fiboblast-like synoviocytes and intimal macrophages, and accumulation of macrophages, T cells, B cells, plasma cells, dendritic cells, mast cells, natural killer cells and neutrophils in the synovial sublining layer [3]. Developments in synovial biopsy techniques, especially arthroscopy, have resulted in easier access to human ST. It is now possible to select ST from many sites within large and small joints, even in the earliest phases of disease, enhancing studies of aetiology, prognosis and response to treatment [4].
Analysis of biomarkers in ST is increasingly being used in the evaluation of new targeted therapies in RA patients [5]. Numerous studies have suggested consistent associations between rapidity and magnitude of both clinical and immunohistological responses. It was shown that, especially within the ST, the number of infiltrating sublining macrophages can be used as a biomarker of clinical efficacy in relatively small studies of short duration [6,7]. Therefore, change in synovial sublining macrophages may be used as a biomarker for the evaluation of novel antirheumatic therapies. In addition to screening for possible efficacy, this approach provides insight into the mechanism of action of treatment.
Within this setting, reliable and validated methods for studying the ST are pivotal. The use of computerized or digital image analysis (DIA) has greatly facilitated the evaluation of ST. The major advantage of DIA is standardization of image acquisition and processing, minimizing variance, and the ability to quantify the actual stained area together with staining intensity in a time efficient manner [8,9]. This allows analysis of large numbers of stained sections. Strong correlations were observed between CIA, semiquantitative scoring and manual counting for analysis of ST cellular markers, cytokines and adhesion molecules [10,11]. Although the reproducibility of measuring cytokine and cell adhesion molecule staining by DIA was reported to be within 10% [8], no formal studies investigating intrarater and inter-rater variability have yet been reported. Therefore, we designed a study to determine the intrarater and inter-rater reliability of this approach for the analysis of synovial biopsies from RA patients participating in clinical trials.
Materials and methods
Patients and samples
Arthroscopic synovial biopsies were obtained before and 2 weeks after treatment in 19 patients who participated in a double-blind, placebo-controlled, single-centre study with prednisolone, as reported earlier [6]. All patients included had RA according to the 1987 criteria proposed by the American College of Rheumatology [12] and were on a stable regimen of disease-modifying antirheumatic drugs (methotrexate, sulphasalazine, hydroxychloroquine or leflunomide, or a combination of these) for at least 28 days before inclusion in the study. Ten out of the 19 patients received prednisolone and nine received placebo treatment. Needle arthroscopy of an actively inflamed joint (knee, ankle, or wrist) was performed under local anaesthesia in all patients before treatment and in the same joint after treatment. The procedures for needle arthroscopy were performed as described previously in detail [13,14]. During each procedure, biopsies were taken from six or more sites throughout the joint to minimize sampling error [15,16]. These specimens were directly collected en bloc in a mold embedded in Tissue Tek OCT (Miles diagnostics, Elkhart, IN, USA) and subsequently snap frozen by immersion in methylbutane (-80°C). The frozen blocs were stored in liquid nitrogen until they were processed. The study was approved by the Medical Ethics Committee of the Academic Medical Center, Amsterdam, The Netherlands, and all patients provided informed consent before start of the study.
Immunohistochemical analysis
From each tissue sample, consisting of six different biopsy samples, serial sections were cut with a cryostat (5 μm) and stained with the following antibodies to analyze the major cell populations in the synovium: anti-CD68 (EMB11; Dako, Glostrup, Denmark), anti-CD38 (HB-7; Becton Dickinson) and anti-CD3 (SK7; Becton Dickinson, Erembodegem, Belgium). Sections with nonassessable tissue, defined as the absence of an intimal lining layer, were not analyzed. For control sections, the primary antibodies were omitted or irrelevant antibodies were applied. Staining for cellular markers was performed using a three-step immunoperoxidase method, as was previously described [17].
Digital image analysis
After immunohistochemical staining, all coded sections were randomly analyzed by computer-assisted image analysis (Fig. 1). For all markers, 18 high-power fields were analyzed. The images of the high-power fields were analyzed using the Qwin analysis system (Leica, Cambridge, UK), as described previously in detail [10,11].
For determination of intrarater reliability, one observer performed the acquisition and analysis twice with an interval of 4 weeks in between (OB1 t0 and OB1 t1, respectively). To determine the inter-rater reliability, acquisition of images and analysis were performed independently by two other experienced observers (OB2 and OB3). All observers were blinded regarding clinical data. For each measurement all observers independently set their own threshold levels regarding the detection of stained antigen, nuclear staining and background staining. After the analysis, all observers independently calculated the mean change in the total number of positive cells per square millimetre of ST for each marker.
Statistical analysis
The nonparametric Friedman test and the Wilcoxon signed rank test were used to identify differences in the detection of the change in positive cell numbers per marker in the whole patient group, between observers and between assessments. The intrarater and inter-rater reliability was quantified by means of the intraclass correlation coefficient (ICC) of agreement [18]. In addition, scatter plots, in accordance with methods reported by Bland and Altman [19], were constructed to show differences in the change in positive cells between two measurements from one observer. The smallest detectable changes (SDCs), representing the smallest change in scores that can be deemed to be a 'real' change [20], for the intra-observer variances was calculated and used to evaluate their disciminatory power. The nonparametric Mann–Whitney U-test was used to determine whether each analysis detected differences in the change of positive cell numbers when the placebo group was compared with the prednisolone-treated group.
Results
Intrarater reliability
The mean numbers of CD3+ T lymphocytes, CD38+ plasma cells and CD68+ sublining macrophages before and after intervention for two analyses by the same observer at different time points (OB1 t0 and OB1 t1) are shown in Table 1. There were no significant differences in the mean change in T cells, plasma cells and macrophages in the total population between the two measurements.
The overall correlations between the first and second analysis by the same observer were good. For the measurement of the change in CD3+ T lymphocytes, CD38+ plasma cells and CD68+ macrophages, the single rater and average of rater ICCs were calculated and are shown in Table 2. The relations between the two measurements by the single observer are plotted in Fig. 2. There were no systemic differences between the two measurements for each marker, but the variation was rather large. An analysis of the between patient variances and within patient variances is provided in Table 2.
The SDC, averaged for the number of readings, for CD3+ lymphocytes was 182, for CD38+ plasma cells it was 128, and for CD68+ macrophages it was 306. When these estimates were used to identify those patients who responded to the treatment (i.e. had a reduction in positive cell numbers exceeding the SDC), for CD3+ lymphocytes four of the 10 patients in the prednisolone group responded versus none of the nine patients in the placebo group; for CD38+ plasma cells four of the 10 patients in the prednisolone group responded versus one of the nine patients in the placebo group; and for CD68+ macrophages seven out of the 10 patients in the prednisolone group responded versus none of the nine patients in the placebo group.
To determine whether the same observer identified the same differences in the synovial infiltrate after treatment at different time points, we determined whether there were significant differences in the change in T cells, plasma cells and macrophages between the placebo group and the prednisolone-treated group for each measurement. At both time points there was, on average, a significant reduction in the number of CD3+ lymphocytes and CD68+ macrophages in the prednisolone-treated patients as compared with placebo (Table 1), whereas on average there were no significant changes in the number of CD38+ plasma cells.
Interrater reliability
The mean number of T cells, plasma cells and macrophages before and after intervention measured by the other two observers (OB2 and OB3) are also shown in Table 1. There were no statistically significant differences in the mean change in positive cells between the analyses by the three observers (OB1 t0, OB2 and OB3).
When the overall correlations between the analyses of the three observers were calculated the ICCs (single and average of raters) appeared to be good for CD3+ lymphocytes, CD38+ plamsa cells and CD68+ macrophages (Table 2). An analysis of between patient variances and the within patient variances is also provided in Table 2.
To determine whether all three observers identified the same differences in the synovial infiltrate after treatment, we determined whether there were significant differences in the change in T cells, plasma cells and macrophages between the placebo group and the prednisolone-treated group for each measurement. The measurements by all three observers showed, on average, a significant reduction in the number of CD3+ lymphocytes and CD68+ macrophages in the prednisolone-treated patients versus placebo (Table 1), whereas, on average, there were no significant changes in the number of CD38+ plasma cells.
Discussion
This study investigated the intra- and interobserver reliability of assessment of the change in ST T cells, plasma cells, and macrophages quantified by DIA. Tissue samples were obtained from RA patients participating in a single-centre, placebo-controlled clinical trial with prednisolone. There were no significant differences in measurement of the mean change in T cells, plasma cells and macrophages between the three observers, or for different measurements by one observer. ICCs revealed good agreement between measurements. All observers and all measurements identified, on average, significant reductions in T cells and macrophages but not in plasma cells in the prednisolone group compared with placebo.
It can be anticipated that there will be an upsurge in randomized controlled trials investigating novel biological agents and small molecules in terms of their safety and efficacy. Thus, sensitive, validated and reliable measurements to screen for potential efficacy in an early phase of drug development are clearly needed. Clinical outcome measures have historically been used as primary end-points, but their reliability may be limited in small proof-of-principle studies. For clinical measurements such as the tender and swollen joint count, ICCs have been reported to vary between 0.15 and 0.85 for inter-rater variability and between 0.67 and 0.95 for intrarater variability [21]. Radiographic measurements, with the use of conventional X-ray films, show good reliability in most studies but they are not useful in short-term clinical trials [21]. The use of magnetic resonance images is promising, with acceptable inter-rater ICC for global synovitis scores and bone erosions, although optimal scoring systems are yet to be developed [22].
In light of the need to screen various compounds for potential efficacy in small numbers of patients and because of recent technical developments, we believe that our thinking about clinical trials is about to change dramatically. Clinical studies conducted during early phases of drug development will increasingly consist of small trials with a high density of biological data [23]. Consistent with this notion, serial ST analysis with evaluation of biomarkers was recently included in several randomized clinical trials of both disease-modifying anti-rheumatic drugs and biological agents [6,13,24-27]. These and other studies showed consistent relationships between the magnitude of synovial changes and clinical response. In particular, the change in infiltrating sublining macrophages was identified to be a potent and sensitive synovial biomarker [6,7].
ST can easily and safely be obtained as a result of the introduction of small-bore arthroscopes and the development of local and regional anaesthesia protocols. Despite heterogeneity in the ST within a single joint, it has been shown that representative measures of synovial inflammation can be obtained by examining a limited area of tissue [15,28,29]. Previous work [10,11] has also shown that DIA is a sensitive, time efficient method for quantifying both the number of stained cells and the staining intensity, with good correlations with both manual counting and semiquantative scoring.
Although DIA is described as reliable and objective, little is known about the variability and reliability of this tool. Variation in measurements may result from a limited number of factors with this approach. In our system the observer selects three different areas of each six high-power fields from one slide, which is composed of six biopsy samples from six different sites in the joint. This is done in such a way that a representative area is selected, and this requires extensive training and experience with the histopathological morphology of ST. After scanning the representative high-power fields, the images are analyzed by setting threshold values for the stained antigen, nuclear staining and background staining [10]. These thresholds are kept constant for all measurements with the same marker within a study, but could theoretically give rise to variation when set by different observers or by one observer at different times. In the present study it was shown that these variables did not result in different outcomes. There were good ICCs when the findings of three experienced observers or the findings of the same observer at different times were compared. Analysis by Bland–Altman plots showed no systemic differences with regard to the intra-observer measurements, and the SDCs showed good discriminatory power when applied to the treatment groups. In addition, all observers and all measurements identified the same cell types (T cells and macrophages) as decreasing significantly in the active treatment group compared with placebo. All measurements also identified a consistent trend toward reduced plasma cell numbers after corticosteroid treatment, which did not reach statistical significance, possibly because of the relative small number of patients included. Although this method does exhibit good agreement in detecting changes in histological markers, this does not necessarily mean that these results can be extrapolated to the expression of a given marker at a given time point, as used in cross-sectional studies of ST. In addition, it remains to be seen whether the same reliability holds true for determination of changes in secreted proteins, such as cytokines and chemokines.
Conclusion
In conclusion, the findings of the present study show the reliability of ST analysis using a DIA system for the evaluation of serial synovial biopsy samples before and after treatment. This approach may be used for efficient quantification of synovial biomarkers in small proof-of-principle clinical trials.
Competing Interests
The author(s) declare that they have no competing interests.
Abbreviations
DIA = digital image analysis; ICC = intraclass correlation coefficient; RA = rheumatoid arthritis; SDC = smallest detectable change; ST = synovial tissue.
Authors' contributions
JJH contributed to experiments, was responsible for data analysis and interpretation, and wrote the manuscript. MV and TJMS were responsible for both the set-up and performance of the experiments. DMG was responsible for including patients and collecting materials and data. AHZ coordinated and assisted in the statistical analysis of the data. PPT was responsible for planning the work and contributed to data analysis, interpretation and write up.
Acknowledgements
This study was supported by a grant from Zon-Mw (The Netherlands Organisation for Health Research and Development), grant number 902-37-123.
Figures and Tables
Figure 1 Acquisition, analysis and output for a digital image analysis system. Acquisition and analysis of immunohistochemical staining of CD3+ T lymphocytes in synovial tissue using a digital image analysis system [10]. Three different areas from each biopsy sample, which are representative of the whole tissue section are selected. During analysis, staining thresholds are set for primary staining (i.e. CD3+ T lymphocytes), nuclear staining and background staining. The output is generated in a spreadsheet as the total number of positive cells per square millimetre of synovial tissue.
Figure 2 Mean change in number of positive cells versus the difference in change in positive cells. Shown are scatter plots of the mean change in the number of positive cells versus the difference in change of positive cells between two measurements by observer 1 for (a) CD3+ lymphocytes, (b) CD38+ plasma cells and (c) CD68+ macrophages. The dotted line represents the mean ± 2 × standard deviation.
Table 1 Numbers of positive cells before and after intervention
Cell type Treatment Observer 1 t0 Observer 1 t1 Intra-observer comparison Observer 2 Observer 3 Inter- observer comparison
Placebo Prednisolone Pa Placebo Prednisolone Pa Pb Placebo Prednisolone Pa Placebo Prednisolone Pa Pc
CD3+ T lymphocytes Before 192 ± 246 358 ± 413 <0.05 110 ± 140 227 ± 225 <0.05 NS 179 ± 185 285 ± 250 <0.05 89 ± 110 112 ± 98 <0.05 NS
After 387 ± 391 140 ± 150 299 ± 356 44 ± 62 470 ± 618 101 ± 94 196 ± 224 54 ± 66
CD38+ plasma cells Before 56 ± 87 99 ± 130 NS 73 ± 134 116 ± 166 NS NS 246 ± 307 397 ± 498 NS 145 ± 151 309 ± 380 NS NS
After 96 ± 127 37 ± 57 119 ± 149 42 ± 78 335 ± 411 132 ± 187 315 ± 416 83 ± 121
CD68+ macrophages Before 804 ± 422 973 ± 419 <0.03 441 ± 422 572 ± 404 <0.03 NS 937 ± 292 1151 ± 254 <0.03 621 ± 445 724 ± 360 <0.03 NS
After 972 ± 151 553 ± 342 632 ± 686 222 ± 278 984 ± 354 796 ± 306 720 ± 527 313 ± 291
Shown are the mean numbers (± standard deviation) of CD3+ T lymphocytes, CD38+ plasma cells and CD68+ sublining macrophages per square millimetre of synovial tissue before and after intervention, measured by one observer at two different time points (OB1 t0 and OB1 t1) and by two other observers (OB2 and OB3) for placebo-treated patients and prednisolone-treated patients. aNonparametric, unpaired, Mann–Whitney U-test for the comparison between placebo and prednisolone treatment. bNonparametric, paired, Wilcoxon signed rank test, for the comparison between OB1 t0 and OB1 t1 (intra-observer comparison). cNonparametric, paired, Friedman test, for the comparsion between the three observers (OB1 t0, OB2 and OB3).
Table 2 Estimates of the variance components (between and within patients) and of the intraclass correlations (single rater and average of raters)
Cell type Intra-observer Inter-observer
Between patients Within patients ICC ICC of the mean of two observations Between patients Within patients ICC ICC of the mean of three observations
CD3+ cells 11.59 1.73 0.87 0.93 10.13 4.85 0.68 0.86
CD38+ cells 1.35 0.82 0.62 0.77 8.65 3.83 0.69 0.87
CD68+ cells 20.32 7.35 0.73 0.85 18.92 7.19 0.72 0.89
ICC, intraclass correlation coefficient.
==== Refs
O'Dell JR Therapeutic strategies for rheumatoid arthritis N Engl J Med 2004 350 2591 2602 15201416 10.1056/NEJMra040226
Firestein GS Evolving concepts of rheumatoid arthritis Nature 2003 423 356 361 12748655 10.1038/nature01661
Tak PP Bresnihan B The pathogenesis and prevention of joint damage in rheumatoid arthritis: advances from synovial biopsy and tissue analysis Arthritis Rheum 2000 43 2619 2633 11145019 10.1002/1529-0131(200012)43:12<2619::AID-ANR1>3.0.CO;2-V
Bresnihan B Tak PP Synovial tissue analysis in rheumatoid arthritis Baillieres Best Pract Res Clin Rheumatol 1999 13 645 659 10652645 10.1053/berh.1999.0051
Tak PP Lessons learnt from the synovial tissue response to anti-rheumatic treatment Rheumatology (Oxford) 2000 39 817 820 10952734
Gerlag DM Haringman JJ Smeets TJ Zwinderman AH Kraan MC Laud PJ Morgan S Nash AF Tak PP Effects of oral prednisolone on biomarkers in synovial tissue and clinical improvement in rheumatoid arthritis Arthritis Rheum 2004 50 3783 3791 15593225 10.1002/art.20664
Haringman JJ Gerlag DM Zwinderman AH Smeets TJ Kraan MC Baeten D McInnes IB Bresnihan B Tak PP Synovial tissue macrophages: highly sensitive biomarkers for response to treatment in rheumatoid arthritis patients Ann Rheum Dis 2005 64 834 838 15576415
Youssef PP Triantafillou S Parker A Coleman M Roberts-Thomson PJ Ahern MJ Smith MD Variability in cytokine and cell adhesion molecule staining in arthroscopic synovial biopsies: quantification using color video image analysis J Rheumatol 1997 24 2291 2298 9415630
Norazmi MN Hohmann AW Jarvis LR Skinner JM Stoll P Bradley J The use of computer-assisted video image analysis in the enumeration of immuno-stained cells in tissue sections J Immunol Methods 1990 131 223 227 1697316 10.1016/0022-1759(90)90193-Y
Kraan MC Haringman JJ Ahern MJ Breedveld FC Smith MD Tak PP Quantification of the cell infiltrate in synovial tissue by digital image analysis Rheumatology (Oxford) 2000 39 43 49 10662872
Kraan MC Smith MD Weedon H Ahern MJ Breedveld FC Tak PP Measurement of cytokine and adhesion molecule expression in synovial tissue by digital image analysis Ann Rheum Dis 2001 60 296 298 11171698 10.1136/ard.60.3.296
Arnett FC Edworthy SM Bloch DA McShane DJ Fries JF Cooper NS Healey LA Kaplan SR Liang MH Luthra HS The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis Arthritis Rheum 1988 31 315 324 3358796
Kraan MC Reece RJ Barg EC Smeets TJ Farnell J Rosenburg R Veale DJ Breedveld FC Emery P Tak PP Modulation of inflammation and metalloproteinase expression in synovial tissue by leflunomide and methotrexate in patients with active rheumatoid arthritis. Findings in a prospective, randomized, double-blind, parallel-design clinical trial in thirty-nine patients at two centers Arthritis Rheum 2000 43 1820 1830 10943872 10.1002/1529-0131(200008)43:8<1820::AID-ANR18>3.0.CO;2-D
Kraan MC Reece RJ Smeets TJ Veale DJ Emery P Tak PP Comparison of synovial tissues from the knee joints and the small joints of rheumatoid arthritis patients: Implications for pathogenesis and evaluation of treatment Arthritis Rheum 2002 46 2034 2038 12209505 10.1002/art.10556
Dolhain RJ Ter Haar NT De Kuiper R Nieuwenhuis IG Zwinderman AH Breedveld FC Miltenburg AM Distribution of T cells and signs of T-cell activation in the rheumatoid joint: implications for semiquantitative comparative histology Br J Rheumatol 1998 37 324 330 9566675 10.1093/rheumatology/37.3.324
Boyle DL Rosengren S Bugbee W Kavanaugh A Firestein GS Quantitative biomarker analysis of synovial gene expression by real-time PCR Arthritis Res Ther 2003 5 R352 R360 14680510 10.1186/ar1004
Tak PP van der Lubbe PA Cauli A Daha MR Smeets TJ Kluin PM Meinders AE Yanni G Panayi GS Breedveld FC Reduction of synovial inflammation after anti-CD4 monoclonal antibody treatment in early rheumatoid arthritis Arthritis Rheum 1995 38 1457 1465 7575695
Deyo RA Diehr P Patrick DL Reproducibility and responsiveness of health status measures. Statistics and strategies for evaluation Control Clin Trials 1991 12 142S 158S 1663851
Bland JM Altman DG Statistical methods for assessing agreement between two methods of clinical measurement Lancet 1986 1 307 310 2868172
Bruynesteyn K Boers M Kostense P van der LS van der HD Deciding on progression of joint damage in paired films of individual patients: smallest detectable difference or change Ann Rheum Dis 2005 64 179 182 15286006 10.1136/ard.2003.018457
Lassere MN van der HD Johnson KR Boers M Edmonds J Reliability of measures of disease activity and disease damage in rheumatoid arthritis: implications for smallest detectable difference, minimal clinically important difference, and analysis of treatment effects in randomized controlled trials J Rheumatol 2001 28 892 903 11327273
Lassere M McQueen F Ostergaard M Conaghan P Shnier R Peterfy C Klarlund M Bird P O'Connor P Stewart N OMERACT Rheumatoid Arthritis Magnetic Resonance Imaging Studies. Exercise 3: an international multicenter reliability study using the RA-MRI Score J Rheumatol 2003 30 1366 1375 12784419
Liu ET Karuturi KR Microarrays and clinical investigations N Engl J Med 2004 350 1595 1597 15084689 10.1056/NEJMp048050
Cunnane G Madigan A Murphy E Fitzgerald O Bresnihan B The effects of treatment with interleukin-1 receptor antagonist on the inflamed synovial membrane in rheumatoid arthritis Rheumatology (Oxford) 2001 40 62 69 11157143
Smeets TJ Kraan MC van Loon ME Tak PP Tumor necrosis factor alpha blockade reduces the synovial cell infiltrate early after initiation of treatment, but apparently not by induction of apoptosis in synovial tissue Arthritis Rheum 2003 48 2155 2162 12905468 10.1002/art.11098
Haringman JJ Kraan MC Smeets TJ Zwinderman KH Tak PP Chemokine blockade and chronic inflammatory disease: proof of concept in patients with rheumatoid arthritis Ann Rheum Dis 2003 62 715 721 12860725 10.1136/ard.62.8.715
Katrib A Smith MD Ahern MJ Slavotinek J Stafford L Cuello C Bertouch JV McNeil HP Youssef PP Reduced chemokine and matrix metalloproteinase expression in patients with rheumatoid arthritis achieving remission J Rheumatol 2003 30 10 21 12508384
Rooney M Condell D Quinlan W Daly L Whelan A Feighery C Bresnihan B Analysis of the histologic variation of synovitis in rheumatoid arthritis Arthritis Rheum 1988 31 956 963 2457377
Bresnihan B Cunnane G Youssef P Yanni G Fitzgerald O Mulherin D Microscopic measurement of synovial membrane inflammation in rheumatoid arthritis: proposals for the evaluation of tissue samples by quantitative analysis Br J Rheumatol 1998 37 636 642 9667617 10.1093/rheumatology/37.6.636
| 15987488 | PMC1175038 | CC BY | 2021-01-04 16:02:37 | no | Arthritis Res Ther. 2005 May 12; 7(4):R862-R867 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1757 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17581598749010.1186/ar1758Research ArticlePerforin deficiency attenuates collagen-induced arthritis Bauer Kristin [email protected] Annika [email protected] Hoang [email protected] Dirk [email protected] Hans-Jürgen [email protected] Horst [email protected] Eilhard [email protected] Hans-Juergen [email protected] Rikard [email protected] Saleh M [email protected] Institute of Immunology, University of Rostock, Rostock, Germany2 Institute of Pathology, University of Rostock, Rostock, Germany3 Institute of Neurology, University of Rostock, Rostock, Germany4 Section for Medical Inflammation Research, Lund University, Lund, Sweden2005 20 5 2005 7 4 R877 R884 24 9 2004 14 10 2004 22 3 2005 15 4 2005 Copyright © 2005 Bauer et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Collagen-induced arthritis (CIA), an approved animal model for rheumatoid arthritis, is thought to be a T cell-dependent disease. There is evidence that CD8+ T cells are a major subset controlling the pathogenesis of CIA. They probably contribute to certain features of disease, namely tissue destruction and synovial hyperplasia. In this study we examined the role of perforin (pfp), a key molecule of the cytotoxic death pathway that is expressed mainly in CD8+ T cells, for the pathogenesis of CIA. We generated DBA/1J mice suffering from mutations of the pfp molecule, DBA/1J-pfp-/-, and studied their susceptibility to arthritis. As a result, pfp-deficient mice showed a reduced incidence (DBA/1J-pfp+/+, 64%; DBA/1J-pfp-/-, 54%), a slightly delayed onset (onset of disease: DBA/1J-pfp+/+, 53 ± 3.6; DBA/1J-pfp-/-, 59 ± 4.9 (mean ± SEM), and milder form of the disease (maximum disease score: DBA/1J-pfp+/+, 7.3 ± 1.1; DBA/1J-pfp-/-, 3.4 ± 1.4 (mean ± SEM); P < 0.05). Concomitantly, peripheral T cell proliferation in response to the specific antigen bovine collagen II was increased in pfp-/- mice compared with pfp+/+ mice, arguing for an impaired killing of autoreactive T cells caused by pfp deficiency. Thus, pfp-mediated cytotoxicity is involved in the initiation of tissue damage in arthritis, but pfp-independent cytotoxic death pathways might also contribute to CIA.
==== Body
Introduction
Collagen-induced arthritis (CIA) is an experimental model of arthritis inducible in susceptible strains of mice, for example DBA1/J, by immunization with bovine collagen type II (CII) in complete Freund's adjuvant (CFA) [1-4]. The development of CIA is known to depend on T cells, and disease susceptibility is linked to the major histocompatibility complex (MHC) region [5]. After T cell activation an inflammatory cascade involving T cells, macrophages/monocytes, B cells, and activated synoviocytes, is triggered. Different types of leucocytes and synovial cells produce a complex array of cytokines and other soluble mediators, which are thought to be responsible for cartilage destruction and bone erosion [6-9]. Some of the main features of disease are synovial hyperplasia and mononuclear cell infiltration. Factors contributing to this phenomenon are unknown; however, an imbalance between rates of cell proliferation and cell death (apoptosis) has been suggested by recent studies of rheumatoid synovium demonstrating that apoptosis of synovial cells and infiltrating lymphocytes was common in situ [10,11]. In the immune system, apoptosis is involved in development and in the negative selection of lymphocytes. It is also crucial in downregulating immune responses to foreign antigens. Cytotoxic T lymphocytes and other killer cells can eliminate their targets by the induction of cell death. All of these functions are primarily mediated through the receptors Fas/APO-1, tumor necrosis factor receptor 1 (TNFR1; p55 TNFR) and the perforin (pfp)/granzyme pathway [12,13].
Perforin is expressed mainly in activated cytotoxic T lymphocytes (CTLs) and natural killer (NK) cells, although some reports suggest its expression in microglia as well [14]. In CTLs, pfp is stored in cytoplasmic granules and is a major effector of cytolysis by these cells. On pfp release it inserts itself into the plasma membrane of target cells and polymerizes into pore forming aggregates. Pores of pfp lead to osmotic lysis of target cells and induce apoptosis by allowing granzymes to enter the target cells. Perforin-deficient mice have confirmed its function as an effector molecule and in the immune response to viruses and tumors as well as in other aspects of immune regulation such as activation-induced cell death (AICD), antibody production and spontaneous autoimmunity [15-17]. NOD mice, an animal model of insulin-dependent diabetes mellitus with a mutation of the pfp gene (NOD/pfp-mice), develop diabetes with highly reduced incidence and markedly delayed onset, pointing to a role of the pfp death pathway in tissue damage in this disease [18].
The role of pfp in arthritis is not clear, although some observations suggest a role in disease pathogenesis, for example pfp-expressing CTLs has been demonstrated in the rheumatoid synovium, and CD8-deficient mice seem to be less susceptible to induction of collagen-induced arthritis [19].
It is conceivable that the pfp/granzyme pathway could contribute to the pathology of arthritis in at least two ways: promotion of autoimmunity by blocking peripheral tolerance and AICD or destruction of target tissues. In this study we attempted to evaluate the role of the pfp-mediated death pathway in the pathogenesis of CIA with the use of pfp-deficient mice (pfp-/-) by examining the effect of the mutation on the clinical course of disease, immune response to collagen and on joint pathology.
Materials and methods
Mice
Perforin-deficient mice were not available on the DBA/1J background. We obtained pfp-/- C57BL/6J (B6) mice, generated as described previously [15], from the Jackson Laboratories (Bar Harbor, ME, USA). These mice were backcrossed onto the CIA-susceptible DBA/1J background (Harlan-Winkelmann, Borchen, Germany) for at least 14 generations. The mice were propagated as hemizygous mutants and the mutations were followed by PCR analysis of tail DNA. To produce homozygous pfp-/- mice for the experiments, heterozygous pfp-deficient mice were intercrossed. Successful backcrossing to the DBA/1J background was assessed by PCR analysis of the MHC-H2 locus [20]. To select mice heterozygous for the pfp-deficiency the primers 5'-TTT TTG AGA CCC TGT AGA CCC A-3' (pfp1) and 5'-GCA TCG CCT TCT ATC GCC TTC T-3' (pfp2) were used. For selection of homozygous pfp-deficient mice, pfp3 primer (5'-CCG GTC CTG AAC TCC TGG CCA A-3') was used in combination with pfp4 primer (5'-CCC CTG CAC ACA TTA CTG GAA G-3'). Microsatellite markers (Metabion GmbH, Planegg-Martinsried, Germany) surrounding the pfp gene were used for determining the size of the C57BL/6J DNA fragment in the mutant mice. The amplified microsatellites were separated and analyzed on denaturing polyacrylamide gels, and were detected with a LI-COR Model 4200L automated DNA sequencer (LI-COR Inc., Lincoln, NE, USA). For the experiments male mice 8 to 16 weeks old were used. Animals were kept and bred under standard conditions at the facility of the University of Rostock. All experiments were approved by the appropriate authorities in the state of Mecklenburg-Vorpommern, Germany.
Induction and clinical evaluation of collagen-induced arthritis
Age-matched mice were immunized intradermally at the base of the tail with 125 μg of bovine CII (Chondrex, Redmond, WA, USA) emulsified in CFA (incomplete Freund's adjuvant containing 4 mg/ml Mycobacterium tuberculosis; Difco Laboratories, Detroit, IL, USA) or with CFA only. Mice were then boosted with 125 μg of bovine CII in incomplete Freund's adjuvant at day 21. Blood were taken at day 0 and day 21 before boosting, and serum was collected. Clinical scores were assessed immediately before immunization (day 0) and thereafter three times weekly. Inflammation of the four paws was scored as follows: 0, no inflammation; 1, swelling or redness of one joint; 2, swelling or redness of more than one joint or mild inflammation of the whole paw; 3, severe inflammation of whole paw or ankylosis. CFA-immunized mice served as controls.
T cell proliferation response
Popliteal, preperioteneal, inguinal, mesenterial, axillary and cervical lymph nodes were removed under aseptic conditions. Single-cell suspensions of mononuclear cells of pooled lymph nodes from individual mice were prepared. The cells were cultured in triplicates in flat-bottomed 96-well culture plates, at a concentration of 2 × 106 cells in 200 μl of medium (RPMI 1640 with Glutamax-II supplemented with 50 IU/ml penicillin, 60 μg/ml streptomycin (Gibco, Karlsruhe, Germany) and 5% heat-inactivated fetal calf serum). To investigate the antigen-specific response, lymphocytes were stimulated with 10, 1 or 0.1 μg/ml bovine CII; 4 μg/ml concanavalin A (Difco) was used as positive control, and medium only was used as negative control. Cells were incubated for 72 hours at 37°C in a humidified atmosphere containing 5% CO2. To measure the proliferation by DNA synthesis, cells were pulsed with 1 μCi of [3H]thymidine for the last 12 hours of culture. Cells were harvested onto glass fiber filters, and [3H]thymidine incorporation was measured in a liquid β-scintillation counter. The results were expressed as counts per minute.
Cytokine ELISA
To analyze cytokine production, cells were cultured as described above; supernatant was collected after 72 hours in vitro antigen challenge. Concentrations of IFN-γ in the supernatants were determined by the Cytoscreen Immunoassay Kit (BioSource, Camarillo, CA, USA) in accordance with the instructions of the manufacturer. In brief, the lymphocyte supernatant was added to an ELISA-plate coated with a monoclonal antibody specific for mouse IFN-γ. After incubation and washing, a biotinylated polyclonal antibody specific for mouse IFN-γ was added. Then streptavidin–peroxidase and later the substrate solution were added to detect the products of the reaction.
Anti-CII antibody assay
The serum from the mice was analyzed with ELISA for the quantification of IgG antibodies against CII. Micro-ELISA plates were coated overnight at 4°C with 50 μl of PBS containing 5 μg/ml bovine CII in each well. After washing, the sera were added and incubated at 37°C for 2 hours. After 1 hour of incubation at 37°C with anti-mouse-IgG conjugated with alkaline phosphatase (Pharmingen BD), p-nitrophenylphosphate containing substrate buffer (Sigma) was added, and 3 M NaOH was used to stop the reaction. The plates were read at 405 nm.
Histopathology
Mice were killed and paws were cut off and subsequently fixed in 4% paraformaldehyde solution. After decalcification for 2 to 3 weeks in an EDTA solution the paws were embedded in paraffin. The paws were sectioned and stained with H & E. Evaluation of disease was made according to a previously published scale [21]: 1, synovial hyperplasia; 2, start of pannus development; 3, erosions of bone and cartilage; 4, severe inflammation and erosions.
Fluorescence-activated cell sorting analysis
For the determination of T cell, B cell and NK cell populations in the different pfp-mice, lymphocytes were isolated and stained with fluorescein isothiocyanate (FITC)-labeled anti-CD4 antibody (Pharmingen; clone H129.19), phycoerythrin-labeled anti-CD8a antibody (Pharmingen; clone 43-6.7), FITC-labeled anti-CD45R/B220 antibody (Pharmingen; clone RA3-6B2), phycoerythrin-labeled anti-CD90 antibody (Pharmingen; clone OX-7), and anti-IgG-biotin and streptavidin-FITC before being analyzed by FACScan (Becton Dickinson; with Cell Quest software version 1.2.2).
Statistics
Statistical evaluation was performed with SPSS software. For analyzing differences in clinical scores, a Mann–Whitney test was used. For incidence calculations, χ2 and Fisher tests were used. When analyzing differences in cytokine production and T cell proliferation, Student's unpaired t-test was used. Differences were considered significant at P < 0.05.
Results
Characterization of the pfp-deficient DBA/1J mice
The pfp-deficient mice were originally derived from C57BL/6J strain by embryonic stem cell transfer. Because this strain is resistant to CIA induction, we backcrossed the mice for at least 14 generations into the CIA-susceptible DBA/1J background. The pfp gene is located on chromosome 10 at 36 cM, and the mutant gene was inherited as a C57BL/6J fragment of about 10 cM between about 28 and 39 cM (Table 1). To exclude the possibility that backcrossing with the DBA/1J strain revealed a defect in T cell, B cell or NK cell maturation we examined distribution of these different immune cell populations by fluorescence-activated cell sorting. It was found that T lymphocytes expressing CD4 and CD8, as well as CD90+ cells (CD90 is expressed on T cells and NK cells) and CD45R/B220-positive cells (B cells, activated killer cells) were present at comparable percentages in heterozygous and pfp-deficient DBA/1J mice, indicating that the lack of pfp did not affect development of these cell populations in the DBA/1J strain (data not shown). The lack of pfp expression in the activated T cells of the pfp-/- mice was confirmed at the RNA level by RT-PCR (see Additional file 1a).
Pfp-deficient mice are less susceptible to collagen-induced arthritis
To investigate the role of pfp in collagen-induced arthritis, we immunized DBA-pfp-/-, DBA-pfp+/- and DBA-pfp+/+ mice with bovine CII in CFA. As shown in Fig. 1 and Table 2, mice deficient for pfp developed a less severe disease with lower clinical scores than mice with intact pfp. DBA-pfp-/- mice also showed a tendency to a delayed onset of CIA. Mice with intact pfp (pfp+/+ and pfp+/-) had an average incidence of 64%, whereas pfp-/- mice showed a mean incidence of 54% (Fig. 1). The severity of disease was decreased on day 50 after immunization (pfp-/-, 0.9 ± 0.86; pfp+/-, 1.4 ± 0.67; pfp+/+, 1.3 ± 0.47 (means ± SEM)), and on day 75 (pfp-/-, 2.1 ± 1.7; pfp+/-, 3.8 ± 1; pfp+/+, 5.2 ± 1.2), and significantly decreased on day 82 (pfp-/-, 3.0 ± 1.5; pfp+/-, 4.0 ± 0.9; pfp+/+, 5.6 ± 0.9; P < 0.05). Maxscore, calculated as the mean of the maximum score value for each individual mouse of a given genotype, was also significantly decreased in pfp-deficient mice (pfp-/- 3.4 ± 1.45) in comparison with pfp+/+ mice (7.3 ± 1.14; Fig. 1). In addition, the area under the curve, as a measure of severity, onset and chronicity, was significantly lower in pfp-/- mice than in mice carrying pfp (pfp-/-, 31.9 ± 23; pfp+/-, 44.6 ± 12.3).
Histopathological features of CIA do not change with pfp deficiency
To investigate joint histopathology, paws were dissected from pfp-/- and pfp+/- mice with CIA, H & E-stained and evaluated blind for signs of arthritis. Results (Fig. 2) reveal that pfp-/- mice can develop a severe arthritis with typical signs of CIA, namely proliferation of synoviocytes, infiltration of inflammatory cells, pannus development, and erosions of bone and cartilage. Comparing both genotypes, no differences in histopathological development of the disease were visible (Fig. 2).
Lack of pfp does not affect the antibody response to collagen
To investigate the role of pfp on the humoral immune response to collagen, sera were obtained from pfp-/-, pfp+/- and pfp+/+ mice at days 0 and 21 after immunization, and the levels of CII-specific IgG antibodies were measured by quantitative ELISA. No significant variations were seen in the levels of anti-CII-specific IgG between the pfp-/- and pfp+/- or pfp+/+ littermates (Fig. 3). Pfp-deficient mice showed a mean titer of anti-CII antibodies of 233 ± 51.1, and the control group had a mean titer of anti-CII antibodies of 306 ± 50.8 at day 21 after immunization. These data show that even without pfp the mice develop a strong humoral immune response against CII, suggesting that the lack of pfp does not affect the anti-CII antibody response.
Pfp-/- mice mount an enhanced proliferative T cell response toward CII
To investigate whether the T cell response against CII was affected by pfp deficiency, draining the lymph nodes of immunized pfp-/- and pfp+/+ was investigated for the proliferative response toward bovine CII. Perforin-deficient mice showed an enhanced T cell proliferation compared with pfp wild-type mice. As shown in Fig. 3, after stimulation with 0.1 μg/ml CII, pfp-/- mice showed a significantly elevated T cell proliferation of 12,835 ± 2,750 cpm compared with 4,727 ± 1,268 cpm in pfp+/+ mice; at a CII concentration of 1 μg/ml the T cell proliferation was also significantly increased in pfp-deficient mice (32,209 ± 6,764 cpm) compared with pfp-wild-type mice (21,904 ± 2,626 cpm). The IFN-γ production of T cells from these mice was measured by ELISA, but no significant differences between pfp-/- and pfp+/+ mice were observed (data not shown).
Discussion
The primary aim of this study was to investigate whether the delayed onset and mild arthritis observed earlier in CD8-/- mice [19] is due to a lack of cytotoxic ability or a lack of some other aspects of T cell function such as the secretion of cytokines. Indeed, we found that pfp-deficient mice develop disease with reduced incidence, a slightly delayed onset, and significantly decreased severity, suggesting that CTL activity is important for the initiation and maintenance of arthritis in the CIA model. This is in agreement with earlier observations in the insulin-dependent diabetes mellitus (NOD) mouse model. In that model, as in CIA, pfp is a susceptible rather than a protective gene. This is in contrast to other disease models, for example experimental autoimmune encephalomyelitis (EAE), systemic lupus erythematosus (SLE) and autoimmune pancreatitis, in which pfp clearly protects against the disease [16-18,22]. The protective effects of pfp could be explained by its immune regulatory function, namely the killing of autoreactive B cells and functional self-lysis of activated T cells through AICD. Hence, its role in regulation of peripheral tolerance and its role in autoimmunity, which is distinct from, but overlaps, the role of the FasL/Fas and TNF/TNFR pathways, might be the reason for accelerated autoimmunity in pfp-deficient mice in the EAE and SLE models.
In the present study we obtained evidence that pfp can also act as a susceptibility gene rather than a protective gene. A lack of the molecule led to reduced severity throughout the disease, a slightly delayed onset, and a reduced incidence of CIA. However, we also observed a strong anti-collagen B cell response with similar anti-collagen-specific IgG levels in control and pfp-/- mice, and significantly increased T cell proliferation in response to collagen. This might have been due to a reduced killing and accumulation of autoreactive T cells after activation because of impaired AICD. These findings indicate that a diminished adaptive response to collagen was not responsible for the reduced arthritis.
Previous studies demonstrated a strong involvement of pfp in the control of CD8+ T cell homeostasis. Perforin deficiency resulted in enhanced CD8+ T cell expansion because of decreased killing of antigen-presenting cells and consequential prolonged stimulation by antigen [23,24]. Our results support these propositions: we observed a significantly increased T cell proliferation in pfp-deficient mice in comparison with controls. These data also suggest a role of pfp in the regulation of T cell homeostasis. Nevertheless, cytokines produced by cytotoxic T cells are also involved in the regulation of the T cell response. TNF-α, for example, can mediate AICD of CD8+ T cells through TNFR1 and TNFR2 [25].
The histopathology of CIA in wild-type pfp+/+ mice of the present study was similar to that seen in previous studies [26]. The data are therefore not shown again here. In the present study there were no significant differences between histopathological changes of pfp+/- and pfp-/- mice, most probably because of a severe disease in individual pfp-/- mice. This result argues for an additional involvement of pfp-independent death pathways in joint destruction. Indeed, pfp was upregulated in the inflamed arthritic joints of wild-type mice as shown by RT-PCR (Additional file 1b).
This is similar to our recent finding showing that the FasL/Fas pathway has a proinflammatory role in CIA and an activating function on fibroblasts in vitro [27]. Fas-deficient mice developed arthritis that was less severe, probably through a reduced IL-1R1/Toll-like receptor-4 signaling that might contribute to a decreased expression of other cytokines, chemokines and matrix metalloproteinases potentially regulated by this pathway [28]. Previous studies also reported a strong involvement of the TNF/TNFR pathway because CIA only developed with a reduced disease incidence, and the severity and neutralization of TNF led to the prevention of arthritis [29,30].
Taken together, the results show that in CIA the disease-promoting effect of pfp prevails. It is therefore tempting to speculate that pfp could contribute to arthritis in at least two ways. First, it could promote tissue damage by direct cytotoxic effects through CD8+ T cells and NK cells. Second, pfp might have some activating functions on fibroblasts or macrophages, leading to the production of proinflammatory cytokines. Indeed, there are reports indicating that fibroblasts and monocytes can be activated by granzyme A to secrete the proinflammatory cytokines IL-6, IL-8 and TNF-α, which could subsequently severely regulate the inflammatory response [31,32].
There are other indications that argue for a prominent role of pfp in arthritis. Perforin was found to be differentially expressed in lymph nodes and joints of DBA/1J and FVB/N (CIA-resistant strain) mice (SM Ibrahim and D Koczan, unpublished observations) and pfp-expressing cytotoxic T lymphocytes and increased apoptosis were observed in the synovia of patients with rheumatoid arthritis [10,33,34]. The targeted pfp gene is likely to have a major role in the observed effects, on the basis of the absence of pfp production, although it is possible that other polymorphic genes in the linked region could also contribute to CIA reduction. Indeed, the pfp gene is mapped to a CIA-susceptible locus, Cia8 on chromosome 10. This quantitative trait locus is covered by the 12 cM C57BL/6J fragment including the mutant pfp gene. The Cia8 region contains several candidate genes that have been implicated in the modulation of CIA susceptibility, for example the macrophage migration inhibitory factor Mif [35] and the autoimmune regulator Aire [36]. However, the observation that heterozygous littermates and pfp+/+ mice that were also backcrossed and contained the same or smaller C57BL/6J fragments do not show effects on the disease argues against a major role for another gene.
In summary, the CIA in pfp-deficient mice was mild and showed delayed onset and reduced incidence, but some individual pfp-/- mice also developed a severe disease. These results suggest that pfp-dependent cytotoxicity is involved in the initiation of tissue damage in arthritis, but that pfp-independent cytotoxic death pathways, for example the FasL/Fas pathway, might also contribute to CIA.
Conclusion
We report that arthritis developed only with reduced incidence, severity and delayed onset in pfp-deficient DBA/1J mice. These findings suggest that pfp-dependent cytotoxicity is involved in the initiation of tissue damage in arthritis but also that one or several other pfp-independent mechanisms, possibly involving FasL/Fas, contribute to the early phase of joint destruction in CIA.
Abbreviations
AICD = activation-induced cell death; CFA = complete Freund's adjuvant; CIA = collagen-induced arthritis; CII = collagen type II; CTL = cytotoxic T lymphocyte; EAE = experimental autoimmune encephalomyelitis; ELISA = enzyme-linked immunosorbent assay; FITC = fluorescein isothiocyanate; H & E = hematoxylin and eosin; IFN = interferon; IL = interleukin; MHC = major histocompatibility complex; NK = natural killer; RT-PCR = reverse transcriptase polymerase chain reaction; pfp = perforin; SLE = systemic lupus erythematosus; TNF = tumor necrosis factor.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
KB, AK and HTR did the experimental work and contributed to writing the manuscript. KB and AK contributed equally to the work. DK, HJT, RH and SMI participated in the design and co-ordination of the study and drafted the manuscript. HJK and HN did the histopathology. EM carried out the T cell proliferation and cytokine assays. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
A TIFF file showing the perforin mRNA expression level in CD8+ T cells of perforin-deficient, heterozygous and wild-type mice as well as the perforin mRNA expression level of healthy and inflamed joints from wild-type mice.
Click here for file
Acknowledgements
We thank I Klamfuss, E Lorbeer and R Waterstradt for excellent technical assistance. This work was supported by grants from the DFG (DFG 243/1) and EU (EUROME) QL.
Figures and Tables
Figure 1 Perforin-deficient (pfp-/-) mice develop decreased collagen-induced arthritis (CIA). Percentage of CIA-affected animals (a) is decreased and average maximum score (b) is significantly lower in pfp-/- mice than in pfp+/- and pfp+/+ mice. Figures show results from two different experiments, with balanced groups, taken together. Maxscore was calculated as the mean maximum score for each individual mouse of a given genotype. *P < 0.05.
Figure 2 Histopathology of perforin-deficient pfp-/- mice compared with pfp+/- mice. (a) Joint of a pfp-/- mouse with severe inflammation and erosions, stage 4. (b) Paw of a heterozygous mouse in the same stage of disease. (c) Comparison of histological and clinical evaluation of the disease. Data are values of severity of pfp+/- mice (n = 6) and pfp-/- mice (n = 6), and are shown as means ± SEM.
Figure 3 Antibody response and T cell proliferation. (a) No significant variations in the levels of collagen-specific IgG antibodies between the perforin (pfp)-deficient mice (n = 6) and the control littermates (pfp+/- and pfp+/+; n = 10) were seen. Sera were obtained from the mice 21 days after immunization with collagen type II (CII) and complete Freund's adjuvant (CFA). Levels of collagen-specific IgG antibodies were measured by quantitative ELISA. (b) pfp-/- mice show an increased proliferation toward CII. Draining lymph nodes were obtained from pfp-/- (n = 3) and pfp+/+ (n = 3) mice 90 days after immunization with CII in CFA. The cells were cultured for 60 hours with collagen II in different concentrations, and then pulsed with [3H]thymidine. *P < 0.05.
Table 1 Mapping of DBA perforin-deficient mice
Marker Position (cM) Parental
D10Mit38 26.8 DBA
D10Mit44 27 DBA
D10Mit194 29 C57BL/6J
D10Mit61 32 C57BL/6J
D10Mit20 35 C57BL/6J
D10Mit31 36 C57BL/6J
D10Mit32 38 C57BL/6J
D10Mit186 40 DBA
D10Mit174 41 DBA
D10Mit175 41.8 DBA
D10Mit42 44 DBA
D10Mit261 47 DBA
D10Mit67 49 DBA
The mice were backcrossed for 14 generations to the DBA/1J background.
Table 2 Summary of disease-related data of perforin-deficient (pfp-/-) mice compared with pfp+/- and pfp+/+ mice
Genotype (n) Day of onset Incidence (%) Severity Maxscore AUC
Day 50 Day 75 Day 82
Pfp-/- (13) 59 ± 4.9 54 0.9 ± 0.8 2.1 ± 1.6 3.0 ± 1.5a 3.4 ± 1.4b 31.9 ± 23.0c
Pfp+/- (14) 52 ± 3.5 64 1.4 ± 0.6 3.8 ± 0.9 4.0 ± 0.9a 5.6 ± 1.2b 44.6 ± 12.3c
Pfp+/+ (14) 53 ± 3.6 64 1.3 ± 0.4 5.2 ± 1.2 5.6 ± 0.9a 7.3 ± 1.1b 59.4 ± 16.7
Evaluations and analyses were performed as described in Materials and methods. Maxscore was calculated as the mean of the maximum score value for each individual mouse of a given genotype. Data are group mean values ± SEM. Values with the same superscript letter are significantly different (P < 0.05). AUC, area under the curve.
==== Refs
Trentham DE Townes AS Kang AH Autoimmunity to type II collagen: an experimental model of arthritis J Exp Med 1977 146 857 868 894190 10.1084/jem.146.3.857
Courtenay JS Dallman MJ Dayan AD Martin A Mosedale B Immunisation against heterologous type II collagen induces arthritis in mice Nature 1980 283 666 668 6153460 10.1038/283666a0
Holmdahl R Jansson L Gullberg D Rubin K Forsberg PO Klareskog L Incidence of arthritis and autoreactivity of anti-collagen antibodies after immunization of DBA/1 mice with heterologous and autologous collagen II Clin Exp Immunol 1985 62 639 646 4085150
Luross JA Williams NA The genetic and immunopathological processes underlying collagen-induced arthritis Immunology 2001 103 407 416 11529930 10.1046/j.1365-2567.2001.01267.x
Wooley PH Luthra HS Stuart JM David CS Type II collagen-induced arthritis in mice. I. Major histocompatibility complex (I region) linkage and antibody correlates J Exp Med 1981 154 688 700 6792316 10.1084/jem.154.3.688
Cooper SM Sriram S Ranges GE Suppression of murine collagen-induced arthritis with monoclonal anti-Ia antibodies and augmentation with IFN-gamma J Immunol 1988 141 1958 1962 3139747
Hom JT Bendele AM Carlson DG In vivo administration with IL-1 accelerates the development of collagen-induced arthritis in mice J Immunol 1988 141 834 841 3260913
Killar LM Dunn CJ Interleukin-1 potentiates the development of collagen-induced arthritis in mice Clin Sci (Lond) 1989 76 535 538 2785895
Brahn E Peacock DJ Banquerigo ML Liu DY Effects of tumor necrosis factor alpha (TNF-alpha) on collagen arthritis Lymphokine Cytokine Res 1992 11 253 256 1467365
Firestein GS Yeo M Zvaifler NJ Apoptosis in rheumatoid arthritis synovium J Clin Invest 1995 96 1631 1638 7657832
Nakajima T Aono H Hasunuma T Yamamoto K Shirai T Hirohata K Nishioka K Apoptosis and functional Fas antigen in rheumatoid arthritis synoviocytes Arthritis Rheum 1995 38 485 491 7536416
Rouvier E Luciani MF Golstein P Fas involvement in Ca2+-independent T cell-mediated cytotoxicity J Exp Med 1993 177 195 200 7678113 10.1084/jem.177.1.195
Trapani JA Sutton VR Smyth MJ CTL granules: evolution of vesicles essential for combating virus infections Immunol Today 1999 20 351 356 10431154 10.1016/S0167-5699(99)01488-7
Gasque P Jones J Singhrao SK Morgan B Identification of an astrocyte cell population from human brain that expresses perforin, a cytotoxic protein implicated in immune defense J Exp Med 1998 187 451 460 9463395 10.1084/jem.187.4.451
Kagi D Ledermann B Burki K Seiler P Odermatt B Olsen KJ Podack ER Zinkernagel RM Hengartner H Cytotoxicity mediated by T cells and natural killer cells is greatly impaired in perforin-deficient mice Nature 1994 369 31 37 8164737 10.1038/369031a0
Malipiero U Frei K Spanaus KS Agresti C Lassmann H Hahne M Tschopp J Eugster HP Fontana A Myelin oligodendrocyte glycoprotein-induced autoimmune encephalomyelitis is chronic/relapsing in perforin knockout mice, but monophasic in Fas- and Fas ligand-deficient lpr and gld mice Eur J Immunol 1997 27 3151 3160 9464800
Peng SL Moslehi J Robert ME Craft J Perforin protects against autoimmunity in lupus-prone mice J Immunol 1998 160 652 660 9551899
Kagi D Odermatt B Seiler P Zinkernagel RM Mak TW Hengartner H Reduced incidence and delayed onset of diabetes in perforin-deficient nonobese diabetic mice J Exp Med 1997 186 989 997 9314549 10.1084/jem.186.7.989
Tada Y Ho A Koh DR Mak TW Collagen-induced arthritis in CD4- or CD8-deficient mice: CD8+ T cells play a role in initiation and regulate recovery phase of collagen-induced arthritis J Immunol 1996 156 4520 4526 8666829
Saha BK Typing of murine major histocompatibility complex with a microsatellite in the class II Eb gene J Immunol Methods 1996 194 77 83 8690943 10.1016/0022-1759(96)00065-8
Holmdahl R Jansson L Larsson A Jonsson R Arthritis in DBA/1 mice induced with passively transferred type II collagen immune serum. Immunohistopathology and serum levels of anti-type II collagen auto-antibodies Scand J Immunol 1990 31 147 157 1968678
Spielman J Lee RK Podack ER Perforin/Fas-ligand double deficiency is associated with macrophage expansion and severe pancreatitis J Immunol 1998 161 7063 7070 9862744
Badovinac VP Tvinnereim AR Harty JT Regulation of antigen-specific CD8+ T cell homeostasis by perforin and interferon-gamma Science 2000 290 1354 1358 11082062 10.1126/science.290.5495.1354
Sad S Kagi D Mosmann TR Perforin and Fas killing by CD8+ T cells limits their cytokine synthesis and proliferation J Exp Med 1996 184 1543 1547 8879227 10.1084/jem.184.4.1543
Zheng L Fisher G Miller RE Peschon J Lynch DH Lenardo MJ Induction of apoptosis in mature T cells by tumour necrosis factor Nature 1995 377 348 351 7566090 10.1038/377348a0
Ibrahim SM Koczan D Thiesen HJ Gene-expression profile of collagen-induced arthritis J Autoimmun 2002 18 159 167 11908948 10.1006/jaut.2001.0580
Hoang TR Hammermuller A Mix E Kreutzer HJ Goerlich R Kohler H Nizze H Thiesen HJ Ibrahim SM A proinflammatory role for Fas in joints of mice with collagen-induced arthritis Arthritis Res Ther 2004 6 R404 R414 15380040 10.1186/ar1205
Ma Y Liu H Tu-Rapp H Thiesen HJ Ibrahim SM Cole SM Pope RM Fas ligation on macrophages enhances IL-1R1-Toll-like receptor 4 signaling and promotes chronic inflammation Nat Immunol 2004 5 380 387 15004557 10.1038/ni1054
Campbell IK O'Donnell K Lawlor KE Wicks IP Severe inflammatory arthritis and lymphadenopathy in the absence of TNF J Clin Invest 2001 107 1519 1527 11413159
Mori L Iselin S De Libero G Lesslauer W Attenuation of collagen-induced arthritis in 55-kDa TNF receptor type 1 (TNFR1)-IgG1-treated and TNFR1-deficient mice J Immunol 1996 157 3178 3182 8816431
Sower LE Klimpel GR Hanna W Froelich CJ Extracellular activities of human granzymes. I. Granzyme A induces IL6 and IL8 production in fibroblast and epithelial cell lines Cell Immunol 1996 171 159 163 10.1006/cimm.1996.0187
Sower LE Froelich CJ Allegretto N Rose PM Hanna WD Klimpel GR Extracellular activities of human granzyme A. Monocyte activation by granzyme A versus alpha-thrombin J Immunol 1996 156 2585 2590 8786323
Griffiths GM Alpert S Lambert E McGuire J Weissman IL Perforin and granzyme A expression identifying cytolytic lymphocytes in rheumatoid arthritis Proc Natl Acad Sci USA 1992 89 549 553 1731326
Muller-Ladner U Kriegsmann J Tschopp J Gay RE Gay S Demonstration of granzyme A and perforin messenger RNA in the synovium of patients with rheumatoid arthritis Arthritis Rheum 1995 38 477 484 7536415
Mikulowska A Metz CN Bucala R Holmdahl R Macrophage migration inhibitory factor is involved in the pathogenesis of collagen type II-induced arthritis in mice J Immunol 1997 158 5514 5517 9164975
Anderson MS Venanzi ES Klein L Chen Z Berzins SP Turley SJ von Boehmer H Bronson R Dierich A Benoist C Projection of an immunological self shadow within the thymus by the Aire protein Science 2002 298 1395 1401 12376594 10.1126/science.1075958
| 15987490 | PMC1175039 | CC BY | 2021-01-04 16:02:38 | no | Arthritis Res Ther. 2005 May 20; 7(4):R877-R884 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1758 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17591598749210.1186/ar1759Research ArticleScreening of an endothelial cDNA library identifies the C-terminal region of Nedd5 as a novel autoantigen in systemic lupus erythematosus with psychiatric manifestations Margutti Paola [email protected] Maurizio [email protected] Fabrizio [email protected] Federica [email protected] Mauro 1Alessandri Cristiano [email protected] Alessandra [email protected]ò Rachele [email protected] Elisabetta [email protected] Guido [email protected] Elena [email protected] Dipartimento di Malattie Infettive, Parassitarie e Immunomediate, Istituto Superiore di Sanità, Rome, Italy2 Dipartimento di Medicina Sperimentale e Patologia, Cattedra di Reumatologia, Università 'La Sapienza', Rome, Italy3 Dipartimento di Clinica e Terapia Medica Applicata, Cattedra di Reumatologia, Università 'La Sapienza', Rome, Italy2005 20 5 2005 7 4 R896 R903 23 12 2004 20 1 2005 22 3 2005 18 4 2005 Copyright © 2005 Margutti et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Anti-endothelial-cell antibodies are associated with psychiatric manifestations in systemic lupus erythematosus (SLE). Our primary aim in this study was to seek and characterize molecules that behave as endothelial autoantigens in SLE patients with psychiatric manifestations. By screening a cDNA library from human umbilical artery endothelial cells with serum from an SLE patient with psychosis, we identified one positive strongly reactive clone encoding the C-terminal region (C-ter) of Nedd5, an intracytoplasmatic protein of the septin family. To evaluate anti-Nedd5 serum immunoreactivity, we analyzed by ELISA specific IgG responses in 17 patients with SLE and psychiatric manifestations (group A), 34 patients with SLE without psychiatric manifestations (group B), 20 patients with systemic sclerosis, 20 patients with infectious mononucleosis, and 35 healthy subjects. IgG specific to Nedd5 C-ter was present in 14 (27%) of the 51 SLE patients. The mean optical density value for IgG immunoreactivity to Nedd5 C-ter was significantly higher in patients of group A than in those of group B, those with infectious mononucleosis, or healthy subjects (0.17 ± 0.14 vs, respectively, 0.11 ± 0.07, P = 0.04; 0.11 ± 0.06, P = 0.034; and 0.09 ± 0.045, P = 0.003, on Student's t-test). Moreover, IgG immunoreactivity to Nedd5 C-ter was significantly higher in patients with systemic sclerosis than in patients of group B or healthy subjects (0.18 ± 0.18 vs, respectively, 0.11 ± 0.07, P = 0.046; and 0.09 ± 0.045, P = 0.003). The percentage of patients with anti-Nedd5 C-ter serum IgG was higher in group A than in group B (8 (47%) of 17, vs 6 (17%) of 34, P = 0.045, on Fisher's exact test). In order to clarify a possible mechanism by which Nedd5 might be autoantigenic, we observed that Nedd5 relocated from cytoplasm to the plasma membrane of EAhy926 endothelial cells after apoptotic stimuli. In conclusion, Nedd5 is a novel autoantigen of potential clinical importance that could be successfully used for a more thorough investigation of the pathogenesis of psychiatric manifestations in SLE. Although anti-Nedd5 autoantibodies are not specific to SLE, they are significantly associated with neuropsychiatric SLE and may represent immunological markers of psychiatric manifestations in this pathology.
==== Body
Introduction
Symptoms originating from the central nervous system occur in 14 to 75% of patients with systemic lupus erythematosus (SLE) and are extremely diverse, including neurological and psychiatric syndromes [1]. In 1999, the American College of Rheumathology defined 19 distinct neuropsychiatric syndromes associated with SLE, including psychosis and depression [2,3]. Neuropsychiatric SLE remains an enigmatic manifestation in lupus. In fact, conflicting results have been reported to clarify associations between neuropsychiatric manifestations and serum antibodies against neuronal antigens, ribosomes, and phospholipids [4]. Recently, we demonstrated an association between the presence of anti-endothelial-cell antibodies (AECAs) and psychiatric manifestations, such as psychosis and depression in SLE, suggesting a possible mechanism underlying psychiatric symptoms [5]. By activating endothelial cells, AECAs up-regulate the expression of adhesion molecules as well as the secretion of cytokines and chemokines. Until recently, few published data have been available on the identity of endothelial cell autoantigens in immune disorders [6-11]. Identifying endothelial autoantigens involved in the autoimmune processes during neuropsychiatric SLE could help to explain the pathogenetic mechanisms involved in the initiation and progression of psychiatric symptoms.
Our primary aim in this study was to seek and characterize molecules that behave as endothelial autoantigens in neuropsychiatric SLE. By screening a cDNA library from human umbilical artery endothelial cells (HUAECs) with serum from an SLE patient with psychosis, we identified one strongly reactive clone encoding the C-terminal region (C-ter) of Nedd5, an intracytoplasmatic protein of the septin family. To evaluate anti-Nedd5 serum immunoreactivity, we used ELISA to measure specific IgG responses in patients with SLE classified according to the presence of psychiatric manifestations, such as psychosis and depression. Data were compared with those of patients with systemic sclerosis (an autoimmune disease characterized by endothelium damage and the presence of AECAs), of patients with infectious mononucleosis, and of healthy subjects. Finally, we investigated by immunofluorescence the intracellular redistribution of Nedd5 in endothelial cells after apoptotic stimuli.
Materials and methods
Patients
For the present investigation, we studied sera from an SLE cohort of 51 outpatients attending the Rheumatology Division of the University of Rome 'La Sapienza'. All patients were diagnosed according to the American College of Rheumatology revised criteria for the classification of SLE [2]. This population of SLE patients had been previously characterized with regard to their psychiatric and autoantibody profiles [5]. For our present purposes, we studied 50 of the 51 sera, because the serum from one SLE patient with mood disorder had run out. In this study, we also included the serum from a patient seen in the meantime with SLE and acute psychosis, a very rare manifestation of neuropsychiatric SLE.
Patients were categorized as being in group A or group B on the basis of the clinical psychiatric examination, which was performed by means of the Structured Clinical Interview for Psychiatric Diagnosis [12]. A psychiatric diagnosis was assigned according to the Diagnostic and Statistical Manual of Mental Disorders IV [13]. Group A consisted of patients with psychosis (n = 3) and mood disorders (n = 14). Group B included patients without psychiatric manifestations (n = 18) and patients whose only psychiatric manifestation was anxiety disorder (n = 16). We did not include patients with anxiety disturbance in group A, because in most SLE patients anxiety is considered a secondary stress reaction and not a direct manifestation of neuropsychiatric SLE [5].
Current SLE disease activity was measured using the SLE Disease Activity Index (SLEDAI). We also studied as controls sera from 35 sex- and age-matched healthy subjects; 20 sera from patients with systemic sclerosis (18 female, 2 male; mean age 53 years, range 27 to 72) attending the Rheumatology Division of the University of Rome 'La Sapienza'; and 20 sex- and age-matched patients with infectious mononucleosis from the Department of Experimental Medicine and Pathology of the University of Rome 'La Sapienza'. Informed consent was obtained from each patient, and the local ethics committee approved the study protocol. The sera were stored at -20°C until they were assayed.
Immunoscreening of the cDNA expression library
A commercially available HUAEC cDNA library (Stratagene, Cambridge, UK) was used to screen for clones showing immunoreactivity with a serum from a patient with SLE, acute and active psychosis, and elevated titer of serum AECA. The expression library was screened essentially as previously described [14]. The serum was diluted 1:350 in PBS containing 1% milk and 0.05% Tween-20 and supplemented with 0.02% sodium azide. To reduce nonspecific binding to Escherichia coli (XL1-Blue MRF') (Stratagene) and phage vector, diluted pool was preadsorbed three times on nonrecombinant phage plaques. For primary immunoscreening, the library was plated out at 12,500 plaque-forming units per 140-mm plate, using XL1-Blue MRF' host cells in accordance with the supplier's instructions. In brief, nitrocellulose filters, incubated with 10 mM isopropyl β-D-1-thiogalactopyranoside (Sigma-Aldrich, St Louis, MO, USA) were overlaid onto the plates and incubated for 4 hours at 37°C. After blocking in 5% milk/PBS, the filters were incubated with the preadsorbed serum overnight at room temperature. After four washes with 0.05% PBS, Tween-20 membranes were incubated with a 1:3000 dilution of goat antihuman IgG (Bio-Rad, Richmond, CA, USA) in PBS containing 0.05% Tween-20 and 1% milk for 3 hours at room temperature. After a final four washes in 0.05% PBS Tween-20, membranes were incubated for 20 min with diaminobenzidine substrate (Sigma-Aldrich). Plaques corresponding to immunoreactive regions were cored from the original plate and resuspended in suspension medium containing 10 μl chloroform. Positive plaques were rescreened with the same serum to obtain the clonality.
Cloned phage showing immunoreactivity was recovered as pBluescript by single-stranded rescue using the helper phage (Stratagene) according to the supplier's instructions and used to transform SolR XL1cells. The nucleotide sequence of the cloned cDNA insertion was sequenced with automated sequencer ABI Prism 310 Collection (Applied Biosystems, Foster City, CA, USA) and sequences were then compared with the GenBank sequence database using both Fasta and Blast analysis [15,16]. To predict coiled-coil domain, we used the appropriate software .
Expression and purification of the recombinant antigen
The selected cDNA clone was subcloned into the Bam HI/HindIII restriction site of the QIA express vector, pQE30. To obtain the whole molecule of Nedd5, we amplified the cDNA of the library using specific primers designed from the 5' and 3' termini of sequence obtained in GenBank (accession number BC033559) with Bam HI/HindIII restriction site and cloned in the expression vector.
The fusion protein was expressed in Escherichia coli SG130009 cells, purified by affinity of NI-NTA resin for the 6X histidine tag and eluted under denaturing conditions (urea) in accordance with the supplier's instructions (Qiagen, Hilden, Germany). After purification, urea was removed by dialysis in PBS with decreasing concentrations of urea, with a last change of PBS alone overnight at 4°C. Protein concentration was determined by the Bio-Rad Bradford protein assay (Bio-Rad).
Indirect immunofluorescence assay
Hep-2 cells were directly stained with the mouse anti-Nedd5 polyclonal antiserum, obtained by standard immunization protocol, or with the corresponding mouse preimmune serum, in PBS containing 1% BSA. After washing three times with PBS, fluorescein-isothiocyanate (FITC)-conjugated antimouse IgG (γ-chain specific) (Sigma) were then added and incubation was at 4°C for 30 min.
Alternatively, EAhy926 human vascular endothelial cells [17] were grown to 60 to 70% confluence and seeded at 5 × 106 per well on glass cover slips. Cells, either untreated or treated with 20 ng/ml of tumor necrosis factor α (TNF-α) and 10 μg/ml of cycloheximide for 16 hours [18], were fixed with 4% formaldehyde in PBS for 30 min at 4°C. Alternatively, cells were permeabilized with acetone:methanol 1:1 (vol:vol) for 10 min at 4°C and then soaked in balanced salt solution (Sigma) for 30 min at 25°C. Cells were then incubated for 30 min at 25°C in the blocking buffer (2% bovine serum albumin in PBS, containing 5% glycerol and 0.2% Tween-20). Apoptosis was evaluated by propidium iodide staining, according to the method of Nicoletti and colleagues [19], and by the binding of FITC-conjugated Annexin V, using the Apoptest binding kit, containing annexin V-FITC and binding buffer. After washing three times with PBS, cells were incubated for 1 hour at 4°C with the mouse anti-Nedd5 polyclonal antiserum, obtained by standard immunization protocol, or with the corresponding mouse preimmune serum, in PBS containing 1% BSA. FITC-conjugated antimouse IgG (γ-chain specific) (Sigma) were then added and incubated at 4°C for 30 min. After washing with PBS, fluorescence was analyzed with an Olympus U RFL microscope (Olympus, Hamburg, Germany).
SDS–PAGE and immunoblotting
Immunoblotting, after 12% SDS–PAGE under reducing conditions, was performed as previously described [20]. In brief, Nedd5 C-ter was used as antigen at the concentration of 3 μg/lane and was revealed by human sera diluted 1:100, by a monoclonal antibody specific to six-histidine tail (Qiagen), or by the mouse polyclonal antiserum (1:200). Goat antihuman and antimouse IgG-labelled sera (Bio-Rad) were used as second antibodies. Strips were developed with peroxidase substrate, 3-3' -diaminobenzidine (Sigma). EAhy926 endothelial cells were harvested by mechanical scraping and centrifuged at 10,000 g for 30 min and the pellet was resuspended in the loading buffer under reducing conditions, boiled for 10 min, and loaded (100,000 cells/well) in a 10.5% SDS-polyacrylamide gel. After immunoblotting, the mouse polyclonal antiserum and the corresponding mouse preimmune serum were used to reveal the presence of Nedd5 in the cell preparation.
ELISA
ELISA for specific total IgG was developed essentially as previously described [14]. In brief, polystyrene plates (Dynex, Berlin, Germany) were coated with Nedd5 C-ter 0.5 μg/well in 0.05 μM NaHCO3 buffer, pH 9.5. Coated plates were incubated overnight at 4°C and then washed three times with PBS containing 0.05% Tween-20 in an automated washer (Wellwash 4, Labsystem, Turku, Finland). Plates were blocked with PBS Tween containing 3% gelatin (Bio-Rad), 100 μl/well, for 1 hour at room temperature and washed as previously described. Human sera were diluted in PBS Tween-20 and 1% gelatin (1:100 for total IgG) and pipetted onto plates at 100 μl per well. Plates were incubated for 1 hour at 20°C and washed as described. Peroxidase-conjugated goat antihuman IgG (Bio-Rad) was diluted 1:3000 in the same buffer. These dilutions were used as second antibodies and incubated (100 μl/well) for 1 hour at 20°C. o-Phenylenediamine dihydrochloride (Sigma) was used as a substrate and absorbance was measured at 490 nm. Means + 2 standard deviations (SD) of the absorbance reading of the healthy controls were considered the cutoff levels for positive reactions. All assays were performed in quadruplicate. Data were presented as the mean optical density (OD) corrected for background (wells without coated antigen). The results of unknown samples on the plate were accepted if internal controls (two serum samples, one positive and one negative) had an absorbance reading within mean ± 10% of previous readings. To inhibit specific IgG, the sera from three patients with SLE, anti-Nedd5 C-ter IgG positive, were diluted 1:50 in PBS-Tween and were incubated overnight at room temperature in 10 μg/ml of Nedd5 C-ter according to the method reported by Huang and colleagues [21]. As a negative control, the sera were pre-incubated with 40 μg/ml of BSA.
Cultures of human umbilical-vein-derived endothelial cells at the third to fourth passage were used to detect AECA (IgG), using a cell-surface ELISA on living cells, as previously reported [5]. AECAs were expressed as binding index (BI) equal to 100 × (S-A) / (B-A), where S is the OD of the sample tested, A is the OD obtained with only the secondary antibody, and B the OD of a positive reference serum. AECAs were considered positive when BI was higher than the cutoff value (mean+2 SD of 66 healthy controls) corresponding to 50% of a positive reference serum from a patients with SLE. Antibodies against cardiolipin, β2 glycoprotein I, Ro/SSA, Ro/SSA 52, La/SSB, glial fibrillary acidic protein, ribosomal P protein, and nucleosome IgG were tested as previously described [5]
Statistical analysis
Unless otherwise specified, all values are means ± SD. The Fisher exact test and χ2 analysis were used to evaluate differences between percentages; Student's t-test was used to evaluate differences between arithmetic means. P values less than 0.05 indicated statistically significant differences.
Results
Immunoscreening of HUAEC expression library
Immunoscreening of the HUAEC expression library with IgG from the serum of a patient with SLE and active psychosis identified a strongly reactive clone. The amino acid sequence, predicted from the 335-base-pair open reading frame of this clone, is 109 residues long and has 100% identity with the C-terminal subunit of Nedd5, a protein of the septin family (Fig. 1a,b). The search for possible coiled-coil domains in the database disclosed a coiled-coil domain in this amino acid region. Because preliminary experiments showed that the whole molecule and the C-terminal subunit gave equivalent serological results (data not shown), in serological tests we used the C-terminal subunit (Nedd5 C-ter), which contains the immunoreactive epitopes. The molecular size predicted from the amino acid sequence of 13.6 kDa, the purity, and the immunoreactivity of the expressed recombinant protein was confirmed by 12% SDS–PAGE and immunoblotting (Fig. 1c).
ELISA for anti-Nedd5 C-ter IgG
IgGs specific to Nedd5 C-ter in ELISA were present in 14 (27.4%) of 51 SLE patients and did not correlate with the presence of AECAs previously studied in the same population of patients [5]. Moreover, no significant correlation was found between the presence of anti-Nedd5-C-ter IgG and the presence of antibodies against cardiolipin, β2 glycoprotein I, Ro/SSA, Ro/SSA52, La/SSB, glial fibrillary acidic protein, ribosomal P protein, or nucleosome IgG (data not shown). To assess the specificity of ELISA, we preadsorbed sera from three SLE patients positive to Nedd5 C-ter with Nedd5 C-ter itself, and we observed a complete inhibition of reactivity (data not shown).
The mean OD value for IgG immunoreactivity to Nedd5 C-ter was significantly higher in patients of group A than in those of group B, those with infectious mononucleosis, or healthy subjects (0.17 ± 0.14 vs, respectively, 0.11 ± 0.07, P = 0.04; 0.11 ± 0.06, P = 0.034; 0.09 ± 0.045, P = 0.003, on Student's t-test). Moreover, IgG immunoreactivity to Nedd5 C-ter was significantly higher in patients with systemic sclerosis than in patients of group B or in healthy subjects (0.18 ± 0.18 vs, respectively, 0.11 ± 0.07, P = 0.046; and 0.09 ± 0.045, P = 0.003) (Fig. 2). The percentage of patients with anti-Nedd5 C-ter serum IgG was higher in group A than in group B (8 (47%) of 17, vs 6 (17%) of 34, P = 0.045 on the Fisher exact test). No correlation was observed between the presence of anti-Nedd5 C-ter antibodies and clinical features of SLE or SLEDAI (Table 1).
Localization of Nedd5 in HEp-2 cells and in EAhy926 cells
We analyzed primarily the localization of Nedd5 in HEp-2 cells, a conventional cell line used in clinical laboratories because of their active proliferation. As expected, the immunolabelling with a mouse polyclonal antiserum specific to the recombinant Nedd5 C-ter appeared confined to the cytoplasm, where staining of the contractile ring was evident. No staining was observed on the cell surface (Fig. 3a). In contrast, no staining with the mouse preimmune serum was observed, indicating that the immunolabelling was specific for Nedd5 C-ter (data not shown).
Immunoblotting of EAhy926 cells, revealed with the mouse polyclonal antiserum specific to the recombinant Nedd5 C-ter, showed a single band corresponding to the native Nedd5 protein at the expected molecular size of 42 kDa deduced from the amino acid sequence (Fig. 3b).
We analyzed Nedd5 relocation to the plasma membrane of EAhy926 cells during apoptosis, by immunofluorescence assay with the mouse polyclonal antiserum anti-Nedd5 C-ter (Fig. 3c). Untreated cells showed virtually no staining on the plasma membrane (i). On the contrary, an uneven surface distribution of anti-Nedd5 staining was observed in cells treated with TNF-α plus cycloheximide (ii). These findings were confirmed in permeabilized cells. Indeed, untreated control cells showed anti-Nedd5 staining confined to the cytoplasm, with pronounced perinuclear granularity and without significant staining of the cell surface (iii). Induction of apoptosis by treatment with TNF-α plus cycloheximide changed the cellular distribution of Nedd5, with an uneven surface distribution of the staining that appeared in the cytoplasm and around plasma membrane (iv). No staining with the mouse preimmune serum was observed in any of the samples, indicating that the observed immunolabelling was specific for Nedd5. Apoptosis was checked by testing the exposure of phosphatidylserine on the cell surface by the binding of FITC-conjugated Annexin V, which revealed that up to 80% of the cells were positive (data not shown).
Discussion
In this study, we used a molecular cloning strategy to identify endothelial autoantigens in SLE patients. Results provide evidence that the C-terminal region of Nedd5 is a novel autoantigen with a role in neuropsychiatric manifestations. Nedd5 is a mammalian septin known to associate with actin-based structures such as the contractile ring and stress fibers [22,23]. The septins are a family of cytoskeletal GTPases that play an essential role in cytokinesis in yeast and mammalian cells [24]. Nedd5 is predominantly expressed in the nervous system and may contribute to the formation of neurofibrillary tangles as integral constituents of paired helical filaments in Alzheimer's disease [25,26].
To our knowledge, this is the first report describing an immune response against a protein of the septin family. This study provides evidence that Nedd5 molecules are expressed on the cell surface after apoptotic stimuli, suggesting a possible mechanism by which Nedd5 may be autoantigenic. Indeed, apoptosis may play an important role in bypassing tolerance to intracellular autoantigens. The specific modification of autoantigens and their redistribution into blebs at the surface of apoptotic cells may contribute to the induction of autoimmune responses [27,28]. Moreover, apoptotic defects and impaired removal of apoptotic cells could contribute to an overload of autoantigens in the circulation or in target tissues that could become available to initiate an autoimmune response [29]. In susceptible individuals, this can lead to autoantibody-mediated tissue damage.
Interestingly, the C-terminal region of Nedd-5 displays a coiled-coil domain. Several autoimmune autoantigens are characterized by the presence of such a domain [30]. Coiled-coil proteins may be exposed to the immune system as surface structures in aberrant disease states associated with unregulated cell death and could become autoimmune targets [30].
Even though in this study we used an endothelial cDNA expression library and we screened it with a serum with an elevated AECA titer, we found no significant correlation between the presence of AECAs and the presence of anti-Nedd5 antibodies in patients with SLE (data not shown). This finding is not surprising, since the cell-surface ELISA on living cells used to detect AECAs reveals only plasma membrane antigens, whereas Nedd5, which is normally confined within the cytoplasm, becomes exposed on the cell surface after triggering apoptosis. Interestingly, we found such antibodies in a large proportion of patients with systemic sclerosis, a pathology in which endothelial damage may often occur. However, we cannot exclude the possibility that the autoimmune response we observed was generated against Nedd5 present in other cellular compartments, such as the nervous system.
An association between serum AECAs and psychosis or depression in patients with SLE has been recently reported, strengthening the view of a possible implication of AECAs in the development of psychiatric disorders in SLE [5]. In this study, attempting to identify a possible molecular target of AECAs in an SLE patient with active psychosis, analyzing the same population of patients as in the previous investigation [5], we demonstrated an association between serum IgG specific to the C-terminal region of Nedd5 and psychiatric manifestations in patients with SLE. Notably, all of the three patients with psychosis had serum IgG to Nedd5 C-ter. Overall, although anti-Nedd5 autoantibodies are not specific to SLE, they are significantly associated with neuropsychiatric SLE and could be immunological markers of psychiatric manifestations in this pathology. The unanswered question is whether anti-Nedd5 C-ter antibodies can cause direct damage, thus contributing to the pathogenesis of psychiatric manifestations, or whether they are an epiphenomenon of these disorders. Further studies are in progress in order to clarify the effective role of anti-Nedd5 C-ter antibodies in vivo.
Conclusion
In the present study, we identified Nedd5 C-ter as a novel autoantigen in SLE. This result is of clinical importance and may be a valuable tool in the diagnosis of neuropsychiatric SLE. In addition, having this recombinant antigen may help in defining the precise role that specific autoantibodies may play in the autoimmune mechanisms underlying psychiatric manifestations in SLE.
Abbreviations
AECA = anti-endothelial-cell antibody; BSA = bovine serum albumin; C-ter = C-terminal region; ELISA = enzyme-linked immunosorbent assay; FITC = fluorescein isothiocyanate; HUAEC = human umbilical artery endothelial cell; OD = optical density; PBS = phosphate-buffered saline; SD = standard deviation; SLE = systemic lupus erythematosus; SLEDAI = SLE Disease Activity Index; TNF-α = tumor necrosis factor α.
Competing interests
The author(s) declare that they have no competing interests
Authors' contributions
PM carried out the screening of the library and participated in the design of the study and in the analysis of data. MS carried out the experiments on endothelial cell line and apoptosis and participated in the design of the study and helped to draft the manuscript. FC participated in the design of the study and in analysis of data and helped to draft the manuscript. FD carried out the cloning and sequencing of cDNA and protein expression and contributed in the interpretation of data. MR carried out the ELISA experiments and participated in analysis of data. CA performed the statistical analysis and the clinical associations. AS participated in the analysis and interpretation of data and in the revision of the manuscript. RR participated in the design of the study and in the revision of the manuscript. EP participated in analysis of data. GV participated in the design and revision of the study. EO conceived of the study, participated in its design and coordination, and drafted the manuscript. All authors read and approved the final manuscript.
Acknowledgements
This work was supported by an Istituto Superiore di Sanità grant no. C3N3 and AE13.
Figures and Tables
Figure 1 Nucleotide and amino acid sequences and immunochemical characterization of the C-terminal region of Nedd5. (a) The nucleotide sequence of 335 base pairs of the cloned cDNA insertion was sequenced with automated sequencer ABI Prism 310 Collection. (b) The amino acid sequence predicted from the nucleotide sequence is 109 residues long. The sequence compared with the GenBank sequence database using both Fasta and Blast analysis has 100% identity with the C-terminal subunit of Nedd5 (accession number Q15019). (c) The molecular size and the purity of the expressed protein were confirmed by 12% SDS–PAGE stained by Coomassie blue (lane 1). Immunoreactivity was analyzed by immunoblotting: monoclonal antibody antihistidine tail (lane 2); mouse polyclonal antiserum specific to Nedd5 C-ter (lane 3); representative sera from three patients with SLE IgG positive to Nedd5 (lanes 4,5,6); representative serum from a patient with SLE IgG negative to Nedd5 C-ter (lane 7); representative serum from a healthy subject (lane 8); control lane without serum (lane 9).
Figure 2 Anti- Nedd5 C-ter antibodies in patients with SLE with and without psychiatric manifestations. Dot plot of anti-Nedd5 C-ter IgG in systemic lupus erythematosus (SLE) patients with psychiatric manifestations (group A, n = 17), in SLE patients without psychiatric manifestations other than anxiety (group B, n = 34), in systemic sclerosis (SSc) patients (n = 20), and in patients with infectious mononucleosis (n = 20). Each dot represents a subject. The samples were considered positive when their optical density was higher than the cutoff value (mean + 2 SD for 35 healthy controls). The broken line represents the cutoff (0.18). *Group A vs group B, P = 0.04; #group A vs infectious mononucleosis, P = 0.034; °group A vs healthy controls, P = 0.003; ^SSc patients vs group B, P = 0.046; §SSc patients vs healthy controls, P = 0.003 (Student's t-test).
Figure 3 Localization of Nedd5 in HEp-2 cells and in EAhy926 cells. (a) Immunofluorescence analysis of Nedd5 localization in HEp-2 cells. The mouse polyclonal antiserum specific to Nedd5 C-ter was used to analyze the cellular distribution of Nedd5. (b) 10.5% SDS–PAGE and immunoblotting of EAhy926 cells. EAhy926 cells were centrifuged and the pellets were redissolved in the loading buffer under reducing conditions (100,000 cells/well). 10.5% SDS–PAGE stained by Coomassie blue (1); immunoblotting was performed with the mouse polyclonal antiserum specific to Nedd5 C-ter (2) and with the mouse preimmune serum (3). (c) Immunofluorescence analysis of Nedd5 localization in endothelial cells under physiological conditions and during apoptosis. EAhy926 cells were treated with tumor necrosis factor α (TNF-α) (20 μg/ml) plus cycloheximide (10 μg/ml) for 16 hours to induce apoptosis. The mouse polyclonal antiserum specific to Nedd5 C-ter was used to analyze the cellular distribution of Nedd5. (i) Untreated cells, fixed with 4% formaldehyde in PBS; (ii) TNF-α plus cycloheximide-treated cells, fixed with 4% formaldehyde in PBS; (iii) untreated cells, permeabilized with acetone:methanol 1:1 (vol:vol); (iv) TNF-α plus cycloheximide-treated cells, permeabilized with acetone : methanol 1:1 (vol:vol).
Table 1 Clinical characteristics of SLE patients according to psychiatric symptoms and anti-Nedd5 C-ter IgG
Group Aa (n = 17) Group Ba (n = 34)
Characteristic Anti-Nedd5 C-ter IgG (n = 8) No anti-Nedd5 C-ter IgG (n = 9) Anti-Nedd5 C-ter IgG (n = 6) No anti-Nedd5 C-ter IgG (n = 28)
Age (y), mean (range) 38.6 (26–52) 37.1 (23–50) 34.5 (28–48) 37.2 (14–70)
Sex (males/females) 1/7 2/7 1/5 3/25
Disease duration (y), mean (range) 11.2 (7–21) 6.8 (0.5–19) 9 (5–15) 12.1 (0.5–24)
Arthritis, no. (%) 5 (62.5) 9 (100) 5 (83.3) 18 (64.2)
Neurological involvement, no. (%) 2 (25) 4 (44.4) 2 (33.3) 11 (39.2)
Renal involvement, no. (%) 3 (37.5) 0 3 (50) 9 (32.1)
Cytopenia, no. (%) 5 (62.5) 7 (77.7) 4 (66.6) 16 (57.1)
Serositis, no. (%) 2 (25) 2 (22.2) 2 (33.3) 8 (28.5)
SLEDAI >3, no. (%) 5 (62.5) 4 (44.4) 4 (66.6) 12 (42.8)
aGroup A, patients with depression or psychosis; group B, patients without psychiatric manifestations other than anxiety. Differences between the groups as measured by the χ2 test were not statistically significant. SLE, systemic lupus erythematosus; SLEDAI, Systemic Lupus Erythematosus Disease Activity Index.
==== Refs
Hanly JG Neuropsychiatric lupus Curr Rheumatol Rep 2001 3 205 212 11352789
ACR ad Hoc Committee on Neuropsychiatric Lupus Nomenclature The American College of Rheumatology nomenclature and case definitions for neuropsychiatric lupus syndromes Arthritis Rheum 1999 42 599 608 10211873 10.1002/1529-0131(199904)42:4<599::AID-ANR2>3.0.CO;2-F
Ainiala H Hietaharju A Loukkola J Peltola J Korpela M Metsanoja R Auvinen A Validity of the new American College of Rheumatology criteria for neuropsychiatric lupus syndromes: a population-based evaluation Arthritis Rheum 2001 45 419 423 11642640 10.1002/1529-0131(200110)45:5<419::AID-ART360>3.0.CO;2-X
Greenwood DL Gitlits VM Alderuccio F Sentry JW Toh BH Autoantibodies in neuropsychiatric lupus Autoimmunity 2002 35 79 86 12071439 10.1080/08916930290016547
Conti F Alessandri C Bompane D Bombardieri M Spinelli FR Rusconi AC Valesini G Autoantibody profile in systemic lupus erythematosus with psychiatric manifestations: a role for anti-endothelial-cell antibodies Arthritis Res Ther 2004 6 R366 R372 15225372 10.1186/ar1198
Lee KH Chung HS Kim HS Oh SH Ha MK Baik JH Lee S Bang D Human alpha-enolase from endothelial cells as a target antigen of anti-endothelial cell antibody in Behcet's disease Arthritis Rheum 2003 48 2025 2035 12847697 10.1002/art.11074
Maehnss K Kobarg J Schmitt WH Hansen HP Lange H Csernok E Gross WL Lemke H Vitronectin- and fibronectin-containing immune complexes in primary systemic vasculitis J Autoimmun 2002 18 239 250 12126637 10.1006/jaut.2002.0582
Moscato S Pratesi F Bongiorni F Scavuzzo MC Chimenti D Bombardieri S Migliorini P Endothelial cell binding by systemic lupus antibodies: functional properties and relationship with anti-DNA activity J Autoimmun 2002 18 231 238 12126636 10.1006/jaut.2002.0583
Okawa-Takatsuji M Aotsuka S Uwatoko S Takaono M Iwasaki K Kinoshita M Sumiya Endothelial cell-binding activity of anti-U1-ribonucleoprotein antibodies in patients with connective tissue diseases Clin Exp Immunol 2001 126 345 354 11703381 10.1046/j.1365-2249.2001.01669.x
Frampton G Moriya S Pearson JD Isenberg DA Ward FJ Smith TA Panayiotou A Staines NA Murphy JJ Identification of candidate endothelial cell autoantigens in systemic lupus erythematosus using a molecular cloning strategy: a role for ribosomal P protein P0 as an endothelial cell autoantigen Rheumatology (Oxford) 2000 39 1114 1120 11035132
Yazici ZA Behrendt M Cooper D Goodfield M Partridge L Lindsey NJ The identification of endothelial cell autoantigens J Autoimmun 2000 15 41 49 10936027 10.1006/jaut.2000.0391
First MB Spitzer RL Gibbon M Williams JB Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I/P), Computer Program Handbook Version 20 1996 New York: New York State Psychiatric Institute, Biometrics Research Department
American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders 1994 4 Washington, DC: American Psychiatric Association
Margutti P Delunardo F Sorice M Valesini G Alessandri C Capoano R Profumo E Siracusano A Salvati B Rigano R Screening of a HUAEC cDNA library identifies actin as a candidate autoantigen associated with carotid atherosclerosis Clin Exp Immunol 2004 137 209 215 15196264 10.1111/j.1365-2249.2004.02491.x
Pearson WR Lipman DJ Improved tools for biological sequence comparison Proc Natl Acad Sci USA 1988 85 2444 2448 3162770
Altschul SF Gish W Miller W Myers EW Lipman DJ Basic local alignment search tool J Mol Biol 1990 215 403 410 2231712 10.1006/jmbi.1990.9999
Edgell CJS McDonald CC Graham JB Permanent cell line expressing human factor VIII-related antigen established by hybridization Proc Natl Acad Sci USA 1983 80 3734 3737 6407019
Xu J Yeh CH Chen S He L Sensi SL Canzoniero LM Choi DW Hsu CY Involvement of de novo ceramide biosynthesis in tumor necrosis factor-alpha/cycloheximide-induced cerebral endothelial cell death J Biol Chem 1998 273 16521 16526 9632721 10.1074/jbc.273.26.16521
Nicoletti I Migliorati G Pagliacci MC Grignani F Riccardi C A rapid and simple method for measuring thymocyte apoptosis by propidium iodide staining and flow cytometry J Immunol Methods 1991 139 271 279 1710634 10.1016/0022-1759(91)90198-O
Margutti P Ortona E Vaccari S Barca S Riganò R Teggi A Muhschlegel F Frosch M Siracusano A Cloning and expression of a cDNA encoding an elongation factor 1β/δ protein from Echinococcus granulosus with immunogenic activity Parasite Immunol 1999 21 485 492 10476057 10.1046/j.1365-3024.1999.00246.x
Huang X Johansson SGO Zargari A Zargari A Nordvall SL Allergen cross-reactivity between Pityrosporum orbiculare and Candida albicans Allergy 1995 50 648 656 7503400
Vega IE Hsu SC The septin protein Nedd5 associates with both the exocyst complex and microtubules and disruption of its GTPase activity promotes aberrant neurite sprouting in PC12 cells Neuroreport 2003 14 31 37 12544826 10.1097/00001756-200301200-00006
Surka MC Tsang CW Trimble WS The mammalian septin MSF localizes with microtubules and is required for completion of cytokinesis Mol Biol Cell 2002 13 3532 3545 12388755 10.1091/mbc.E02-01-0042
Martinez C Ware J Mammalian septin function in hemostasis and beyond Exp Biol Med (Maywood) 2004 229 1111 1119 15564437
Peng XR Jia Z Zhang Y Ware J Trimble WS The septin CDCrel-1 is dispensable for normal development and neurotransmitter release Mol Cell Biol 2002 22 378 387 11739749 10.1128/MCB.22.1.378-387.2002
Kinoshita A Kinoshita M Akiyama H Tomimoto H Akiguchi I Kumar S Noda M Kimura J Identification of septins in neurofibrillary tangles in Alzheimer's disease Am J Pathol 1998 153 1551 1560 9811347
Cline AM Radic MZ Apoptosis, subcellular particles, and autoimmunity Clin Immunol 2004 112 175 182 15240161 10.1016/j.clim.2004.02.017
White S Rosen A Apoptosis in systemic lupus erythematosus Curr Opin Rheumatol 2003 15 557 562 12960480 10.1097/00002281-200309000-00006
Pittoni V Valesini G The clearance of apoptotic cells: implications for autoimmunity Autoimmun Rev 2002 1 154 161 12849009 10.1016/S1568-9972(02)00032-0
Stinton LM Eystathioy T Selak S Chan EKL Fritzler MJ Autoantibodies to protein transport and messenger RNA processing pathways: endosomes, lysosomes, Golgi complex, proteasomes, assemblyosomes, exosomes, and GW bodies Clinical Immunol 2004 110 30 44 14962794 10.1016/j.clim.2003.10.005
| 15987492 | PMC1175040 | CC BY | 2021-01-04 16:02:37 | no | Arthritis Res Ther. 2005 May 20; 7(4):R896-R903 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1759 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17601598749110.1186/ar1760Research ArticleNADPH-oxidase-driven oxygen radical production determines chondrocyte death and partly regulates metalloproteinase-mediated cartilage matrix degradation during interferon-γ-stimulated immune complex arthritis van Lent Peter LEM [email protected] Karin CAM 1Blom Arjen B 1Sloetjes Annet 1Holthuysen Astrid EM 1Kolls Jay 2Van De Loo Fons AJ 1Holland Steven M 3Van Den Berg Wim B 11 Department of Rheumatology, University Medical Centre, St Radboud, Nijmegen, The Netherlands2 Department of Pediatrics, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA3 Department of Host Defenses, National Institute of Allergy and Infectious Diseases, Bethesda, MD, USA2005 20 5 2005 7 4 R885 R895 21 7 2004 24 9 2004 5 4 2005 19 4 2005 Copyright © 2005 van Lent et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
In previous studies we have found that FcγRI determines chondrocyte death and matrix metalloproteinase (MMP)-mediated cartilage destruction during IFN-γ-regulated immune complex arthritis (ICA). Binding of immune complexes (ICs) to FcγRI leads to the prominent production of oxygen radicals. In the present study we investigated the contribution of NADPH-oxidase-driven oxygen radicals to cartilage destruction by using p47phox-/- mice lacking a functional NADPH oxidase complex. Induction of a passive ICA in the knee joints of p47phox-/- mice resulted in a significant elevation of joint inflammation at day 3 when compared with wild-type (WT) controls as studied by histology. However, when IFN-γ was overexpressed by injection of adenoviral IFN-γ in the knee joint before ICA induction, a similar influx of inflammatory cells was found at days 3 and 7, comprising mainly macrophages in both mouse strains. Proteoglycan depletion from the cartilage layers of the knee joints in both groups was similar at days 3 and 7. Aggrecan breakdown in cartilage caused by MMPs was further studied by immunolocalisation of MMP-mediated neoepitopes (VDIPEN). VDIPEN expression in the cartilage layers of arthritic knee joints was markedly lower (between 30 and 60%) in IFN-γ-stimulated arthritic p47phox-/- mice at day 7 than in WT controls, despite significant upregulation of mRNA levels of various MMPs such as MMP-3, MMP-9, MMP-12 and MMP-13 in synovia and MMP-13 in cartilage layers as measured with quantitative RT-PCR. The latter observation suggests that oxygen radicals are involved in the activation of latent MMPs. Chondrocyte death, determined as the percentage of empty lacunae in articular cartilage, ranged between 20 and 60% at day 3 and between 30 and 80% at day 7 in WT mice, and was completely blocked in p47phox-/- mice at both time points. FcγRI mRNA expression was significantly lower, and FcγRII and FcγRIII were higher, in p47phox-/- mice than in controls. NADPH-oxidase-driven oxygen radical production determines chondrocyte death and aggravates MMP-mediated cartilage destruction during IFN-γ-stimulated IC-mediated arthritis. Upregulation of FcγRI by oxygen radicals may contribute to cartilage destruction.
==== Body
Introduction
During rheumatoid arthritis (RA), large numbers of inflammatory cells, mainly macrophages, migrate into the synovial layer [1]. Many of these macrophages become activated by mechanisms that are as yet unknown. Activated macrophages produce cytokines such as tumour necrosis factor-α (TNFα) and interleukin-1 (IL-1) and enzymes such as the metalloproteinase family, which can mediate severe cartilage destruction. A strong correlation was found between the number of activated macrophages and cartilage erosion [2]. Important triggers of macrophages are IgG-containing immune complexes, which are found in large amounts in the joints of many RA patients [3]. In previous studies we have found, by comparing various experimental arthritis models, that severe cartilage destruction developed mainly when immune complexes were present [4]. Severe cartilage destruction is thereby defined as chondrocyte death and cartilage matrix destruction. The latter is induced predominantly by metalloproteinases (MMPs), which are released in a latent form. Upon activation these enzymes degrade the collagen type II network in the cartilage resulting in irreversible erosion [5]. During immune complex (IC)-mediated arthritides, synovial macrophages seemed to be dominant factors in the induction of severe cartilage destruction [6].
IgG-containing ICs communicate with macrophages with FcγR. Three classes have been described, and previous studies in our laboratory showed that absence of the activating FcγRI and FcγRIII completely abrogated severe cartilage destruction [7-9].
The mechanism of FcγR-mediated chondrocyte death and MMP-mediated cartilage destruction is not known. However, we found recently that FcγRI is the dominant activating FcγR causing cartilage destruction [10,11]. In T cell-driven immune complex arthritis (ICA), chondrocyte death in FcγRI-/- was completely abrogated, whereas MMP-mediated cartilage destruction was significantly diminished [12]. Moreover, ICA stimulated by local overexpression of the T cell factor IFN-γ showed pronounced chondrocyte death that was also completely mediated by FcγRI [13].
Binding of ICs to FcγRI causes intracellular signalling and triggers activation of the multicomponent enzyme NADPH oxidase, which catalyses the production of oxygen species [14]. The latter have been shown to be involved in cell death [15,16] and in the activation of metalloproteinases [17]. The active central role in NADPH oxidase is as the transmembrane cytochrome b556, which comprises two subunits, gp91phox and p22phox. p47phox is the cytosolic component of the NADPH oxidase complex that translocates to the membrane and associates with cytochrome b556 to form the active complex that catalyses the reduction of oxygen to superoxide. Functionally, p47phox increases the binding of p67phox to cytochrome b556 about 100-fold [18-20]. IFN-γ strongly stimulates p91 and also the expression of FcγRI. Binding of ICs to FcγRI further increases NADPH oxidase activity [21]. Phospholipase D-1 has been shown to be an important mediator between FcγRI signalling and the activation of NADPH oxidase [14,22]. The combination of IFN-γ and FcγRI stimulation might therefore result in a strong stimulation of NADPH oxidase, catalysing the production of large amounts of superoxide.
In the present study we investigated the effect of NADPH-oxidase-driven oxygen radicals in the generation of severe cartilage destruction during IFN-γ-accelerated ICA. For that purpose mice in which the p47phox gene had been knocked out were used; they are unable to form a functional NADPH oxidase complex [23] and are therefore unable to make oxygen species by the NADPH oxidase pathway. However, other oxygen-radical-producing pathways remain intact. We found that chondrocyte death was completely abrogated, whereas MMP-mediated cartilage destruction was significantly inhibited. FcγRI expression was significantly downregulated; in contrast, MMP gene expression in the synovium was higher, suggesting that oxygen radicals are involved in the activation step of MMPs.
Materials and methods
Animals
NADPH-oxidase-deficient (C57BL/6-p47phox- /-) mice were generated as described previously [23], and lack the cytosolic p47phox subunit of the NADPH oxidase multicomponent system. The knockout mice were backcrossed to the C57BL6 background for 15 generations; C57BL/6 mice (obtained from the Jackson Laboratory, Bar Harbor, ME, USA) were used as controls. In some experiments p47phox- /- mice of intercross progeny (C57BL/6 × 129Sv) were used with their proper controls. Colonies were maintained at the National Institutes of Health (Bethesda, MD, USA). All mice were housed under specified pathogen-free conditions during breeding and experiments. Mice received autoclaved chow and acidified water ad libitum. Only healthy mice were used in the experiments and were age-matched (10 to 20 weeks) and sex-matched for each set of experiments. All experiments were approved by local authorities of the Animal Care and Use Committee (DEC 98.22) and performed by personnel certified by the Dutch Ministry of Well-being, Public Health and Culture.
Overexpression of IFN-γ in vivo with an adenoviral construct
The recombinant adenovirus encoding murine IFN-γ (AdIFN-γ) was generated as described previously [24]. Knee joints of naive mice were injected intra-articularly with 6 μl of AdIFN-γ (107 plaque-forming units). At different time points (days 3 and 7), patellae with adjacent synovium were dissected in a standardised manner [25] and synovium biopsies were taken using a biopsy punch with a diameter of 3 mm. Total RNA was extracted in 1 ml of TRIzol reagent and used for quantitative PCR as described below. AdIFN-γ was injected intra-articularly 1 day before arthritis induction.
Induction of immune complex arthritis
ICA was passively induced by injecting 3 μg of poly-(L-lysine)-coupled lysozyme into the knee joints of mice that had previously (16 hours earlier) received, intravenously, polyclonal antibodies directed against lysozyme. These antibodies were raised in rabbits.
Histology of arthritic knee joints
Total knee joints of mice were isolated 3 and 7 days after arthritis onset. Mice were killed by cervical dislocation, knee joints were decalcified, dehydrated, and embedded in paraffin. Tissue sections (7 μm) were stained with haematoxylin and eosin. Seven sections spaced 70 μm apart representing the whole knee joint were measured to obtain a statistically justified result. Histopathological changes were scored by grading the inflammation on a scale from 0 (no inflammation) to 3 (severe inflamed joint) as the influx of inflammatory cells into synovium and joint cavity. To study proteoglycan (PG) depletion from cartilage matrix, sections were stained with safranin O followed by counterstaining with fast green. PG depletion (loss of red staining) from various cartilage layers was determined by using an arbitrary scale from 0 to 3. Normal cartilage was assigned the value 0, and cartilage fully depleted of PGs was taken as 3. Chondrocyte death was determined in total knee joint sections stained with haematoxylin and eosin. Chondrocyte death was determined as the percentage of the area of the cartilage containing empty lacunae in relation to the total area. All experiments were scored separately and independently from each other.
Immunohistochemical detection of the identification marker of macrophages
F4/80, a murine macrophage membrane antigen, was detected with a specific rat anti-mouse F4/80 IgG. Primary antibodies were detected with rabbit anti-rat IgG and avidin–horseradish peroxidase conjugate. Finally, sections were counterstained with Mayer's haematoxylin (Merck, Darmstadt, Germany).
Immunolocalisation of MMP-induced neoepitope (VDIPEN)
For immunohistochemical analysis of MMP-induced neoepitopes, sections were deparaffinised, rehydrated and digested with chondroitinase ABC (Sigma; 0.25 U/ml in 0.1 M Tris-HCl, pH 8.0) for 1 hour at 37°C, to remove chondroitin sulphate from the PGs. Sections were then treated for 20 min with 1% hydrogen peroxide in methanol and subsequently for 5 min with 0.1% (v/v) Triton X-100 in phosphate-buffered saline. After incubation for 20 min with 1.5% (v/v) normal goat serum, sections were incubated with affinity-purified anti-VDIPEN IgG overnight at 4°C. These antibodies were kindly provided by Irwin Singer and Ellen Bayne (Merck Research Laboratories, Rahway, NJ, USA) and have been extensively characterised previously [26,27]. In addition, sections were incubated with biotinylated goat anti-rabbit IgG and binding was detected by avidin-peroxidase staining (Elite kit; Vector Labs, Inc., Burlingame, CA, USA). Development of the peroxidase product was performed by nickel enhancement, and counterstaining was performed with Orange G (2%) for 5 min.
Quantitative RT-PCR of synovium and cartilage
Synovial biopsies were taken from tissue adjacent to the suprapatellar ligament with a biopsy punch (diameter 3 mm). The cartilage layers from patellae and tibiae were isolated after decalcification with 5% EDTA for 4 hours at 4°C. Subsequently, patellae and tibiae were washed in 0.9% NaCl and the cartilage layer was carefully removed from the underlying bone with forceps and a dissection microscope. RNA was isolated with 1 ml of TRIzol reagent (Life Technologies, Breda, The Netherlands). Specific mRNA levels for various MMPs (MMP-2, MMP-3, MMP-9, MMP-12 and MMP-13), their inhibitors (TIMP-1, TIMP-2, TIMP-3 and TIMP-4) and FcγR (FcγRI, FcγRII and FcγRIII) were quantified with the ABI/PRISM 7000 Sequence Detection System (ABI/PE, Foster City, CA, USA). In brief, 1 μg of synovial RNA was used for RT-PCR. mRNA was reverse-transcribed to cDNA with the use of oligo(dT) primers; 1/20 of the cDNA was used in one PCR amplification. PCR was performed in SYBR Green Master Mix by using the following amplification protocol: 2 min at 50°C followed by 40 cycles of 15 s at 95°C and 1 min at 60°C, with data collection in the last 30 s. Message for murine glyceraldehyde-3-phosphate dehydrogenase, MMPs, MMP inhibitors and FcγR was amplified with specific primers (Biolegio, Malden, The Netherlands) for these molecules at a final concentration of 300 nM. Relative quantification of the PCR signals was performed by comparing the cycle threshold value (Ct) of the various molecules in the different samples after correction of the glyceraldehyde-3-phosphate dehydrogenase content for each individual sample to rule out confounding by variation of the RNA purification and reverse transcriptase steps.
Results
During ICA, joint inflammation is downregulated by oxygen radicals, which is compensated for by IFN-γ
To investigate the effect of the NADPH-oxidase-driven production of oxygen radicals on joint inflammation, ICA was induced in knee joints of p47phox-/- mice and their wild-type (WT) controls. Total knee joint sections were stained with haematoxylin and eosin and the numbers of inflammatory cells present within the synovium (infiltrate) and joint cavity (exudate) were determined by using an arbitrary scale of 0 to 3. At day 3 after the induction of ICA, joint inflammation was significantly higher in p47phox-/- mice than in their WT controls (Fig. 1), indicating that oxygen radicals inhibit IC-mediated joint inflammation.
In addition, the effect of IFN-γ on joint inflammation was investigated by injecting an adenoviral IFN-γ construct into the knee joints of p47phox-/- mice and their WT controls, one day before ICA induction. Previous studies had shown that IFN-γ does not elevate joint inflammation during ICA, whereas cartilage destruction is significantly enhanced [13]. The latter is strongly correlated with upregulation of FcγRI. We find that the amount of inflammatory mass was comparable in IFN-γ-stimulated knee joints of p47phox-/- mice and in their WT controls at both day 3 and day 7 after ICA induction (Fig. 2), suggesting that IFN-γ compensates for the aggravating effect of oxygen radicals on joint inflammation.
Apart from the amount, the cell type of the infiltrated cells might also be different. To determine the contribution of macrophages, the dominant cell type involved in cartilage destruction within this model, sections were stained with antibodies directed against F4/80. At day 7 after IFN-γ-stimulated ICA induction, high but comparable amounts of F4/80-positive macrophages were detected in both p47phox-/- mice and their WT controls. Between 70 and 80% of the inflammatory cells, in both infiltrate and exudate, showed clear F4/80 staining (Fig. 3a). Moreover, large amounts of F4/80-positive macrophages were attached to cartilage surfaces at sites where erosion was detected (Fig. 3b).
Oxygen radicals are not involved in mediating early PG depletion
During ICA, mild cartilage destruction starts with the release of PGs from the surface of the cartilaginous layers. To investigate this early cartilage destruction, which is mainly mediated by aggrecanases, total knee joint sections were stained with safranin-O. Loss of red staining (a measure of PG loss), was scored on an arbitrary scale from 0 to 3 in various cartilage layers of the knee joint (medial and lateral femur, tibia and patella). At day 3 after IFN-γ-stimulated arthritis induction, PG loss in arthritic WT controls varied from 1 in the patella to 3 in the lateral and medial femur. At day 7 after ICA induction, nearly maximal PG loss was found in all cartilage layers investigated. Comparable PG depletion was found in arthritic p47phox-/- knee joints at both day 3 and day 7 after arthritis induction (compare Fig. 4a with Fig. 4b), suggesting that NADPH-oxidase-driven oxygen radicals do not alter the aggrecanase activity responsible for PG loss. Arthritic knee joints not previously stimulated by IFN-γ also showed maximal PG loss that was not different between the two strains (data not shown).
Oxygen radicals aggravate MMP-mediated cartilage destruction during IFN-γ-accelerated ICA
Because PG loss was not different between p47phox-/- mice and WT controls, we additionally investigated the more severe cartilage matrix destruction mediated by MMPs. For this purpose, the amount of MMP-specific neoepitope VDIPEN expressed within various cartilage layers within the knee joint was determined by immunostaining with specific anti-VDIPEN antibodies. A progressive amount of VDIPEN staining was observed at day 7 when compared with day 3 in cartilage layers of IFN-γ-stimulated arthritic knee joints but not in p47phox-/- mice (compare Fig. 5a with Fig. 5b).
In WT controls, the amount of VDIPEN staining varied from 5% in the patella to 55% in the lateral femur 3 days after arthritis induction. In p47phox-/- mice, VDIPEN staining in various cartilage layers was comparable to WT controls at that time point (Fig. 5a). At day 7 after arthritis induction, VDIPEN staining varied between 10 and 80% in WT controls. Interestingly, in knee joints of arthritic p47phox-/- mice, VDIPEN staining was significantly lower in the lateral femur, medial femur and lateral tibia (50%, 60% and 50% reduction, respectively) (Fig. 5b, and compare Fig. 5c with Fig. 5d) and values were not different from those found at day 3. These results indicate that oxygen radicals aggravate MMP-mediated cartilage damage during IC-mediated arthritis.
Oxygen radicals downregulate MMP mRNA levels within cartilage layers and inflamed synovium during ICA
One important source of MMPs involved in cartilage destruction might be derived from activated chondrocytes. To investigate whether oxygen radicals alter the expression of MMPs within the cartilage of an inflamed knee joint, patellar and tibial cartilage layers were isolated at day 3 and day 7 after arthritis induction, and mRNA levels of various MMPs (MMP-2, MMP-3, MMP-9, MMP-12 and MMP-13) and their inhibitors (TIMP-1, TIMP-2, TIMP-3 and TIMP-4) were determined by quantitative RT-PCR. IFN-γ-stimulated arthritis induced a marked increase in MMP-3, MMP-12 and MMP-13 mRNA levels in the cartilage layers of both WT controls and p47phox-/- mice, in comparison with the cartilage of naive knee joints (Δ Ct ranging from 3 to 9) (Fig. 6a,b). However, MMP-12 and MMP-13 levels were significantly increased in the cartilage of arthritic p47phox-/- knee joints at day 3 and day 7, respectively, in comparison with WT controls (Fig. 6a,b). TIMP-1 was the only inhibitor moderately expressed in cartilage after the induction of IFN-γ-stimulated arthritis, and no differences were observed between arthritic knee joints of WT controls and those of p47phox-/- mice (Fig. 6c,d).
Another important source of MMPs might be the inflamed synovium. Well-defined synovial specimens were isolated at days 3 and 7 after arthritis induction. At days 3 and 7 after IFN-γ accelerated ICA, MMP mRNA levels were evidently present in the cartilage of WT controls and p47phox-/- mice, when compared with naive knee joints (Fig. 7a,b). Interestingly, at day 7, the cartilage of p47phox-/- mice showed a significant elevation of MMP-3, MMP-9 and MMP-13 in comparison with that of WT controls (Fig. 7b). The expression of TIMP mRNA was determined in WT controls and p47phox-/- mice: TIMP-1 and TIMP-2 were moderately expressed after the induction of arthritis in both groups (Fig. 7c,d). Moreover, the expression of TIMP-1 at day 7 after arthritis onset was significantly higher in p47phox-/- mice than in the WT controls (Fig. 7d). Our data suggest that MMP mRNA levels are higher, and are certainly not decreased, in the cartilage layers and synovium of inflamed knee joints of p47phox-/- mice.
Oxygen radicals upregulate FcγRI and downregulate FcγRII and FcγRIII during IFN-γ-stimulated ICA
In previous studies we found that activating FcγR (mainly FcγRI), predominantly expressed by haemopoietic cells present in the synovium, are important in the activation step of latent MMPs [10,11]. To investigate further whether oxygen radicals are involved in the regulation of FcγR, mRNA levels of the three FcγR classes were determined in synovia of day 7 IFN-γ-stimulated ICA. In WT mice, FcγRI and FcγRII were still upregulated 16 and 4 times, respectively whereas the expression of FcγRIII was four times lower than at zero time. In p47phox-/- mice, FcγRI was downregulated (four cycli), whereas FcγRII and FcγRIII were both strongly upregulated (difference 64 times from zero time in both strains; Fig. 8).
Oxygen radicals determine chondrocyte death during IFN-γ-driven IC-mediated arthritis
Apart from MMP-mediated cartilage destruction, chondrocyte death is an important parameter of severe cartilage destruction. In earlier studies we found that during IFN-γ-accelerated ICA, chondrocyte death was completely dependent on FcγRI. Because the binding of ICs to FcγRI results in the substantial production of oxygen radicals [14] and subsequently leads to a significant upregulation of this receptor, we further investigated whether NADPH-oxidase-driven oxygen radical production is indeed responsible for chondrocyte death in this model. The numbers of empty lacunae (resulting from chondrocyte death) present within various cartilage layers of the knee joint were determined and expressed as a percentage of the total numbers of chondrocytes present. Without IFN-γ overexpression, ICA did not induce chondrocyte death in normal WT and p47phox-/- knee joints (Fig. 9a). In contrast, when AdIFN-γ was injected before arthritis induction, chondrocyte death increased tremendously and varied between 40 and 60% in the lateral and medial femur and between 20 and 40% in the lateral and medial tibia (Fig. 9b) at day 3 in WT mice. At day 7, chondrocyte death was even higher (between 60 and 70% in the femur and between 20 and 70% in the tibia; Fig. 9c,d). Interestingly, in arthritic knee joints of IFN-γ-stimulated p47phox-/- mice, although joint inflammation was comparable to that found in WT mice, chondrocyte death was completely absent at day 3 and was only very low at day 7 (between 2 and 5% in the tibia and between 5 and 8% in the femur; Fig. 9b,c,e). Chondrocyte death does lead to cartilage erosion. However, at day 7 after IFN-γ-stimulated arthritis induction, erosion was still mild in knee joint cartilage layers of arthritic WT mice. Erosion pits were found only in the superficial layers of the medial and lateral tibia. Clear attachment of macrophages to the cartilage surface was observed. Cartilage layers in the knee joints of arthritic p47phox-/- mice showed similar attachment of macrophages and only mild erosion, whereas no chondrocyte death was observed (Fig. 9e).
Discussion
In the present study we found that in the absence of NADPH-oxidase-generated oxygen radicals, IC-mediated joint inflammation was significantly enhanced in p47phox-/- mice. This might be due to a disruption in IC clearance because the removal of ICs from the joint determines the severity of arthritis [28]. This is in line with a previous study in which it was shown that oxygen radicals are crucial in the clearance of foreign particles such as cell walls of microorganisms [29]. Previously we found that injecting zymosan directly into the knee joint of p47phox-/- mice caused a strongly elevated joint inflammation due to retarded clearance and resulted in prominent granuloma formation within the synovia of these mice [30]. In the present study we found that IFN-γ overexpression in the knee joint of p47phox-/- mice before ICA induction prevented the increase in joint inflammation, and no granuloma formation was found. IFN-γ is a potent upregulator of receptors involved in phagocytosis, such as FcγR and complement receptors, and might lead to an efficient removal of the small amount of ICs responsible for continuing arthritis within the knee joints of p47phox-/- mice. Macrophages form the dominant cell type within this model and these cells express large quantities of FcγR, largely responsible for IC clearance but also for the activation of the lining cells, driving arthritis [31]. Interestingly, synovial expression of the inhibitory FcγRII, which has been shown to be the dominant FcγR involved in IC clearance [32], was upregulated in the synovium of IFN-γ-stimulated p47phox-/- mice, and because FcγRII does not need oxygen radicals for efficient clearance this might lead to a more efficient IC clearance.
Although the amount of infiltrated macrophages was not different between arthritic p47phox-/- mice and their WT controls, destruction of the cartilage matrix by MMPs was lower in the absence of oxygen radicals. Cytokines such as IL-1 and TNFα activate chondrocyte and synoviocytes to produce MMPs, which are released in an inactive form. These latent enzymes need an activation step to become able to degrade the cartilage matrix. MMP-3 is the crucial MMP involved in the activation of MMP-13, which forms the rate-limiting enzyme in the degradation of the collagen type II matrix, leading to erosion of the cartilage matrix [5]. IFN-γ overexpression strongly increased MMP expression both in cartilage layers and in the synovium. This might be regulated directly by IFN-γ or indirectly in the synovium by the upregulation of FcγR and their subsequent activation by ICs. In the present study we found that inflamed synovia of IFN-γ-stimulated p47phox-/- mice showed a strong upregulation of various MMPs such as MMP-9, MMP-12 and MMP-13, whereas only a minor upregulation of MMP-3 and MMP-12 was found within the cartilage. In the synovium, only TIMP-1 and TIMP-2 were marginally upregulated, whereas in the cartilage no differences in TIMP expression were found. Because MMP-mediated cartilage destruction was lower in arthritic p47phox-/- mice, whereas MMP expression in the synovium and cartilage layers seemed higher, this might indicate that oxygen radicals, apart from inhibiting the gene expression of MMPs, are involved in their activation. Oxygen radicals have previously been shown to activate latent MMPs such as MMP-2 [17]. In the present study we also found that oxygen radicals upregulate FcγRI. Binding of ICs to FcγRI leads to more oxygen radical production [14] and might form an amplification step in the activation of pro-MMPs.
An interesting difference in the contribution of oxygen radicals to MMP-mediated cartilage damage in p47phox-/- mice was found between arthritis induced by zymosan (ZIA) and that by ICs. During IFN-γ-stimulated ICA, oxygen radicals enhance MMP-cartilage damage, whereas during ZIA they inhibit it. An explanation for this discrepancy might be the cell type involved in mediating cartilage destruction. During ZIA, many polymorphonuclear cells (PMNs) infiltrate into the joint. Crucial enzymes released by PMNs are elastase and cathepsin G, which because of their highly positive charge are highly capable of penetrating cartilage and are then able to stimulate pro-MMPs into their active form, to generate VDIPEN neoepitopes [33]. Under normal circumstances elastase activity is inhibited by synovial fluid inhibitors such as α2-macroglobulin, and no VDIPEN staining can be detected within the cartilage layers [4]. However, in the absence of oxygen radicals the number of infiltrated PMNs was strongly increased during ZIA [30] and the amount of elastase might then overrule the inhibiting capacity of the synovial fluid.
In contrast to ZIA, during IFN-γ-stimulated ICA the dominant infiltrating cell is the macrophage, which strongly attaches to the surface of the cartilage. The production of oxygen radicals such as hydrogen peroxide generated after the stimulation of FcγR by ICs [14] and the presence of superoxide dismutase might then be of crucial importance in regulating the activation of pro-MMPs in the cartilage matrix (Fig. 10). Hydrogen peroxide has a relatively long half-life and is able to activate pro-MMPs [17]. Synovial fluid contains large amounts of inhibitors of hydrogen peroxide such as catalase [34]. However, because of the close proximity of the activated macrophage to the cartilage surface, hydrogen peroxide can escape from this inhibitor, which owing to its large size (240 kDa) is not able to penetrate into the cartilage matrix [35].
Another parameter of severe cartilage destruction is chondrocyte death, which was completely abrogated in the absence of NADPH-oxidase-driven oxygen radicals. Chondrocyte death might be mediated by oxygen radicals released by the chondrocyte itself or by the inflamed synovium. Chondrocytes do express NADPH oxidase [36] and cytokines such as IL-1 are potent inducers of oxygen radicals in chondrocytes [37]. The production of intracellular hydrogen peroxide inside the chondrocyte can cause disruption of the mitochondrial membrane, leading to apoptosis [38]. However, earlier studies in our laboratory showed that FcγR activated synovium is of crucial importance in mediating chondrocyte death [39]. During IFN-γ-accelerated ICA, the infiltrated macrophages become activated by ICs, mainly via FcγR. In the mouse knee joint, FcγRI is expressed not by chondrocytes but exclusively by macrophages (and not neutrophils) and becomes strongly upregulated by IFN-γ. Binding of ICs to FcγRI in particular and, to a lesser extent, to FcγRIII leads to the activation of oxygen radical production (Fig. 10). Apart from FcγRI stimulation, IFN-γ itself has been shown to upregulate the p91 and p47 components of the NADPH oxidase and might contribute to the enhanced superoxide generation [40].
p47phox-/- mice might also produce oxygen radicals by pathways other than the NADPH oxidase pathway [23]. However, IFN-γ alone had no effect on chondrocyte death. Moreover, it has been shown that IFN-γ does not upregulate alternative ways of oxygen radical production in p47phox-/- mice [23]. This indicates that chondrocyte death is completely mediated via NADPH oxidase. IFN-γ induces the upregulation of NADPH oxidase components and FcγRI [41]. Stimulation of FcγRI by ICs might also lead to an enormous increase in oxygen radical production, mediating cartilage destruction (Fig. 10). Hydrogen peroxide might again be the most plausible oxygen species mediating chondrocyte death. Hydrogen peroxide can easily penetrate through cell membranes. Previous studies have shown that hydrogen peroxide, when injected into mouse knee joints, was able to induce considerable chondrocyte death, which might be induced by apoptosis [42]. Hydrogen peroxide activates the opening of the mitochondrial permeability transition pore and the release of cytochrome c [43]. In the cytoplasm, cytochrome c, in combination with Apaf-1, activates caspase-9, leading to the activation of caspase-3 and subsequent apoptosis [44].
NADPH oxidase and p47phox phosphorylation is strongly increased in leucocytes derived from synovial fluid of RA patients [45]. Cytokines such as IFN-γ are potent candidates for the upregulation of NADPH oxidase [41]. Moreover, ICs are found in considerable amounts in joints of many RA patients. These ICs might be responsible for a large part of NADPH oxidase activation via FcγRI stimulation, resulting in large quantities of oxygen radicals. The latter might mediate part of the severe cartilage destruction. Because FcγRI-mediated oxygen radical production might have a major function in mediating cartilage destruction during arthritis, this receptor might form a crucial target in combating this crippling disease.
Conclusion
FcγR are central to the regulation of severe cartilage destruction during arthritis mediated by ICs. These ICs bind to FcγR, and the stimulation of activating FcγR, especially on synovial macrophages, leads to the production of as yet unknown products responsible for cartilage destruction. Th1 cytokines such as IFN-γ strongly upregulate FcγR – mainly FcγRI – and its stimulation leads to an enhanced production of oxygen radicals via NADPH oxidase. Using p47-/- mice, which fail to produce oxygen radicals via NADPH oxidase, we have shown that during IFN-γ-stimulated IC-mediated arthritis, oxygen radicals completely determine chondrocyte death and aggravate MMP-mediated cartilage destruction. Blockade of signalling pathways regulating oxygen radical production via FcγR or by neutralising oxygen radicals directly may form new therapeutic methods of preventing severe cartilage destruction.
Abbreviations
IC = immune complex; ICA = immune complex arthritis; IFN-γ = interferon-γ; IL = interleukin; MMP = matrix metalloproteinase; PG = proteoglycan; PMN = polymorphonuclear cells; RA = rheumatoid arthritis; RT-PCR = reverse transcriptase polymerase chain reaction; TIMP = tissue inhibitor of metalloproteinase; WT = wild-type; ZIA = zymosan-induced arthritis.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
PLEMvL conceived of the study, organised its design and coordination and drafted the manuscript. KCAMN participated in coordination, drafting of the manuscript, and performing the in vivo experiments and quantitative RT-PCR. ABB participated in coordination. AS performed the histology. AEMH performed the in vivo experiments. JK manufactured the adenoviral IFN-γ adenoviral construct. FAJVDL participated in drafting the manuscript. SMH manufactured the P47-/- mice. WBVDB participated in drafting the manuscript. All authors read and approved the final manuscript.
Figures and Tables
Figure 1 Knee joints, 3 days after induction of ICA in p47phox-/- mice and their WT controls. The numbers of cells present in the synovium (infiltrate) and in the joint cavity (exudate) were determined on an arbitrary scale from 0 to 3: 0, no cells; 1, few; 2, moderate; 3, maximal. The number of cells was determined by two blinded observers. Data are means ± SD for eight animals. Significance was tested with the Wilcoxon rank test (*P < 0.05).
Figure 2 Inflamed knee joints, various days after IFN-γ-stimulated ICA in p47phox-/- mice and controls. The numbers of cells present in the synovium (infiltrate) and in the joint cavity (exudate) were determined on an arbitrary scale from 0 to 3: 0, no cells; 1, few; 2, moderate; 3, maximal. (a) Numbers of cells; (b) wild-type controls; (c) p47phox-/- mice. The number of cells was determined by two blinded observers. Data are means ± SD for eight animals. Significance was tested with the Wilcoxon rank test (*P < 0.05). Original magnifications × 100. F, femur; P, patella. Note that there is a comparable cell mass in arthritic knee joints of p47phox-/- mice and wild-type controls.
Figure 3 Expression of macrophage marker F4/80 in knee joints of day 7 IFN-γ-stimulated ICA. Note that 70 to 80% of the infiltrated cells within the synovium consist of macrophages (a). Arrowheads in (b) indicate F4/80-positive macrophages attached to the cartilage surface and found in the lacunae of erosion pits. Original magnification × 400. F, femur; JS, joint space.
Figure 4 Proteoglycan loss from various cartilage layers of inflamed knee joints. Loss of red staining was scored in tibia, femur and patella on an arbitrary scale from 0 to 3. Data are expressed as loss of red staining in comparison with control cartilage layers, and are means ± SD for eight mice; they were tested for significance with the Wilcoxon rank test (*P < 0.05). No significant difference in proteoglycan loss was found on day 3 (a) or day 7 (b) between wild-type controls and p47phox-/- mice. LF, lateral femur; LT, lateral tibia; MF, medial femur; MT, medial tibia; P, patella.
Figure 5 VDIPEN expression inday 3 (a) and day 7 (b) inflamed knee joints. Positive VDIPEN staining was determined in various cartilage layers (LF, lateral femur; LT, lateral tibia; MF, medial femur; MT, medial tibia; P, patella) at an original magnification of ×100 by using automated image analysis and was expressed as a percentage of the total cartilage surface. VDIPEN staining was significantly lower at day 7 in lateral and medial femur and lateral tibia of p47phox-/- mice (d) than in wild-type mice on day 7 (c) Data are means ± SD for eight mice.
Figure 6 mRNA levels of various MMPs and TIMPs in cartilage layers derived from inflamed knee joints. Cartilage layers were isolated from patellae and tibia, three and seven days after IFN-γ accelerated arthritis. The cycle threshold value (Ct) of the various MMP and TIMP genes was corrected for glyceraldehyde-3-phosphate dehydrogenase content and t = 0. Note that at day 3, MMP-12 and at day 7 MMP-3 mRNA levels were significantly elevated in p47phox-/- mice when compared to controls (6A and B). TIMP mRNA levels were not altered (6C and D).
Figure 7 mRNA levels of various MMPs and TIMPs in synovia derived from inflamed knee joints. Inflamed synovia were isolated 3 days (a,c) and 7 days (b,d) after IFN-γ-accelerated arthritis. The cycle threshold value (Ct) of the various MMP (a,b) and TIMP (c,d) genes was corrected for glyceraldehyde-3-phosphate dehydrogenase content and for values at zero time. Note that at day 7 significantly elevated levels of MMP-3, MMP-9 and MMP-13 were found in p47phox-/- mice in comparison with controls (b). TIMP-1 and 2 were also somewhat elevated, although to a lesser extent than MMP (d). No significant differences were found at day 3 (a,c).
Figure 8 mRNA levels of various FcγRs in synovia derived from inflamed knee joints. Inflamed synovia were isolated 7 days after IFN-γ-stimulated arthritis. The cycle threshold value (Ct) of the various FcγR genes was corrected for glyceraldehyde-3-phosphate dehydrogenase content and for values at zero time. Note that at day 7 significantly lower levels of FcγRI and elevated levels of FcγRII and FcγRIII were found in p47phox-/- mice than in controls.
Figure 9 Chondrocyte death in cartilage layers of inflamed knee joints of p47phox-/- and wild-type mice). At day 3 after ICA induction (a) and at day 3 (b) and day 7 (c) after IFNγ-stimulated ICA. Chondrocyte death was determined in various cartilage layers of the knee joint (LF, lateral femur; LT, lateral tibia; MF, medial femur; MT, medial tibia; P, patella). Chondrocyte death was expressed as a percentage of empty lacunae. Note that without IFN-γ no chondrocytes were observed. At day 7 after IFNγ-stimulated ICA, chondrocyte death was clearly present in wild-type controls (d), whereas in p47phox-/- mice chondrocyte death was completely absent (e). Original magnifications ×400.
Figure 10 Possible involvement of NADPH oxidase in mediating cartilage destruction during ICA. Immune complexes (ICs) bind to FcγRI and FcγRIII expressed on macrophages, and cause the activation of NADPH oxidase. IFN-γ strongly elevates the expression of mainly FcγRI and NADPH oxidase. Oxygen radical production causes chondrocyte (CH) death and the activation of matrix metalloproteinases (MMP), leading to matrix degradation in the cartilage expressed as VDIPEN epitopes.
==== Refs
Yanni G Whelan A Feighery C Bresnihan B Synovial tissue macrophages and joint erosion in rheumatoid arthritis Ann Rheum Dis 1994 53 39 44 8311554
Mulherin D Fitzgerald O Bresnihan B Synovial tissue macrophage populations and articular damage in rheumatoid arthritis Arthritis Rheum 1996 39 115 124 8546720
Cooke TD Richer S Hurd E Jasin HE Localization of antigen-antibody complexes in intra-articular collagenous tissues Ann NY Acad Sci 1975 256 10 24 126032
van Meurs JB van Lent PL Holthuysen AE Singer II Bayne EK van den Berg WB Kinetics of aggrecanase- and metalloproteinase-induced neoepitopes in various stages of cartilage destruction in murine arthritis Arthritis Rheum 1999 42 1128 1139 10366105 10.1002/1529-0131(199906)42:6<1128::AID-ANR9>3.0.CO;2-2
van Meurs J van Lent P Stoop R Holthuysen A Singer I Bayne E Mudgett JS Poole R Billinghurst C van der Kraan P Cleavage of aggrecan at the Asn341-Phe342 site coincides with the initiation of collagen damage in murine antigen-induced arthritis: a pivotal role for stromelysin 1 in matrix metalloproteinase activity Arthritis Rheum 1999 42 2074 2084 10524678 10.1002/1529-0131(199910)42:10<2074::AID-ANR7>3.0.CO;2-5
Van Lent PL Holthuysen AE Van Den Bersselaar LA Van Rooijen N Joosten LA Van De Loo FA Van De Putte LB Van Den Berg WB Phagocytic lining cells determine local expression of inflammation in type II collagen-induced arthritis Arthritis Rheum 1996 39 1545 1555 8814067
Ravetch JV Fc receptors: rubor redux Cell 1994 78 553 560 8069908 10.1016/0092-8674(94)90521-5
Ravetch JV Bolland S IgG Fc receptors Annu Rev Immunol 2001 19 275 290 11244038 10.1146/annurev.immunol.19.1.275
Verbeek JS Hazenbos WL Capel PJ Van De Winkel JG The role of FcR in immunity: lessons from gene targeting in mice Res Immunol 1997 148 466 474 9498008 10.1016/S0923-2494(97)82673-9
Fossati-Jimack L Ioan-Fascinay A Reininger L Chicheportiche Y Watanabe N Saito T Hofhuis FM Gessner JE Schiller C Schmidt RE Markedly different pathogenicity of four immunoglobulin G isotype-switch variants of an anti erythrocyte autoantibody is based on their capacity to interact in vivo with the low-affinity Fcγ receptor III J Exp Med 2000 191 1293 1302 10770797 10.1084/jem.191.8.1293
Van Lent PL Nabbe K Blom AB Holthuysen AE Sloetjes A Van De Putte LB Verbeek S Van Den Berg WB Role of activatory Fc gamma RI and Fc gamma RIII and inhibitory Fc gamma RII in inflammation and cartilage destruction during experimental antigen-induced arthritis Am J Pathol 2001 159 2309 2320 11733380
Nabbe KC Blom AB Holthuysen AE Boross P Roth J Verbeek S Van Lent PL Van Den Berg WB Coordinate expression of activating Fcγ receptors I and III and inhibiting Fcγ receptor II in the determination of joint inflammation and cartilage destruction during immune complex-mediated arthritis Arthritis Rheum 2003 48 255 265 12528127 10.1002/art.10721
Nabbe KC van Lent PL Holthuysen AE Kolls JK Verbeek S van den Berg WB FcγRI up-regulation induced by local adenoviral-mediated interferon-γ production aggravates chondrocyte death during immune complex-mediated arthritis Am J Pathol 2003 163 743 752 12875993
Melendez AJ Bruetschy L Floto RA Harnett MM Allen JM Functional coupling of FcγRI to nicotinamide adenine dinucleotide phosphate (reduced form) oxidative burst and immune complex trafficking requires the activation of phospholipase D1 Blood 2001 98 3421 3428 11719383 10.1182/blood.V98.12.3421
Valencia A Moran J Reactive oxygen species induce different cell death mechanisms in cultured neurons Free Radic Biol Med 2004 36 1112 1124 15082065 10.1016/j.freeradbiomed.2004.02.013
Nair VD Yuen T Olanow CW Sealfon SC Early single bifurcation of pro- and anti-apoptotic states during oxidative stress J Biol Chem 2004 279 27494 27501 15078887 10.1074/jbc.M312135200
Yoon SO Park SJ Yoon SY Yun CH Chung AS Sustained production of H2O2 activates pro-matrix metalloproteinase-2 through receptor tyrosine inases/phosphatidylinositol 3-kinase/NF-κB pathway J Biol Chem 2002 277 30271 30282 12058032 10.1074/jbc.M202647200
Deleo FR Allen LA Apicella M Nauseef WM NADPH oxidase activation and assembly during phagocytosis J Immunol 1999 163 6732 6740 10586071
Park HS Kim IS Park JW Phosphorylation induces conformational changes in the leukocyte NADPH oxidase subunit p47phox Biochem Biophys Res Commun 1999 259 38 42 10334912 10.1006/bbrc.1999.0721
Vignais PV The superoxide-generating NADPH oxidase:structural aspects and activation mechanism Cell Mol Life Sci 2002 59 1428 1459 12440767 10.1007/s00018-002-8520-9
Woolhiser MR Okayama Y Gilfillan AM Metcalfe DD IgG-dependent activation of human mast cells following up-regulation of FcγRI by IFN-γ Eur J Immunol 2001 31 3298 3307 11745347 10.1002/1521-4141(200111)31:11<3298::AID-IMMU3298>3.0.CO;2-U
Melendez AJ Harnett MM Allen JM Crosstalk between ARF6 and protein kinase Calpha in FcγRI-mediated activation of phospholipase D1 Curr Biol 2001 5 11 869 874 10.1016/S0960-9822(01)00260-3
Jackson SH Gallin JI Holland SM The p47phox mouse knock-out model of chronic granulomatous disease J Exp Med 1995 182 751 758 7650482 10.1084/jem.182.3.751
Lei D Lancaster JR JrJoshi MS Nelson S Stoltz D Bagby GJ Odom G Shellito JE Kolls JK Activation of alveolar macrophages and lung host defenses using transfer of the interferon-gamma gene Am J Physiol 1997 272 L852 L859 9176248
Van Meurs JB Van Lent PL Joosten LA Van der Kraan PM Van den Berg WB Quantification of mRNA levels in joint capsule and articular cartilage of the murine knee joint by RT-PCR: kinetics of stromelysin and IL-1 mRNA levels during arthritis Rheumatol Int 1997 16 197 205 9032819 10.1007/BF01330296
Singer II Kawka DW Bayne EK Donatelli SA Weidner JR Williams HR Ayala JM Mumford RA Lark MW Glant TT VDIPEN, a metalloproteinase-generated neoepitope, is induced and immunolocalized in articular cartilage during inflammatory arthritis J Clin Invest 1995 95 2178 2186 7537757
Singer II Scott S Kawka DW Bayne EK Weidner JR Williams HR Mumford RA Lark MW McDonell J Christen AJ Aggrecanase and metalloproteinase-specific aggrecan neo-epitopes are induced in the articular cartilage of mice with collagen II-induced arthritis Osteoarth Cartil 1997 5 407 418
Wipke BT Wang Z Nagengast W Reichert DE Allen PM Staging the initiation of autoantibody-induced arthritis: a critical role for immune complexes J Immunol 2004 172 7694 7702 15187152
Mardiney M 3rdJackson SH Spratt SK Li F Holland SM Malech HL Enhanced host defense after gene transfer in the murine p47phox-deficient model of chronic granulomatous disease Blood 1997 89 2268 2275 9116268
Van de Loo FA Bennink MB Arntz OJ Smeets RL Lubberts E Joosten LA van Lent PL Coenen-de Roo CJ Cuzzocrea S Segal BH Deficiency of NADPH oxidase components p47phox and gp91phox caused granulomatous synovitis and increased connective tissue destruction in experimental arthritis models Am J Pathol 2003 163 1525 1537 14507659
Van Lent PL Holthuysen AE Van Rooijen N van de Putte LB van den Berg WB Local removal of phagocytic synovial lining cells by clodronate-liposomes decreases cartilage destruction during collagen type II arthritis Ann Rheum Dis 1998 57 408 413 9797567
van Lent PL Nabbe KC Boross P Blom AB Roth J Holthuysen A Sloetjes A Verbeek S van den Berg W The inhibitory receptor FcγRII reduces joint inflammation and destruction in experimental immune complex-mediated arthritides not only by inhibition of FcγRI/III but also by efficient clearance and endocytosis of immune complexes Am J Pathol 2003 163 1839 1848 14578184
van Meurs J van Lent P Holthuysen A Lambrou D Bayne E Singer I van den Berg W Active matrix metalloproteinases are present in cartilage during immune complex-mediated arthritis: a pivotal role for stromelysin-1 in cartilage destruction J Immunol 1999 163 5633 5639 10553093
Biemond P Swaak AJ Koster JF Protective factors against oxygen free radicals and hydrogen peroxide in rheumatoid arthritis synovial fluid Arthritis Rheum 1984 27 760 765 6743361
Van Lent PL van den Berg WB Schalkwijk J van de Putte LB van den Bersselaar L The impact of protein size and charge on its retention in arthritic cartilage J Rheumatol 1987 14 798 805 2822925
Moulton PJ Goldring MB Hancock JT NADPH oxidase of chondrocytes contains an isoform of the gp91phox subunit Biochem J 1998 329 449 451 9445369
Lo YY Conquer JA Grinstein S Cruz TF Interleukin-1 beta induction of c-fos and collagenase expression in articular chondrocytes: involvement of reactive oxygen species J Cell Biochem 1998 69 19 29 9513043 10.1002/(SICI)1097-4644(19980401)69:1<19::AID-JCB3>3.0.CO;2-Y
Tada-Oikawa S Hiraku Y Kawanishi M Kawanishi S Mechanism for generation of hydrogen peroxide and change of mitochondrial membrane potential during rotenone-induced apoptosis Life Sci 2003 73 3277 3288 14561532 10.1016/j.lfs.2003.06.013
van Lent PL van Vuuren AJ Blom AB Holthuysen AE van de Putte LB van de Winkel JG van den Berg WB Role of Fc receptor gamma chain in inflammation and cartilage damage during experimental antigen-induced arthritis Arthritis Rheum 2000 43 740 752 10765918 10.1002/1529-0131(200004)43:4<740::AID-ANR4>3.0.CO;2-0
Yang S Madyastha P Ries W Key LL Characterization of interferon gamma receptors on osteoclasts: effect of interferon gamma on osteoclastic superoxide generation J Cell Biochem 2002 84 645 654 11813269 10.1002/jcb.10074
Gupta JW Kubin M Hartman L Cassatella M Trinchieri G Induction of expression of genes encoding components of the respiratory burst oxidase during differentiation of human myeloid cell lines induced by tumor necrosis factor and gamma-interferon Cancer Res 1992 52 2530 2537 1568222
Schalkwijk J van den Berg WB van de Putte LB Joosten LA Hydrogen peroxide suppresses the proteoglycan synthesis of intact articular cartilage J Rheumatol 1985 12 205 210 3928892
Yang JC Cortopassi GA Induction of the mitochondrial permeability transition causes release of the apoptogenic factor cytochrome c Free Radic Biol Med 1998 24 624 631 9559874 10.1016/S0891-5849(97)00367-5
Budihardjo I Oliver H Lutter M Luo X Wang X Biochemical pathways of caspase activation during apoptosis Annu Rev Cell Dev Biol 1999 15 269 290 10611963 10.1146/annurev.cellbio.15.1.269
El Benna J Hayem G Dang PM Fay M Chollet-Martin S Elbim C Meyer O Gougerot-Pocidalo MA NADPH oxidase priming and p47phox phosphorylation in neutrophils from synovial fluid of patients with rheumatoid arthritis and spondylarthropathy Inflammation 2002 26 273 278 12546136 10.1023/A:1021460517468
| 15987491 | PMC1175041 | CC BY | 2021-01-04 16:02:38 | no | Arthritis Res Ther. 2005 May 20; 7(4):R885-R895 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1760 | oa_comm |
==== Front
Arthritis Res TherArthritis Research & Therapy1478-63541478-6362BioMed Central London ar17651598749310.1186/ar1765Research ArticleCatabolic stress induces expression of hypoxia-inducible factor (HIF)-1α in articular chondrocytes: involvement of HIF-1α in the pathogenesis of osteoarthritis Yudoh Kazuo [email protected] Hiroshi 1Masuko-Hongo Kayo 1Kato Tomohiro 1Nishioka Kusuki 11 Department of Bioregulation, Institute of Medical Science, St. Marianna University School of Medicine, Kawasaki, Japan2005 27 5 2005 7 4 R904 R914 20 11 2004 16 12 2004 23 4 2005 5 5 2005 Copyright © 2005 Yudoh et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Transcription factor hypoxia-inducible factor (HIF)-1 protein accumulates and activates the transcription of genes that are of fundamental importance for oxygen homeostasis – including genes involved in energy metabolism, angiogenesis, vasomotor control, apoptosis, proliferation, and matrix production – under hypoxic conditions. We speculated that HIF-1α may have an important role in chondrocyte viability as a cell survival factor during the progression of osteoarthritis (OA). The expression of HIF-1α mRNA in human OA cartilage samples was analyzed by real-time PCR. We analyzed whether or not the catabolic factors IL-1β and H2O2 induce the expression of HIF-1α in OA chondrocytes under normoxic and hypoxic conditions (O2 <6%). We investigated the levels of energy generation, cartilage matrix production, and apoptosis induction in HIF-1α-deficient chondrocytes under normoxic and hypoxic conditions. In articular cartilages from human OA patients, the expression of HIF-1α mRNA was higher in the degenerated regions than in the intact regions. Both IL-1β and H2O2 accelerated mRNA and protein levels of HIF-1α in cultured chondrocytes. Inhibitors for phosphatidylinositol 3-kinase and p38 kinase caused a significant decrease in catabolic-factor-induced HIF-1α expression. HIF-1α-deficient chondrocytes did not maintain energy generation and cartilage matrix production under both normoxic and hypoxic conditions. Also, HIF-1α-deficient chondrocytes showed an acceleration of catabolic stress-induced apoptosis in vitro. Our findings in human OA cartilage show that HIF-1α expression in OA cartilage is associated with the progression of articular cartilage degeneration. Catabolic-stresses, IL-1β, and oxidative stress induce the expression of HIF-1α in chondrocytes. Our results suggest an important role of stress-induced HIF-1α in the maintenance of chondrocyte viability in OA articular cartilage.
==== Body
Introduction
The breakdown or absence of oxygen homeostasis is a hallmark of many common diseases, such as cancer, myocardial infarction, and arthritis. In most normal and tumor tissues, adaptation to hypoxic conditions is critical for successful tissue expansion [1,2]. In response to down-regulation of oxygen homeostasis, cells during hypoxic challenge transiently or chronically tolerate lowered oxygen levels by means of adaptive mechanisms [1]. In mitochondrial oxidative phosphorylation, oxygen is the terminal electron acceptor during ATP production. Several enzymatic reactions require oxygen as a substrate [3,4]. Responses to hypoxia include a metabolic shift to anaerobic glycolysis as well as the initiation of neoangiogenesis via the expression of angiogenic factors to increase the opportunity for oxygen to reach the tissue [1-5]. Oxygen homeostasis and its down-regulation are involved in the pathogenesis of common diseases [3].
It is well known that the transcription factor hypoxia-inducible factor 1 (HIF-1) appears to be one of the major regulators of the hypoxic response [3,6]. HIF-1 controls hypoxic expression of erythropoietin, as well as the expression of genes with metabolic functions such as glucose transport and metabolism, and angiogenic factors such as vascular endothelial cell growth factor (VEGF) [6-8]. HIF-1 is a heterodimer of the PAS subfamily of basic-helix-loop-helix transcription factors, and it consists of the subunit HIF-1α (120 kDa), produced in response to hypoxia, and the constitutively expressed HIF-1α (91 to 94 kDa) subunit [9]. HIF-1 protein accumulates and activates the transcription of genes that are of fundamental importance for oxygen homeostasis, including genes involved in energy metabolism, angiogenesis, vasomotor control, apoptosis, proliferation and matrix production, under hypoxic conditions [6,8,9].
Articular cartilage is an avascular tissue lacking a capillary network, in which oxygen is limited due to its delivery via diffusion through the synovial fluid. It is well known that there is a physiological gradient of oxygenation within articular cartilage [10-12]. It has been reported that the partial pressure of O2 in synovial fluid in joints affected by osteoarthritis (OA) is between 40 and 85 mmHg, corresponding to an oxygen concentration of approximately 6 to 11% [13]. Since O2 must enter from the cartilage surface, the concentration of oxygen is approximately 6% at the surface zone of the articular tissue and less than 1% in the deep zone. We histologically examined the oxygen gradation in articular cartilage tissue by immunofluorescence staining with a specific probe. We performed the analysis in human articular cartilage tissue in patients undergoing arthroplastic knee surgery. The levels of immunostaining revealed an O2 tension (approximately 3 to 8%) at the surface of the cartilage similar to that in positive control tumor tissues with already known O2 tension. There is a general consensus that articular chondrocytes are adapted to hypoxic conditions. Since HIF-1α expression is associated with low O2, this factor may play a role in chondrocyte survival and the maintenance of fundamental homeostasis in the normally hypoxic articular cartilage. In addition, degeneration of articular cartilage may directly influence the chondrocyte microenvironment, especially cellular adaptation to hypoxic conditions, in articular cartilage. Even a slight change may affect the adaptative hypoxic conditions of chondrocytes, resulting in alteration of the cellular microenvironment that is involved in the maintenance of articular cartilage. Indeed, more recently it has been demonstrated that HIF-1α is expressed in OA articular cartilage [14]. However, the exact role of this factor in the pathogenesis of OA remains unclear.
We postulated that HIF-1α could play an important role as a survival factor protecting tissue against catabolic changes during the progression of OA. Our data show here for the first time a correlation between the levels of expression of HIF-1α and degeneration of articular cartilage in patients with OA. To clarify the role of HIF-1α in the pathogenesis of OA, we investigated whether or not hypoxia and catabolic factors (IL-1β and H2O2) affected the expression of HIF-1α, energy generation, cartilage matrix production, and apoptosis in OA chondrocytes.
We also report evidence for the action of HIF-1α as a chondrocyte survival factor in OA.
Materials and methods
Preparation of human articular cartilage samples
Donor OA cartilage samples were obtained from knee joints of OA patients undergoing arthroplastic knee surgery (seven OA patients) after obtaining the patients' informed consent. The characteristics of patients are summarized in Table 1. Each sample was cut and divided into two pieces: one was used for histological evaluation and the other was stored at -30°C for later analysis by real-time PCR analysis.
Each cartilage sample was evaluated histologically and macroscopically for the degree of degeneration according to the scales of Mankin and colleagues and of Collins [15,16]. Articular cartilage samples with subchondral bones were fixed for 2 days in 4% paraformaldehyde solution and then decalcified in 4% paraformaldehyde containing 0.85% sodium chloride and 10% acetic acid. Tissues were dehydrated in a series of ethanol solutions and infiltrated with xylene and before being embedded in paraffin and cut into 6-μm sections. Sections were deparaffinized through sequential immersion in xylene and a graded series of ethanol solutions in accordance with conventional procedures. Sections were also stained with safranin O-fast green to determine the loss of proteoglycans [17].
Chondrocyte isolation and culture
Human articular cartilage samples were obtained from knee joints during arthroplastic surgery for OA (n = 7, one male, six females, 61, 62, 64, 66, 67, 68, 72 years old) after obtaining the patients' informed consent. Cartilage tissues were cut into small pieces, washed in PBS, and digested in Dulbecco's modified Eagle's medium (DMEM; Sigma, St. Louis, MO) containing 1.5 mg/ml collagenase B (Sigma). Digestion was carried out at 37°C overnight on a shaking platform. Cells were centrifuged, washed with PBS, and plated with fresh DMEM. Chondrocytes were cultured in DMEM supplemented with 10% heat-inactivated fetal calf serum, 2 mM L-glutamine, 25 mM HEPES (2-[4-(2-hydroxyethyl)-1-piperazinyl] ethanesulfonic acid), and 100 units/ml penicillin and streptomycin at 37°C in a humidified 5% CO2 atmosphere [18].
Chondrocyte culture under hypoxic conditions
Human chondrocytes were dispensed into a 10-cm culture plate. The plates were placed in a sealed hypoxia chamber (Billups-Rothenberg, Del Mar, CA, USA) equilibrated with a humidified 5% CO2 atmosphere or with certified gas containing 1% O2, 5% CO2, and 94% N2 [19,20]. In this hypoxia chamber system, approximately 5 to 6% O2 tension was observed after 15 min of gas flow (20 l/min). The O2 tension in the culture medium was monitored with an oxygen meter (Fuso Rekaseihin Ltd, Tokyo, Japan) as described by the manufacturer. We monitored the O2 tension with an oxygen meter to maintain the concentration of approximately 6%. When so indicated, recombinant human IL-1β (10 ng/ml; Sigma) or H2O2 (10.0 μM; Wako Pure Industries, Tokyo, Japan) was added, and the cells were incubated under normoxic or hypoxic culture conditions at 37°C. As a positive control, COCl2 (150 μM; Sigma), a chemical inducer of HIF-1, was added to the cells during the incubation time in normoxia or hypoxia [20].
In other experiments, human chondrocytes were cultured in the presence or absence of a phosphatidylinositol 3-kinase (PI3K) inhibitor LY294002 (Sigma), a p38 mitogen-activated protein kinase (MAPK) inhibitor SB203580 (Sigma), and extracellular signal-regulated kinase (ERK 1/2) inhibitor PD98059 (Wako).
Immunoblotting
Cells were lysed in boiling sample buffer as suggested by the manufacturer (Sigma). Samples were then homogenized by repeated aspiration through a 26-gauge needle. Cellular proteins were resolved by SDS-PAGE (12.5% acrylamide) and were transferred to nitrocellulose membranes. Blots were incubated for 2 hours in Tris-buffered saline/Tween 20 (TBST; 10 mM Tris/HCL, pH 8.0, 150 mM NaCl, and 0.2% Tween 20) containing 2% powdered skimmed milk and 1% bovine serum albumin. After three washes with TBST, membranes were incubated for 2 hours with the primary antibody to HIF-1α (diluted 1000-fold in TBST) (Santa Cruz Biotechnology Inc, Santa Cruz, CA) and for 1 hour with horseradish-peroxidase-conjugated goat antimouse IgG (diluted 1000-fold) (DAKO, Glostrup, Denmark). Bound antibodies were detected using an ECL detection kit (Amersham Bioscience KK, Tokyo, Japan). Densitometry of the signal bands was analyzed with Image Gauge version 4.0 (FUJI Photo Film, Tokyo, Japan).
Proteoglycan production in chondrocytes
Chondrocyte activity was measured by the production of glycosaminoglycan (GAG) from cultured chondrocytes. Chondrocytes were cultured under either normoxic or hypoxic conditions using the sealed hypoxia chamber. After 24 hours of incubation, we collected the cells and supernatant. The amount of GAG in the supernatant was measured by using a spectrophotometric assay with dimethylmethylene blue (Aldrich Chemical, Milwaukee, WI, USA) measured at 540 nm using shark chondroitin sulfate (Sigma) as a standard [19].
Measurement of lactic acid in cultured chondrocytes
Supernatants from chondrocyte cultures were collected after 24 hours under normoxic or hypoxic conditions. Lactic acid was determined by a colorimetric assay (Sigma) at 540 nm in accordance with the manufacturer's instructions. Lactic acid levels were normalized to total protein content as measured by the Bradford assay (Bio-Rad, Hercules, CA, USA) [21].
ATP levels in cultured chondrocytes
Chondrocytes were collected after a 24-hour incubation under normoxic or hypoxic conditions. The ATP Bioluminescence assay kit CLS II (Roche, Heidelberg, Germany) was used. The assay is based on the light-emitting oxidation of luciferin by luciferase in the presence of extremely low levels of ATP. After collecting the chondrocytes by scraping, cells were centrifuged for 10 min at 500 × g in the cold. Chondrocytes pellets were extracted by boiling 100 mM Tris (tris(hydroxymethyl)aminomethane) buffer containing 4 mM EDTA (ethylenediaminetetraacetic acid) for 2 min in order to inactivate NTPases. Cell remnants were removed by centrifugation at 1000 × g. Supernatants were removed and placed on ice. Determination of free ATP was as outlined in the manufacturer's protocol. Light emission was measured at 562 nm using a luminometer. ATP levels were normalized to protein content as measured by the Bradford assay (Bio-Rad) [19].
RT-PCR
Total RNA was extracted from articular cartilage by acid guanidine–phenol–chloroform extraction using ISOGEN ® (Nippon Gene Inc, Tokyo, Japan). First-strand complementary DNA (cDNA) was synthesized with Superscript II reverse transcriptase. PCR amplification was performed using specific primers (Table 2). The PCR products were analysed by electrophoresis in 2% agarose gels stained with ethidium bromide, and bands were visualized and photographed under ultraviolet excitation.
Real-time PCR
For PCR analyses, cDNA from triplicate dishes from four independent experiments (24 hours of hypoxia or normoxia) were diluted to a final concentration of 10 ng/ μl. Quantitative real-time RT-PCR was performed with a TaqMan Universal Mastermix (Biosystems Inc, Foster City, CA). cDNA (50 ng) was used as template to determine the relative amounts of mRNA by real-time PCR (ABI 7700 sequence detection system) using specific primers and probes (Table 2). The reaction was conducted as follows: 95°C for 4 min, and 40 cycles of 15s at 95°C and 1 min at 60°C (21). To standardize mRNA levels, we amplified 18S rRNA as an internal control and calculated using Microsoft Excel.
Antisense oligonucleotide treatment of chodrocytes
HIF-1α depletion in chondrocytes was accomplished by using antisense oligonucleotide (ODN) loading using phosphorothioate derivatives of antisense (5'-GCCGGCGCCCTCCAT-3') or control sense (5'-ATGGAGGGCGCCGGC-3') oligonucleotides. Antisense HIF-1α ODN and control ODN were designed and synthesized by BIOGNOSTIK (Göttingen, Germany). Scrambled oligonucleotide was used as control. Chondrocytes were washed in serum-free medium and then in medium containing 20 mg/ml transfection reagent (Qiagen Inc, Valencia, CA, USA) with 2 μM HIF-1α antisense or control ODN. Cells were incubated for 4 hours at 37°C and then replaced with medium containing growth factors. The cellular uptake efficiency was monitored by fluorescein-isothiocyanate-labeled ODN (transfection efficiency approximately 60 to 70% after 4 hours of treatment). The transfection efficiency detected by fluorescein-isothiocyanate-labeled ODN was maintained after a further 24 hours of incubation. Treated cells were cultured in hypoxic or normoxic conditions for the indicated periods of time (24 hours) in each experiment. HIF-1α mRNA was quantified by RT-PCR and western blotting analysis as described above. Data were analyzed for four independent experiments.
Apoptosis
Human subconfluent chondrocytes were cultured in the presence of 10 ng/ml IL-1β for 24 hours under the normoxic or hypoxic conditions described above. Cellular apoptosis was detected using the Apoptosis detection kit (TdT in situ apoptosis detection kit: R&D systems Inc., MN, USA) in chondrocyte cell cultures in accordance with the manufacturer's protocol. The kit was used to identify apoptotic cells by detecting DNA fragmentation through a combination of enzymology and immunohistochemistry techniques. Biotinylated nucleotides are incorporated into the 3'-OH ends of the DNA fragments by terminal deoxynucleotidyl transferase (TdT). Cells containing fragmented nuclear chromatin characteristic of apoptosis exhibit a brown nuclear staining. Apoptosis was assessed by measuring the percentage of apoptotic nuclei in each sample [22,23].
Statistical analysis
Results were expressed as means ± standard deviations. Data were analyzed by a nonparametric statistical analysis. An analysis resulting in value of P < 0.05 was considered statistically significant.
Results
HIF-1α mRNA expression in articular cartilage from patients with OA
To clarify the expression of HIF-1α mRNA in human OA cartilage, the real-time PCR analysis for HIF-1α was performed with donor-matched pairs of intact and degenerated articular cartilage isolated from the same OA sample. Fig. 1a shows a representative safranin-O staining in the degenerated region and intact region of articular cartilage from OA patients. The levels of HIF-1α mRNA in all seven donor articular cartilage samples were higher in the degenerated regions than in the intact regions (Fig. 1b).
Catabolic factors induce the expression of HIF-1α in human articular cartilage
To clarify the effect of catabolic factors on HIF-1α expression in human articular cartilage, the quantitative real-time PCR and western blotting analysis were performed under normoxic and hypoxic culture conditions. In normoxic culture conditions, mRNA levels of HIF-1α were observed in cultured chondrocytes, whereas HIF-1α protein was undetected regardless of stimulation of IL-1β and H2O2 (Fig. 2). Under hypoxic culture conditions, both HIF-1α mRNA and protein were detected in cultured chondrocytes (Figs 2, 3). The expression of HIF-1α was significantly accelerated by the chondrocyte catabolic factors IL-1β and H2O2 (Figs 2, 3). Under hypoxic conditions, the inhibitors of PI3K and p38 kinase caused a significant decrease in the catabolic-factor-induced HIF-1α expression (Fig. 3a, b). Data from four independent experiments were analyzed.
Role of HIF-1α in free ATP production in human articular chondrocytes
To study the role of HIF-1α in chondrocyte energy production, we measured the free ATP levels of cultured chondrocytes under normoxic and hypoxic culture conditions. In control chondrocytes, the levels of free ATP in hypoxia were significantly higher than in normoxia. Under hypoxic conditions, HIF-1α-deficient chondrocytes showed a significant decrease of free ATP in comparison with control ODN-treated chondrocytes (Fig. 4b). In HIF-1α-deficient chondrocytes, free ATP production was approximately 20% of control cells under hypoxic conditions. In contrast, although HIF-1α-deficient chondrocytes showed a slight decrease of energy generation under normoxic conditions, there was no statistically significant difference in energy generation between the three groups (normal chondrocytes, antisense ODN-treated chondrocytes, and chondrocytes treated with the scrambled ODN). Data for four independent experiments were analyzed.
Influence of HIF-1α deficiency on glycolysis in human chondrocytes
As shown in Fig. 4c, d, significant increases of lactic acid (c) and glucose transporter-1 (d) were observed in control ODN-treated chondrocytes under hypoxic culture conditions compared with normoxic culture conditions. In contrast, HIF-1α-deficient chondrocytes showed a complete loss of the induced increases in glycolytic activities even under hypoxic culture conditions.
Under normoxic conditions, HIF-1α-deficient chondrocytes showed a slight decrease of energy generation; however, there was no statistically significant difference in energy generation between control chondrocytes and antisense HIF-1α-treated chondrocytes. Data for four independent experiments were analyzed.
Proteoglycan production from chondrocytes in different oxygen tension
To test whether HIF-1α-mediated alteration affects the potential to produce matrix proteins in chondrocytes, we determined the amount of GAG produced by cultured chondrocytes. We observed large increases in the concentration of GAG in control ODN-treated cultures under hypoxia compared with normoxia. GAG levels were decreased in HIF-1α-deficient chondrocytes under hypoxia to approximately 35% of control levels (Fig. 5). Data for four independent experiments were analyzed.
Apoptosis induction by catabolic factors in HIF-1α-deficient chondrocytes
Under hypoxic conditions, IL-1β-induced apoptosis was increased in hypoxic chondrocytes lacking HIF-1α, to twice that of control ODN-treated chondrocytes (Fig. 6). Even under normoxic conditions, HIF-1α-deficient chondrocytes showed significantly increased levels of apoptosis compared with their control counterparts. Data for four independent experiments were analyzed.
Discussion
Our findings show the potential involvement of HIF-1α expression in the progression of articular cartilage degeneration. In patients with OA, stronger expression of HIF-1α mRNA in chondrocytes was observed in degenerating regions than in intact regions from the same articular cartilage samples. Our findings in human articular cartilage tissues indicate for the first time that expression of HIF-1α mRNA is closely involved in the progression of articular cartilage degeneration.
The HIF-1 complex is ubiquitous, and the presence of this complex in growth-plate chondrocytes has been documented recently [24-26]. Schipani and colleagues reported that in HIF-1α-null mice, hypoxic chondrocytes showed massive cell death in cartilaginous elements such as the chondrosternal junction of the ribs and growth plate, suggesting that HIF-1α is not only crucial for survival of hypoxic chondrocytes, but also modulates the process of chondrocyte proliferation, differentiation, and growth arrest in growth-plate chondrocytes [26]. More recently, Coimbra and colleagues also showed that HIF-1α is expressed in cultured cartilage and chondrocytes under both normoxic and hypoxic conditions [14]. Their findings of HIF-1α expression in chondrocytes are basically consistent with our results from both human and rat OA cartilages. However, from their data, it remained unclear whether HIF-1α expression in chondrocytes is related to the degeneration of articular cartilage in vivo. Indeed, Coimbra and colleagues showed that HIF-1α was expressed not only in normal chondrocytes and cartilage but also in OA chondrocytes, under both hypoxic and normoxic conditions in vitro. In our present study, HIF-1α protein was undetected in chondrocytes under normoxic conditions. It is well known that cellular HIF-1α is not detected in normoxia [27-29]. Under normoxic conditions, the HIF-1α protein undergoes ubiquitination and rapid degeneration in proteasomes [30].
Our data suggest that chondrocyte catabolic factors IL-1β and oxidative stress (oxidative free radicals) may induce the expression of HIF-1α in articular chondrocytes. IL-1 has been shown both to inhibit chondrocyte anabolic activity, including the down-regulation of proteoglycan synthesis, and to stimulate catabolic activity, including production of metalloproteinases [31,32]. IL-1 also stimulates chondrocyte expression of inducible nitric oxide synthesis, iNOS, which results in an increase in NO production [33]. Numerous reports have already demonstrated that oxidative stress acts as a catabolic factor in articular cartilage [34-38]. Articular chondrocytes actively produce endogenous reactive oxygen species, O2- [35], NO [36], -HO [37], and H2O2 [3]). Oxidative damage in cartilage may affect chondrocyte function, resulting in changes in cartilage homeostasis that are relevant to cartilage aging and the development of OA. Our data indicated that in cultured chondrocytes, both mRNA and protein levels of HIF-1α were up-regulated by both IL-1β and H2O2 under hypoxic but not normoxic conditions. These findings suggest that OA-related catabolic stresses (IL-1β, H2O2) induce the expression of HIF-1α in the degenerated articular cartilages as degeneration progresses.
Interestingly, besides hypoxia, many cytokines and growth factors have been shown to be capable of stabilizing and activating HIF-1α under normoxic conditions. Stimulation of cultured synovial fibroblasts with IL-1β and TNFα increases levels of HIF-1α mRNA. Moreover, incubation with IL-1β leads to stabilization of HIF [39]. Our results of catabolic stress-induced expression of HIF-1α in chondrocytes are consistent with these findings. These findings suggest that HIF-1α may, at least in part, have some role in the pathogenesis of inflammatory arthritis even under normoxic conditions, although further studies are needed to clarify this issue. Also, these findings, including our results, provide evidence to support the idea that OA-related catabolic factors (IL-1β etc.) induce HIF-1α during the progression of cartilage degeneration.
In this context, we also studied the signal transduction pathways involved in stress-induced HIF-1α expression in chondrocytes. It has been reported that p38 MAPK, PI3K, and ERK MAPK pathways are responsible for the stress-induced responses in a variety of cells [40-42]. We found that both IL-1β and H2O2 induced a prolonged activation of p38 in chondrocytes (data not shown). Under hypoxic conditions, the inhibitors of PI3K and p38 kinase caused a significant decrease in catabolic-factor-induced HIF-1α expression; this finding supports the idea that PI3K and p38 kinase, but not ERK, activation are required for catabolic stress-induced HIF-1α expression in chondrocytes in hypoxia. Local accumulation of a regulating protein to adapt to hypoxia may be mediated, at least in part, by p38 and PI3K in articular chondrocytes. In addition, we have focused on the redox factor 1 (Ref-1, also known as APE, HAP1, and APEX), a ubiquitous multifactorial protein that is a redox-sensitive regulator of mutifactorial transcription factors, including nuclear factor κB, c-myc gene, activating protein-1, and HIF-1α. Ref-1 may play a critical role in the regulation of endothelial cell fate in response to pathophysiological stimuli such as hypoxia [43]. We have studied the interactions between Ref-1 and HIF-1 activity in OA chondrocytes (data not shown).
Our in vitro data clearly indicate that expression of HIF-1α is responsible for the energy generation and cellular survival of hypoxic chondrocytes. We have shown that HIF-1α activity is essential for regulation of glycolysis, energy generation, synthesis of cartilage matrix proteins, and cell survival in OA chondrocytes under hypoxic conditions. HIF-1α-null chondrocytes did not maintain their viability; energy generation, and matrix production under normoxic and hypoxic conditions. In addition, HIF-1α-null chondrocytes showed accelerated apoptosis induction by IL-1β, suggesting that HIF-1α has an important role in the survival of tissues that lack a functional vasculature, such as articular cartilage.
Articular cartilage adapts to hypoxic conditions, since the cartilage is an avascular tissue. Nutrition and oxygen for articular cartilage are supplied from the synovial fluid. Even a surface zone of articular cartilage has lower oxygen tension (approximately 6%) than synovial fluid (approximately 6 to 15%) [10-13]. There is an oxidative gradient in articular cartilage. Oxygen homeostasis in normal articular cartilage is maintained under hypoxic conditions. During the progression of cartilage degeneration, OA-related catabolic stresses, mechanical and chemical, including IL-1β and oxidative stress, could induce the degradation of the extracellular matrix and decrease chondrocyte viability, resulting in the down-regulation of chondrocyte environment and the further degeneration of articular cartilage. The OA-related changes may also affect oxygen tension and the hypoxic conditions in articular cartilage. Breakdown of oxygen homeostasis in articular cartilage may influence the chondrocytes adapted to hypoxic conditions within the articular cartilage. Although further studies are needed to clarify the exact mechanism of HIF-1α expression, HIF-1α may be expressed in response to change of cellular microenvironment, especially O2 tension, in the tissue. Jewell and colleagues reported that HIF-1α was up-regulated by reoxygenation [44]. Their findings suggest that expression of HIF-1α may be influenced by the cellular microenvironment. When there is deviation from the stable condition in terms of O2 tension, HIF-1α expression may be influenced to maintain the cellular homeostasis. Articular chondrocytes that are well adapted to hypoxia into the tissue may express HIF-1α with the deviating from adaptation to hypoxia. We postulated that HIF-1α is expressed in response to catabolic change in articular cartilage to maintain the cell viability and readaptation to hypoxia. Catabolic stress during the development of OA could influence the chondrocyte adaptation to hypoxia in the tissue.
Our findings of catabolic-factor-induced HIF-1α expression in chondrocytes provide evidence to support the idea that HIF-1α expression is up-regulated in response to catabolic degeneration of articular cartilage. Although genes under the control of HIF-1α have not yet been analyzed in OA, stress-induced HIF-1α may lead to the expression of anti-apoptotic factors or act as a chondroprotective factor to maintain chondrocyte viability in the face of catabolic changes in articular cartilage. Many molecular aspects of HIF-1 signaling should be also studied to further clarify the role of HIF-1 in the pathogenesis of OA.
Conclusion
Taken together, our findings show for the first time that exposure to IL-1β and oxidative stress induce HIF-1α expression in degenerated articular cartilage, which is mediated, at least in part, by activation of PI3K and p38 kinase. Our data suggest a potential for HIF-1α in the maintenance of chondrocyte viability in articular cartilage challenged by progressive articular cartilage degeneration, and they provide new insights into pathogenesis of OA.
Abbreviations
DMEM = Dulbecco's modified Eagle's medium; ERK = extracellular signal-regulated kinase; GAG = glycosaminoglycan; HIF-1α = hypoxia-inducible factor 1α; IL-1 = interleukin-1; MAPK = mitogen-activated protein kinase; MCL = medial collateral ligament; OA = osteoarthritis; ODN = oligonucleotide; PBS = phosphate-buffered saline; PI3K = phosphatidylinositol 3-kinase; TBST = Tris-buffered saline/Tween 20; TdT = terminal deoxynucleotidyl transferase; Tris = tris(hydroxymethyl)aminomethane.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
KY carried out in vitro studies (cell culture), participated in the design of the study, conducted sequence alignment, and drafted the manuscript. HN, K H-M, TK, and KN conceived the study, participated in its design and coordination, and helped to draft the manuscript. All authors read and approved the final manuscript.
Acknowledgements
This study was supported by grants from the Ministry of Education, Culture, Sports, Science and Technology of Japan, the Ministry of Health, Labour and Welfare of Japan, and the Japan Rheumatism Foundation.
Figures and Tables
Figure 1 Levels of HIF-1α mRNA in the articular cartilage from patients with osteoarthritis (OA). (a)Representative x-ray film of knee joint and safranin-O staining for hypoxia-inducible factor 1α(HIF-1α) in the degenerated region and intact region of articular cartilage from a 66-year-old woman with OA. Original magnification of histological sections × 200. (b) The mRNA levels of HIF-1α were higher in the degenerated regions than in the intact regions from the same OA sample.
Figure 2 IL-1β and H2O2 induce the expression of HIF-1α mRNA in human articular chondrocytes. Under normoxic culture conditions, mRNA levels of hypoxia-inducible factor 1α (HIF-1α) were observed in cultured chondrocytes, whereas HIF-1α protein was undetected in the cells. HIF-1α mRNA was accelerated by IL-1β or H2O2 in cultured chondrocytes under hypoxic conditions. Cobalt chloride (CoCl2), chemical inducer of HIF-1, was used for the positive control. *P < 0.05, **P < 0.01 compared with the control. Cont., control.
Figure 3 Catabolic factors induce the expression of HIF-1α protein in human articular cartilage. (a)Hypoxia-inducible factor 1α (HIF-1α) protein was accelerated by IL-1β or H2O2in cultured chondrocytes under hypoxic conditions. (b)Under hypoxic conditions, the inhibitors of PI3K and p38 mitogen-activated protein kinase (MAPK) reduced protein levels of IL-1β-induced HIF-1α expression. Cobalt chloride (CoCl2), chemical inducer of HIF-1, was used for the positive control. *P < 0.05, **P < 0.01 compared with the control. LY294002: phosphatidylinositol 3-kinase inhibitor; SB203580: p38 mitogen-activated protein kinase inhibitor; PD98059: extracellular signal-regulated kinase inhibitor.
Figure 4 Effect of HIF-1α on ATP production and glycolysis in human articular cartilage. (a)Hypoxia-inducible factor 1α (HIF-1α) depletion by antisense oligonucleotide was assessed by RT-PCR and western blotting analyses. HIF-1α mRNA and protein expressions were reduced in antisense HIF-1α-treated chondrocyte populations. Scrambled oligonucleotide was used as control oligonucleotide. Representative data from four independent experiments are shown. (b)In hypoxia, HIF-1α-deficient chondrocytes showed a significant decrease of free ATP in comparison with control oligonucleotide-treated chondrocytes. Statistical differences were calculated using data from four independent experiments. (c, d) The levels of lactate (c) and glucose transporter-1 (Glu-1) (d) were increased in the scrambled ODN-treated chondrocytes under hypoxic culture conditions compared with normoxic culture condition. In HIF-1α-deficient chondrocytes, both glycolytic activities were reduced under hypoxic conditions. Statistical differences were calculated using data from four independent experiments. aP < 0.01, control oligonucleotide hypoxia vs HIF-1α-deficient hypoxia; *P < 0.05, **P < 0.01.
Figure 5 Glycosaminoglycan production and apoptosis induction in HIF-1α-deficient chondrocytes. In the scrambled oligonucleotide-treated groups, the amount of glycosaminoglycan (GAG) produced by cultured chondrocytes was higher under hypoxic conditions than under normoxic conditions. Under hypoxic conditions, GAG levels decreased in chondrocytes deficient in hypoxia-inducible factor 1α (HIF-1α). Statistical differences were calculated using data from four independent experiments. aP < 0.01, control oligonucleotide hypoxia vs HIF-1α-antisense hypoxia; *P < 0.05; **P < 0.01.
Figure 6 Apoptosis induction in HIF-1α-deficient chondrocytes. IL-1β was used for apoptosis induction under hypoxic or normoxic conditions in chondrocytes lacking hypoxia-inducible factor 1α (HIF-1α), treated with oligonucleotide, and cultured without oligonucleotide, and also not exposed to the oligonucleotide or HIF-1α antisense nucleotides. IL-1β-induced apoptosis was significantly increased in HIF-1α-deficient chondrocytes compared with chondrocytes treated with scrambled oligonucleotide under both normoxic and hypoxic culture conditions. Statistical differences were calculated using data from four independent experiments. **P < 0.01, control oligonucleotide group vs HIF-1α-antisense group.
Table 1 Characteristics of patients with osteoarthritis
Mankin grade
Donor Age (y) Sex Disease duration (y) Intact region Degenerated region
1 61 Female 5.5 3 7
2 62 Female 6.0 4 9
3 64 Male 8.4 3 8
4 66 Female 11.5 2 9
5 67 Female 9.6 2 9
6 68 Female 8.4 2 8
7 72 Female 9.0 3 9
Table 2 Sequences of PCR primers and probes
Primer (5'-3') Probe
c-Jun fw: TGCATGCTATCATTGGCTCATAC CCCGGCAACACACA-MGB
rv: CACACCATCTTCTGGTGTACAGTCT
HIF-1α fw:CTATGGAGGCCAGAAGAGGGTAT AGATCCCTTGAAGCTAG-MGB
rv:CCCACATCAGGTGGCTCATAA
Glucose transporter-1 fw:GGGCATGTGCTTCCAGTATGT CAACTGTGCGGCCCCTACGTCTTC
rv:ACGAGGAGCACCGTGAAGAT
fw, forward; rv, reverse
==== Refs
Maxwell PH Ratcliffe PJ Oxygen sensors and angiogenesis Semin Cell Dev Biol 2002 13 29 37 11969369 10.1006/scdb.2001.0287
Maxwell PH HIF-1's relationship to oxygen: simple yet sophisticated Cell Cycle 2004 3 156 159 14712080
Distler JHW Wenger RH Gassmann M Kurowska M Hirth A Gay S Distler O Physiologic responses to hypoxia and implications for hypoxia-inducible factors in the pathogenesis of rheumatoid arthritis Arthritis Rheum 2004 50 10 23 14730595 10.1002/art.11425
Semenza GL Roth PH Fang HM Wang GL Transcriptional regulation of genes encoding glycolytic enzymes by hypoxia-inducible factor 1 J Biol Chem 1994 269 23757 23763 8089148
Kung AL Wang S Klco JM Kaelin WG Livingston DM Suppression of tumor growth through disruption of hypoxia-inducible transcription Nat Med 2000 6 1335 1340 11100117 10.1038/82146
Semenza GL Regulation of mammalian O2 homeostasis by hypoxia-inducible factor 1 Annu Rev Cell Dev Biol 1999 15 551 578 10611972 10.1146/annurev.cellbio.15.1.551
Richard DE Berra E Pouyssegur J Angiogenesis: how a tumor adapts to hypoxia Biochem Biophys Res Commun 1999 266 718 722 10603309 10.1006/bbrc.1999.1889
Wenger RH Mammalian oxygen sensing, signaling and gene regulation J Exp Biol 2000 203 1253 1263 10729275
Wang GL Jiang BH Rue EA Semenza GL Hypoxia-inducible factor 1 is a basic-helix-loop-helix-PAS heterodimer regulated by cellular O2 tension Proc Natl Acad Sci USA 1995 92 5510 5514 7539918
Silver IA Measurement of pH and ionic composition of pericellular sites Philos Trans R Soc Lond B Biol Sci 1975 271 261 272 239420
Cernanec J Guilak F Weinberg JB Pisetsky DS Fermor B Influence of hypoxia and reoxygenation on cytokine-induced production of proinflammatory mediators in articular cartilage Arthritis Rheum 2002 46 968 975 11953974 10.1002/art.10213
Treuhaft PS McCarty DJ Synovial fluid pH, lactate, oxygen and carbon dioxide partial pressure in various joint diseases Arthritis Rheum 1971 14 475 484 5564921
Grimshaw MJ Mason RM Bovine articular chondrocyte function in vitro depends upon oxygen tension Osteoarthritis Cartilage 2000 8 386 392 10966846 10.1053/joca.1999.0314
Coimbra IB Jimenez SA Hawkins DF Piera-Velazquez S Stokes DG Hypoxia inducible factor-1 alpha expression in human normal and osteoarthritic chondrocytes Osteoarthritis Cartilage 2004 12 336 345 15023385 10.1016/j.joca.2003.12.005
Collins DH The Pathology of Articular and Spinal Diseases 1949 London: Edward Arnold and Co 76 79
Mankin HJ Dorfman H Lippiello L Zarins A Biochemical and metabolic abnormalities in articular cartilage from osteo-arthritic human hips. II. Correlation of morphology with biochemical and metabolic data J Bone Joint Surg Am 1971 53 523 537 5580011
Dudler J Renggli-Zulliger N Busso N Lotz M So A Effect of interleukin 17 on proteoglycan degradation in murine knee joints Ann Rheum Dis 2000 59 529 532 10873962 10.1136/ard.59.7.529
Piret JP Lecocq C Toffoli S Ninane N Raes M Michiels C Hypoxia and CoCl2 protect HepG2 cells against serum deprivation- and t-BHP-induced apoptosis: a possible anti-apoptotic role for HIF-1 Exp Cell Res 2004 295 340 349 15093734 10.1016/j.yexcr.2004.01.024
Farndale RW Buttle DJ Barrett AJ Improved quantitation and discrimination of sulphated glycosaminoglycans by use of dimethylmethylene blue Biochim Biophys Acta 1986 883 173 177 3091074
Parsch D Brummendorf TH Richter W Fellenberg J Replicative aging of human articular chondrocytes during ex vivo expansion Arthritis Rheum 2002 46 2911 2916 12428231 10.1002/art.10626
Pfander D Cramer T Schipani E Johnson RS HIF-1alpha controls extracellular matrix synthesis by epiphyseal chondrocytes J Cell Sci 2003 116 1819 1826 12665562 10.1242/jcs.00385
Hall JL Matter CM Wang X Gibbons GH Hyperglycemia inhibits vascular smooth muscle cell apoptosis through a protein kinase C-dependent pathway Circ Res 2000 87 574 580 11009562
Hall JL Wang X Van Adamson Zhao Y Gibbons GH Overexpression of Ref-1 inhibits hypoxia and tumor necrosis factor-induced endothelial cell apoptosis through nuclear factor-kappab-independent and -dependent pathways Circ Res 2001 88 1247 1253 11420300
Wiener CM Booth G Semenza GL In vivo expression of mRNAs encoding hypoxia-inducible factor 1 Biochem Biophys Res Commun 1996 225 485 488 8753788 10.1006/bbrc.1996.1199
Jain S Maltepe E Lu MM Simon C Bradfield CA Expression of ARNT, ARNT2, HIF1 alpha, HIF2 alpha and Ah receptor mRNAs in the developing mouse Mech Dev 1998 73 117 123 9545558 10.1016/S0925-4773(98)00038-0
Schipani E Ryan HE Didrickson S Kobayashi T Knight M Johnson RS Hypoxia in cartilage: HIF-1alpha is essential for chondrocyte growth arrest and survival Genes Dev 2001 15 2865 2876 11691837
Jeong JW Bae MK Ahn MY Kim SH Sohn TK Bae MH Yoo MA Song EJ Lee KJ Kim KW Regulation and destabilization of HIF-1alpha by ARD1-mediated acetylation Cell 2002 111 709 720 12464182 10.1016/S0092-8674(02)01085-1
Ivan M Kondo K Yang H Kim W Valiando J Ohh M Salic A Asara JM Lane WS Kaelin WG Jr HIFalpha targeted for VHL-mediated destruction by proline hydroxylation: implications for O2 sensing Science 2001 292 464 468 11292862
Huang J Zhao Q Mooney SM Lee FS Sequence determinants in hypoxia-inducible factor-1alpha for hydroxylation by the prolyl hydroxylases PHD1, PHD2, and PHD3 J Biol Chem 2002 277 39792 39800 12181324 10.1074/jbc.M206955200
Maxwell PH Wiesener MS Chang GW Clifford SC Vaux EC Cockman ME Wykoff CC Pugh CW Maher ER Ratcliffe PJ The tumour suppressor protein VHL targets hypoxia-inducible factors for oxygen-dependent proteolysis Nature 1999 399 271 275 10353251 10.1038/20459
Benton HP Tyler JA Inhibition of cartilage proteoglycan synthesis by interleukin I Biochem Biophys Res Commun 1988 154 421 428 3260778 10.1016/0006-291X(88)90703-6
Gowen M Wood DD Ihrie EJ Meats JE Russell RG Stimulation by human interleukin 1 of cartilage breakdown and production of collagenase and proteoglycanase by human chondrocytes but not by human osteoblasts in vitro Biochim Biophys Acta 1984 797 186 193 6320902
Stadler J Stefanovic-Racic M Billiar TR Curran RD McIntyre LA Georgescu HI Simmons RL Evans CH Articular chondrocytes synthesize nitric oxide in response to cytokines and lipopolysaccharide J Immunol 1991 147 3915 3920 1658153
Loeser RF Carlson CS Del Carlo M Cole A Detection of nitrotyrosine in aging and osteoarthritic cartilage: Correlation of oxidative damage with the presence of interleukin-1beta and with chondrocyte resistance to insulin-like growth factor 1 Arthritis Rheum 2002 46 2349 2357 12355482 10.1002/art.10496
Hiran TS Moulton PJ Hancock JT Detection of superoxide and NADPH oxidase in porcine articular chondrocytes Free Radic Biol Med 1997 23 736 743 9296450 10.1016/S0891-5849(97)00054-3
Stefanovic-Racic M Morales TI Taskiran D McIntyre LA Evans CH The role of nitric oxide in proteoglycan turnover by bovine articular cartilage organ cultures J Immunol 1996 156 1213 1220 8558000
Tiku ML Liesch JB Robertson FM Production of hydrogen peroxide by rabbit articular chondrocytes: enhancement by cytokines J Immunol 1990 145 690 696 2114447
Tiku ML Yan YP Chen KY Hydroxyl radical formation in chondrocytes and cartilage as detected by electron paramagnetic resonance spectroscopy using spin trapping reagents Free Radic Res 1998 29 177 187 9802549
Thornton RD Lane P Borghaei RC Pease EA Caro J Mochan E Interleukin 1 induces hypoxia-inducible factor 1 in human gingival and synovial fibroblasts Biochem J 2000 350 307 312 10926858 10.1042/0264-6021:3500307
Iwasa H Han J Ishikawa F Mitogen-activated protein kinase p38 defines the common senescence-signalling pathway Genes Cells 2003 8 131 144 12581156 10.1046/j.1365-2443.2003.00620.x
Lin AW Barradas M Stone JC van Aelst L Serrano M Lowe SW Premature senescence involving p53 and p16 is activated in response to constitutive MEK/MAPK mitogenic signaling Genes Dev 1998 12 3008 3019 9765203
Geng Y Valbracht J Lotz M Selective activation of the mitogen-activated protein kinase subgroups c-Jun NH2 terminal kinase and p38 by IL-1 and TNF in human articular chondrocytes J Clin Invest 1996 98 2425 2430 8941662
Hall JL Wang X Adamson V Zhao Y Gibbons GH Overexpression of Ref-1 inhibits hypoxia and tumor necrosis factor-induced endothelial cell apoptosis through nuclear factor-kB-independent and -dependent pathways Circ Res 2001 88 1247 1253 11420300
Jewell UR Kvietikova I Scheid A Bauer C Wenger RH Gassmann M Induction of HIF-1a in response to hypoxia is instantaneous FASEB J 2001 15 1312 1314 11344124
| 15987493 | PMC1175045 | CC BY | 2021-01-04 16:02:37 | no | Arthritis Res Ther. 2005 May 27; 7(4):R904-R914 | utf-8 | Arthritis Res Ther | 2,005 | 10.1186/ar1765 | oa_comm |
==== Front
Breast Cancer ResBreast Cancer Research1465-54111465-542XBioMed Central London bcr10111598744610.1186/bcr1011Research ArticleOverexpression of β1-chain-containing laminins in capillary basement membranes of human breast cancer and its metastases Fujita Manabu [email protected] Natalya M [email protected] Shikha [email protected] Kiyotoshi [email protected] Takako [email protected] William G [email protected] Alexander V [email protected] Keith L [email protected] Julia Y [email protected] Maxine Dunitz Neurosurgical Institute, Cedars–Sinai Medical Center, Los Angeles, California, USA2 Department of Pathology, Cedars–Sinai Medical Center, Los Angeles, California, USA3 Institute of Protein Research, Osaka University, Osaka, Japan4 Max-Planck-Institut für Biochemie, Martinsried, Germany5 Fred Hutchinson Cancer Research Center and Department of Pathobiology, University of Washington, Seattle, Washington, USA6 Ophthalmology Research Laboratories, Cedars–Sinai Medical Center, Los Angeles, California, USA2005 6 4 2005 7 4 R411 R421 12 11 2004 27 1 2005 9 2 2005 17 2 2005 Copyright © 2005 Fujita et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Introduction
Laminins are the major components of vascular and parenchymal basement membranes. We previously documented a switch in the expression of vascular laminins containing the α4 chain from predominantly laminin-9 (α4β2γ1) to predominantly laminin-8 (α4β1γ1) during progression of human brain gliomas to high-grade glioblastoma multiforme. Here, differential expression of laminins was studied in blood vessels and ductal epithelium of the breast.
Method
In the present study the expressions of laminin isoforms α1–α5, β1–β3, γ1, and γ2 were examined during progression of breast cancer. Forty-five clinical samples of breast tissues including normal breast, ductal carcinomas in situ, invasive ductal carcinomas, and their metastases to the brain were compared using Western blot analysis and immunohistochemistry for various chains of laminin, in particular laminin-8 and laminin-9.
Results
Laminin α4 chain was observed in vascular basement membranes of most studied tissues, with the highest expression in metastases. At the same time, the expression of laminin β2 chain (a constituent of laminin-9) was mostly seen in normal breast and carcinomas in situ but not in invasive carcinomas or metastases. In contrast, laminin β1 chain (a constituent of laminin-8) was typically found in vessel walls of carcinomas and their metastases but not in those of normal breast. The expression of laminin-8 increased in a progression-dependent manner. A similar change was observed from laminin-11 (α5β2γ1) to laminin-10 (α5β1γ1) during breast tumor progression. Additionally, laminin-2 (α2β1γ1) appeared in vascular basement membranes of invasive carcinomas and metastases. Chains of laminin-5 (α3β3γ2) were expressed in the ductal epithelium basement membranes of the breast and diminished with tumor progression.
Conclusion
These results suggest that laminin-2, laminin-8, and laminin-10 are important components of tumor microvessels and may associate with breast tumor progression. Angiogenic switch from laminin-9 and laminin-11 to laminin-8 and laminin-10 first occurs in carcinomas in situ and becomes more pronounced with progression of carcinomas to the invasive stage. Similar to high-grade brain gliomas, the expression of laminin-8 (and laminin-10) in breast cancer tissue may be a predictive factor for tumor neovascularization and invasion.
==== Body
Introduction
Identification of new markers for human breast cancer development, progression and metastases is important for successful breast tumor therapy and management. Ductal carcinoma in situ (DCIS)/ductal intraepithelial neoplasia is a proliferation of malignant epithelial cells within the mammary ductal system without evidence of infiltration. However, incomplete understanding of the natural history of DCIS and inability to identify predictive factors for the development of invasive carcinoma have resulted in a confusing variety of treatments for the disease [1,2]. How often DCIS transforms to invasive carcinoma and what are the factors that predispose to this transformation are unresolved questions. Invasive ductal carcinoma (IDC) is the most common type of breast cancer, accounting for 80% of all cases.
Angiogenesis (the formation of new blood vessels) is a fundamental process associated with normal development but also with tumor growth, invasion, and metastasis. Primary and metastatic breast tumors are dependent on angiogenesis, and they exhibit the greatest angiogenic activity at the beginning of tumor development [3,4]. Therefore, antiangiogenic therapy is currently regarded as a promising and relatively new approach to cancer treatment; a number of antiangiogenic drugs were recently developed, and a new antiangiogenic basis for emerging metronomic therapy is also being established [5]. Unlike dose-dense chemotherapy, which mostly targets proliferating tumor cells, frequent or continuous metronomic chemotherapy mainly targets endothelial cells [6]. It is important to identify novel targets for this therapy, which will probably be combined with classic chemotherapeutic drugs.
Angiogenesis is critical to solid tumor growth and invasion. Newly formed blood vessels participate in tumor formation and provide nutrients and oxygen to the tumor. Angiogenesis, a response to tumor growth, is a dynamic process that is highly regulated by signals from surrounding environment, including growth factors/cytokines and extracellular matrix (ECM). Their cooperative regulation is essential for angiogenesis accompanying the growth of solid tumors [7-9].
The ECM and its specialized structures, basement membranes (BMs), play important roles in tumor progression as barriers to invasion, migration substrata for tumor cells, and as components of tumor blood vessels. Penetration of vascular BMs occurs during tumor dissemination and metastasis. Laminins are major BM components and are important for cell adhesion, migration, and angiogenesis. Dysregulated cell–laminin interactions are major traits of various cancers. In many solid tumors, including breast cancer, BMs are often discontinuous or absent, which correlates with invasive properties [10-14]. The distributions of laminin chains α1, α3, α5, β1–β3, γ1, and γ2, as well as of type IV collagen chains, have been studied in various types of carcinomas and in normal tissues. Corroborating their widespread distribution in normal epithelial tissues, laminin-5 and laminin-10 are the most abundant laminins in the corresponding carcinomas [15]. Recent studies suggest that the expression of laminin-5 receptor, α6β4 integrin, may be a poor prognostic factor for invasive breast carcinoma [16]. Furthermore, the utilization of siRNA to reduce the expression of α6β4 integrin may be a useful approach to prevent carcinoma progression [17]. Cleavage of laminin-5 by matrix metalloproteinases (MMPs) produces a fragment (DIII) that binds to epidermal growth factor receptor and stimulates downstream signaling through mitogen-activated protein kinase, MMP-2 expression, and cell migration. These findings indicate that ECM cues may operate via direct stimulation of receptor tyrosine kinases (e.g. epidermal growth factor receptor) in tissue remodeling and, possibly, cancer invasion [18].
Laminin-8 (α4β1γ1) plays important roles in angiogenesis and migration of endothelial cells [19-21]. Laminin α4-chain-deficient mice exhibit impaired newborn capillary maturation [22]. These reports support the hypothesis on the pivotal role of laminin-8 in the process of neovascularization. In addition, our previous work has shown that laminin-8, a vascular BM component, was overexpressed in high-grade gliomas and their adjacent tissues as compared with normal brain, which correlated with shorter time to glioblastoma recurrence and patient survival [23,24]. Blocking laminin-8 expression resulted in the inhibition of glioma invasion in vitro [25].
Here, we studied the expression of laminins, in particular laminin-8 and laminin-9, in human breast tumors, such as DCIS, invasive ductal carcinoma, and metastases of IDC, in comparison with corresponding normal breast tissues.
Materials and methods
Tissue samples
Samples of breast cancers, breast cancer metastases to the brain, and samples of normal breast were obtained from the Department of Pathology and Laboratory Medicine, Cedars–Sinai Medical Center. The study protocol was approved by the institutional review board and conformed with the guidelines of the 1975 Declaration of Helsinki. Immediately after surgery, each sample was frozen in liquid nitrogen and stored at -80°C until protein extraction or embedding in OCT (optimal cutting temperature) compound for cryosectioning. Before protein extraction, each frozen sample was morphologically evaluated, in accordance with the World Health Organization classification of breast tumors.
A total of 45 samples were analyzed by Western blot analysis and immunohistochemistry, including normal breast tissues (n = 14), DCIS (n = 5), primary IDC, not otherwise specified (n = 23), and carcinomas metastatic to the brain (n = 3). Twenty-seven samples were analyzed using both methods to confirm laminin-8 and laminin-9 chain expression.
Immunohistochemistry
Sections of 38 specimens (14 normal breast, five DCIS, 16 IDC, and three brain metastases of cancer) were analyzed. Tissue samples were snap-frozen in liquid nitrogen by a pathologist immediately after surgery, embedded in OCT compound, and 8 μm sections were cut on a cryostat. Indirect immunofluorescence, photography, and routine negative controls were as described previously [23,24]. Briefly, we used well characterized polyclonal and mAbs to laminin chains α1–α5, β-β3, γ1, and γ2 (Table 1) [26-31]. Secondary cross-species absorbed fluorescein- and rhodamine-conjugated goat anti-mouse, anti-rat, and anti-rabbit antibodies were obtained from Chemicon International (Temecula, CA, USA). Polyclonal antibodies to human von Willebrand factor (Sigma-Aldrich Corp., St. Louis, MO, USA) were used for endothelial cell detection. Mouse mAbs to cytokeratin-8 and cytokeratin-18 (Biomeda, Foster City, CA, USA) were used for epithelial cell detection. The overwhelming majority of carcinomas also expressed these cytoskeletal proteins. mAbs were used as straight hybridoma supernatants or at 10–20 μg/ml when purified, and polyclonal antibodies were used at 20–30 μg/ml. Sections were viewed and photographed using an Olympus BH-40 fluorescence microscope equipped with 6 megapixel Magnafire digital camera. Routine specificity controls (without primary or secondary antibodies) were negative. At least two independent experiments were performed for each marker, with identical results.
Quantitation of tissue staining intensity
Staining intensity was graded as follows: -, no staining; +, weak staining; ++, distinct staining; +++, bright staining; ++++, very strong staining; and /, when vessels in the same specimen exhibited two different categories of staining. The immunofluorescent staining was independently analyzed by three researchers in each case.
Western blot analysis
Twenty-eight tissue samples were analyzed (10 normal breast tissues, four DCIS, 11 IDC, and three brain metastases of breast cancer). Tissue samples were snap-frozen in liquid nitrogen by a pathologist immediately after surgery. Proteins were separated using 10% Tris-glycine SDS-PAGE (Invitrogen, Carlsbad, CA, USA) under reducing conditions. Lysates of human glioma T98G, known to express laminin-8 but not laminin-9 [25,30], were used as positive control. The gels were blotted onto nitrocellulose membrane (Invitrogen). The membranes were probed with primary mAbs followed by chemiluminescent detection using the Immun-Star™ AP kit with alkaline phosphatase-conjugated secondary antibodies (Bio-Rad, Hercules, CA, USA). Antibodies (Table 1) were used to laminin α4 chain (mAb 8B12), β1 chain (mAb LT3), and b2 chain (mAb C4). Antibody to β-actin (Table 1) was used to control for equal loading of gel lanes.
Statistical analysis
Results of the immunostaining data were analyzed by the two-sided Fisher's exact test using the InStat software program (GraphPad Software, San Diego, CA, USA). To this end [23], the number of cases with a certain staining pattern in one experimental group (e.g. normal) was compared with the number of cases with the same staining pattern in another experimental group (e.g. breast cancer or brain metastasis). P < 0.05 was considered statistically significant.
Results
Laminin β1 chain is overexpressed in capillary basement membranes during tumor progression
To study laminin chain expression, serial sections of human breast tumor and normal tissues were stained either with hematoxylin and eosin for morphological observation (Fig. 1a, panels A–D; duplicated in Fig. 1b, panels A–D) or by indirect immunofluorescence with antibodies to different laminin chains. Some sections were double stained using antibodies to an endothelial marker, von Willebrand factor/factor-8 (F8; Fig. 1a, panels E–P), or epithelial cytokeratin-8 and cytokeratin-18 (CK; Fig. 1b, panels Q–X). We first concentrated on chains of laminin-8 (α4β1γ1) and laminin-9 (α4β2γ1) that underwent distinct changes during brain tumor progression [23,24] but that have not previously been studied in breast cancer.
The expression of laminin α4 chain in normal breast and DCIS was detected in the BMs of cytokeratin-8/18-positive epithelial cells of ductal and lobular structures (weak to negative in DCIS), as well as in BMs of factor-8-positive blood vessels (Table 2; Fig. 1a, panels E and F; Fig. 1b, panels Q and R). In invasive tumors, weak epithelial BM staining was only seen in the remnants of pre-existing ducts (not shown) and not around invading groups of epithelial cells (Fig. 1b, panel S). Vascular BMs were positive for α4 chain in all IDCs and metastatic tumors with distinct colocalization of α4 chain and factor-8 (Table 2; Fig. 1a, panels G and H). The staining intensity of α4 chain in vascular BMs of many primary and metastatic carcinomas was stronger than in normal tissue.
In normal breast, the epithelial or vascular expression of laminin β1 chain was nearly absent (Table 2; Fig. 1a, panel I; Fig. 1b, panel U). In DCIS, IDC and metastases, β1 chain appeared in the BMs of tumor vessels (Table 2; Fig. 1a, panels J–L; Fig. 1b, panels V–X).
Laminin β2 chain expression is decreased during tumor progression
In contrast to β1 chain, the expression of β2 chain was readily detected mainly around epithelial structures of normal breast tissue, with some vascular BM staining (Fig. 1a, panel M). This pattern was preserved in all DCIS cases (Fig. 1a, panel N) except one in which β2 chain was not detected. Additionally, β2 chain expression was not observed around invasive carcinoma cells or in vascular BMs of most IDCs and of all metastases (Table 2; Fig. 1a, panels O and P). In these cases, β2 chain could only be detected around remnant ducts within carcinomas.
The data summarized in Table 3 show that laminin-9 (α4β2γ1) is predominant in the vascular BMs of normal breast and DCIS. However, a switch from β2 to β1 chain leads to predominant expression of laminin-8 (α4β1γ1) in IDCs and especially in their metastases.
The expression of laminin-2 and laminin-10 increases in capillary basement membranes during tumor progression, similar to laminin-8
The expression of other laminin chains α1, α2, α3, α5, β3, γ1, and γ2 was also studied in normal and malignant breast tissues (Table 4). The α1 chain was only seen in three cases altogether, either in epithelial (one case; not shown) or in vascular (two cases; Table 4) BMs. The α2 chain, in accordance with previous data obtained in other tumors, was upregulated in vascular BMs of DCIS, invasive breast carcinomas, and metastases compared with normal breast (Table 4). Taking into account the expression of β1 chain, this finding indicates the appearance of laminin-2 (α2β1γ1) in tumor vascular BMs. Chains of laminin-5 (α3β3γ2) were mainly seen in ductal structures but not in blood vessel BMs (Table 4). The ubiquitous laminin α5 chain was seen in both epithelial and vascular BMs of all tissues. This chain is present in laminin-10 (α5β1γ1) and laminin-11 (α5β2γ1). Given the distribution of β1 and β2 chains presented above, there is also a shift from laminin-11 to laminin-10 in vascular BMs of most invasive tumors compared with normal breast (Tables 2 and 4).
Laminin γ1 chain, which is part of more than 10 laminin isoforms, was uniformly and strongly expressed in BMs of epithelial cells and blood vessels of all tissues studied (Table 2).
Western blotting reveals a shift from β2-containing to β1-containing laminins during breast cancer progression
To confirm the expression of select laminin chains, we compared laminin α4, β1, and β2 chains in normal and cancerous breast tissues using semiquantitative Western blot analysis with gel loading normalization by β-actin (Fig. 2). The expression of laminin α4 chain was variable and present in all tumor tissues and in 50% of corresponding normal tissues (two out of 10 normals shown in Fig. 2). The expression of β2 chain was high in all normal tissues, and the signal declined in DCIS, up to complete absence in IDCs and metastases. In contrast, expression of β1 chain was detected in 50% of DCIS (two out of four DCIS shown in Fig. 2) and in all invasive carcinomas and metastases, but only weakly in some normal breast samples. The data suggest that the expression of α4 chain in normal tissues corresponds mostly to laminin-9. In contrast to normal breast, a marked shift from β2 to β1 chain in invasive breast carcinomas and metastases suggests predominant expression of laminin-8 in a tumor grade-dependent manner. The results from Western blot analysis are in agreement with data obtained by immunohistochemistry.
Discussion
Laminins are heterotrimeric glycoproteins composed of α, β, and γ chains, and are commonly found as structural elements of all BMs. To date, five α, three β, and three γ chains have been identified and are known to give rise to at least 15 laminin isoforms [32,33]. Although the functions of laminins may vary by isoform, they serve not only as structural elements and as a scaffold for cell attachment, but also as signaling molecules through their interactions with cell surface receptors [32-34]. Specific transitions of laminin isoforms occur in various tissues at specific stages of development [35-39]. In invasive cancers, laminins usually become discontinuous or absent around tumor foci, which is attributed to either increased degradation or reduced synthesis. At the same time, previously documented changes in the expression of laminin isoforms concerned only α2-chain-containing laminins in basal cell carcinomas, medullary thyroid carcinomas, Schwannomas, and hepatocellular carcinomas [38,40-42]. We have now confirmed these data in breast tumors and their metastases (Table 4).
In this report we document for the first time a shift in α4 and α5 chain-containing laminin isoforms (from laminin-9 to laminin-8, and from laminin-11 to laminin-10, respectively) in invasive breast cancers. Chains of laminin-9 (α4β2γ1) and laminin-11 (α5β2γ1) were detected in vessel BMs of normal breast tissue. In DCIS, both laminin-8 and laminin-9 (plus laminin-10 and laminin-11) chains were expressed in blood vessel BMs. In invasive ductal breast carcinomas and their metastases to the brain, mostly laminin-8 and laminin-10 were expressed in vascular BMs, similar to the situation with brain gliomas, during the appearance of grade IV glioblastoma multiforme. In breast cancer the switch between laminin-9 and laminin-8 occurred in nearly all tumors, and therefore it was even more pronounced than in glioblastoma multiforme, with laminin-8 expression in 75% of cases [23,24]. The same was true for laminin-11 and laminin-10. The only difference between brain and breast tumors appears to be in the relative quantity of laminin α4 chain. It was distinctly upregulated in brain glioblastomas but not as much in invasive breast carcinomas. Laminin isoform switch in invasive breast cancers due to a shift from β2 to β1 chain may be useful for tumor prognosis in terms of further tumor progression and invasion potency.
Angiogenesis is essential for tumor growth and metastasis [43]. Tumor capillaries develop in a dynamic process, starting at the sites of local degradation of the vascular endothelial BMs. Afterward, endothelial cells migrate, proliferate, and differentiate to form a capillary sprout, while interacting with newly secreted ECM proteins from cancer cells and/or endothelial cells [34,43]. This remodeling of the vascular BMs by host endothelial cell is essential for tumor angiogenesis.
It is generally accepted that tumor cells secrete various angiogenic factors that enable endothelial cells to migrate into the tumor tissue and form new capillaries [34]. These factors may provide a mechanism for the observed switch of laminin-9 and laminin-11 in normal vascular BMs to laminin-8 and laminin-10 (plus appearance of laminin-2) in the microvascular BMs of invasive ductal breast carcinomas and of their metastases. In molecular terms, this switch relates to the change in expression of β2 to β1 laminin chain during breast cancer progression. This change may reflect the remodeling process of vessel BMs during progression from normal and DCIS to invasive carcinoma or metastasis. It has been shown that cleavage of laminin-5 γ2 chain by MMP-2 facilitates cell migration [44]. It may be suggested that, in breast carcinoma vessels, laminin β2 chain may also be degraded by some tumor-derived proteinases, which may trigger a compensatory upregulation of laminin-8 and laminin-10 to replace the 'normal' laminin-9 and laminin-11 in tumor tissue, which in turn would promote angiogenesis [9].
Another possible mechanism of laminin β2 to β1 chain switch in breast carcinomas may be related to different regulation of their expression. The TESS database analysis of laminin β1 and β2 chain gene promoter sequences [45] shows that β2 but not β1 promoter has a putative binding site for the early growth response protein Egr-2. This zinc finger DNA-binding transcription factor is a tumor suppressor and is decreased in various cancers [46,47]. Interestingly, Egr-2 expression is upregulated by tumor suppressor PTEN, which may play an important role in cell growth suppression [48,49]. Furthermore, the chromosomal loci of these two respective genes are very close to each other (Egr-2, 10q21-q22; PTEN, 10q23.31). Loss of heterozygosity of this chromosome 10 region and reduced PTEN expression are associated with poor outcome of invasive ductal breast carcinoma [50-52]. It may be suggested that the sequential downregulation of laminin β2 chain after the inactivation of PTEN and its downstream transcription factor Egr-2 in invasive breast cancer may bring about a compensatory increase in β1 chain expression, with the appearance of new laminin isoforms laminin-2, laminin-8, and laminin-10. Further experimentation is needed to support this mechanism.
A change from β2-containing to β1-containing laminins may present a special advantage for breast cancer cells. Laminin-8 and laminin-10 can promote endothelial cell attachment, migration, and tube formation on a BM matrix. Antisense inhibition of laminin-8 expression reduced glioma cell invasion through a BM matrix in vitro [25]. Therefore, accumulation of laminin-2, laminin-8, and laminin-10 in tumor vascular BMs might facilitate invasion of tumor cells through these BMs and subsequent metastasis. Indirect evidence in favor of laminin-10 as another modulator of glioma invasion was obtained in our experiments. Antisense oligonucleotides to β1 chain were more effective than those to laminin α4 chain in inhibiting glioma invasion in vitro [25]. Whereas the α4 antisense would downregulate only laminin-8, the β1 antisense would reduce both laminin-8 and laminin-10, thus supporting the role for laminin-10 in tumor invasion. Additional studies are needed to determine whether laminin-10 indeed has invasion-promoting activity. It would be interesting to determine whether other malignant tumors also have increased expression of laminin-2, laminin-8, and laminin-10. For the purposes of pathological diagnosis and prognosis, only the relative expression of β1 versus β2 chain may need to be determined. Antibodies, antisense oligonucleotides, or siRNA to laminin β1 chain might be useful for future treatment of solid tumors of various sites. In the case of breast cancers, such reagents may complement the existing and clinically useful herceptin antibody to HER-2/neu [53-55].
Conclusion
It may be concluded that laminin-2, laminin-8, and laminin-10 are important components of breast cancer microvessels, and that lack of laminin-9 and laminin-11 may play a role in remodeling of new vessels in breast cancer. The expression of laminin-2, laminin-8, and laminin-10 in cancer microvasculature may be related to the development of breast cancer-induced neovascularization and tumor progression. Similar to high-grade brain gliomas, a switch from vascular laminin-9 and laminin-11 to laminin-8 and laminin-10 in breast cancer tissue (from β2 to β1 chain) may be a predictive factor for tumor neovascularization and a possible target for antiangiogenic therapy. Because expressions of laminin-8 and laminin-10 have now been observed during progression of both gliomas and ductal breast carcinomas, they may have general predictive value in solid human tumors.
Abbreviations
BM = basement membrane; DCIS = ductal carcinoma in situ; ECM = extracellular matrix; IDC = invasive ductal carcinoma; mAb = monoclonal antibody; MMP = matrix metalloproteinase.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MF conducted immunostaining and Western blot analysis experiments. NMK processed tissues and conducted immunostaining experiments. SB provided tissue samples and made diagnoses. KS provided antibodies to laminin α4 chain for Western analysis and participated in manuscript writing. TS provided antibodies to laminin α4 chain for immunohistochemistry and participated in manuscript writing. WGC provided antibodies to laminin α3 and β3 chains and participated in manuscript writing. AVL provided antibody to laminin γ1 chain, participated in study design and conception, and in manuscript writing. KLB participated in study design and conception. JYL conceived the study, participated in its design and coordination, and in the writing of the manuscript. All authors read and approved the final manuscript.
Acknowledgements
This work was supported by the grants from the Maxine Dunitz Neurosurgical Institute, Cedars-Sinai Medical Center. The antibody C4 to laminin β2 chain produced by Dr Joshua Sanes was obtained from the Developmental Studies Hybridoma Bank, Department of Biology, University of Iowa (Iowa City, IA, USA), under contract N01-HD-2-3144 from the NICHD.
Figures and Tables
Figure 1 Immunohistochemistry of human breast tissues including normal, DCIS, primary IDC and metastases. (a) Panels A–D: hematoxylin and eosin staining of normal breast, DCIS, IDC and metastatic tissues, respectively. Panels E–H: double immunostaining with laminin α4 (red) and an endothelial marker, von Willebrand factor/factor-8 (F8; green). Panels I–L: double immunostaining with laminin β1 (red) and an endothelial marker von Willebrand factor (F8, green). Panels M–P: double immunostaining for laminin β2 (red) and F8 (green). For each representative case, serial sections are shown. (b) Panels A–D: hematoxylin and eosin staining (same as in Fig. 1a, panels A–D). Panels Q–T: double immunostaining for laminin α4 chain (red) and lining epithelium markers cytokeratins (CK)-8/18 (green). Panels U–X: double immunostaining for laminin β1 (red) and CK-8/18 (green). For each case, serial sections to Fig. 1a are shown. Because of lack of appropriate antibodies, no double staining could be performed for laminin β2 chain and CK-8/18. In normal breast tissues, laminin-9 chains α4 and β2 are expressed in BMs of mammary gland ducts (arrows in Fig. 1a, panels E and M, and Fig. 1b, panel Q) and blood vessels. In DCIS laminin α4 chain starts disappearing from ductal BMs (Fig. 1a, panel F, and Fig. 1b, panel R) but β2 chain is present (Fig. 1a, panel N [arrows]). Laminin-8 chains α4 and β1 and laminin-9 chains α4 and β2 colocalize in some microvessels. In all invasive ductal carcinomas, laminin-8 α4 and β1 chains are both found in BMs of F8-positive microvessels (Fig. 1a, panels G and K). Laminin-9 is absent (no β2 chain; Fig. 1a, panel O). In metastases of breast carcinoma, laminin-8 chains are seen in microvascular BMs (Fig. 1a, panels H and L; Fig. 1b, panels T and X) but laminin-9 is absent again (no β2 chain; Fig. 1a, panel P). BM, basement membrane; DCIS, ductal carcinoma in situ; IDC, invasive ductal carcinoma.
Figure 2 Western blot analysis. Shown are eight out of 28 samples (two normal breast samples, two DCIS, two IDC and two breast cancer metastases to brain) subjected to Western blot analysis for laminin α4, β1 and β2 chains. Gel loading was normalized by β-actin (lower row). The expression of laminin α4 chain, a constituent of laminin-8 and laminin-9, varies in normal and tumor tissues, with the highest expression detected in metastases. Laminin β2 chain, a constituent of laminin-9, is highly expressed in normal tissues, but its expression is very low in breast cancer tissues. In contrast, expression of laminin β1 chain, a constituent of laminin-8, is high in brain metastases and IDC but low in DCIS and absent in normal tissues. Laminin α4 chain migrates at 200 kDa, β1 chain at 230 kDa, β2 chain at 190 kDa, and β-actin at 47 kDa. The T98G glioblastoma cell line, which is known to express α4 and β1 chains of laminin-8 but no β2 chain, is used as a positive control.
Table 1 Antibodies used in the study
Antigen Antibody Reference/source
Laminin α1 chain Rabbit pAb 1057 (VI/V) [26]
Laminin α2 chain Mouse mAb 1F9 [27]
Laminin α3 chain Mouse mAb D2-1 [28]
Mouse mAb C2-5
Laminin α4 chain Rabbit pAb 1129 (IIIa) [29]
Mouse mAb 8B12 [30]
Laminin α5 chain Mouse mAb 4C7 Chemicon International
Laminin β1 chain Rat mAb LT3 Upstate
Mouse mAb LN26-7 Axxora/Alexis
Laminin β2 chain Mouse mAb C4 Developmental Studies Hybridoma Bank
Laminin β3 chain Mouse mAb A2'-2 [28]
Laminin γ1 chain Rat mAb A5 [31]
Laminin γ2 chain Mouse mAb D4B5 Chemicon International
Cytokeratin-8 and -18 Mouse mAbs B22.1 & B23.1 Biomeda
β-actin Mouse mAb AC15 Sigma-Aldrich
von Willebrand factor Rabbit pAb Sigma-Aldrich
mAb, monoclonal antibody; pAb, polyclonal antibody.
Table 2 Expression of laminin-8 and laminin-9 chains in breast tissue blood vessel basement membranes
Sample Diagnosis Ln-α4 Ln-β1 Ln-β2 Ln-γ1 Ln typea
7 Normal + - +++ +++ 9
9 Normal ++ + +++ +++ 9
11 Normal + ++ ++++ +++ 9
17 Normal ++ - +++ ++ 9
21 Normal - - +++ +++ 9
25 Normal + - +++ +++ 9
32 Normal ++ +/- +++ +++ 9
75 Normal + +/- + ++ 9
77 Normal +++ - +++ +++ 9
79 Normal ++ - + ++ 9
64 Normal + ++ + +++ 8/9
65 Normal + +/- - ++ 9
66 Normal ++ + - +++ 8
67 Normal + + + ++ 9
8 DCIS ++ ++ +++ +++ 8/9
22 DCIS +++ +++ - +++ 8
32 DCIS ++ +/- +++ +++ 9
38 DCIS ++ + ++ ++ 9
41 DCIS ++ + +++ +++ 9
1 IDC +++ ++++ - +++ 8
2 IDC +++ +++ - +++ 8
3 IDC +++ +++ - +++ 8
5 IDC ++ ++ - ++ 8
6 IDC ++++ +++ - +++ 8
12 IDC + - + +++ 9
14 IDC ++ + ++ +++ 9
20 IDC ++++ ++++ - +++ 8
22 IDC +++ ++ - +++ 8
24 IDC ++++ +++ ++ +++ 8/9
28 IDC +++ +++ - +++ 8
30 IDC +++ ++ ++ +++ 8/9
34 IDC +++ ++ +++ +++ 8/9
36 IDC ++++ ++++ - +++ 8
76 IDC +++ ++ - +++ 8
78 IDC ++ ++ - +++ 8
121 Metastasis ++++ +++ +/- +++ 8
146 Metastasis ++++ ++++ - +++ 8
157 Metastasis +++ +++ +/- +++ 8
aPredominant laminin type is shown for each case; when some vessels had one isoform and the others had another, both are shown (see also Table 3). Staining intensity was graded as follows: -, no staining; +, weak staining; ++ distinct staining; +++, bright staining; ++++, very strong staining, /, some vessels in the same sample are in one category and some are in another category. Ln, laminin; DCIS, ductal carcinoma in situ; IDC, invasive ductal carcinoma.
Table 3 Summary of laminin-8 and laminin-9 expression in breast tissues as determined by immunohistochemistry
Histological diagnosis Number of cases Laminin-8 (n [%]) Laminin-8/9 (n [%]) Laminin-9 (n [%])
Normal breast tissue 14 1 (7) 1 (7) 12 (86)
Ductal carcinoma in situ 5 1 (20) 1 (20) 3 (60)
Invasive ductal carcinoma 16 11 (69) 3 (19) 2 (12)
Metastasis to the brain 3 3 (100) 0 0
The percentage of cases with a given predominant laminin isoform was determined using data in Table 1. For both laminin-8 and laminin-9 expression, the difference between normal tissues and carcinomas or metastases is statistically significant (P < 0.015).
Table 4 Expression of different laminin chains in breast tissue blood vessel basement membranes
Sample Diagnosis Ln-α1 Ln-α2 Ln-α3 Ln-α5 Ln-β3 Ln-γ1 Ln-γ2
16 Normal + - - +++ - +++ -
17 Normal - - - +++ - ++ -
75 Normal - + - +++ - +++ +
77 Normal - - - ++ - +++ ++
67 Normal - - - +++ - +++ -
22 DCIS - - - +++ - +++ -
38 DCIS ++ ++ - +++ - ++ -
41 DCIS - + - +++ - +++ -
50 DCIS - ++ - +++ - +++ -
20 IDC - ++/+++ - +++ - +++ -
30 IDC - - - + - +++ -
28 IDC - ++ - +++ - +++ -
52 IDC - - - +++ - +++ -
54 IDC - ++ - +++ - +++ -
121 Metastasis - +++ - +++ - +++ -
146 Metastasis - +++ - +++ - +++ -
157 Metastasis - ++ - +++ - +++ -
membranes Ln, laminin; DCIS, ductal carcinoma in situ; IDC, invasive ductal carcinoma. Staining intensity was graded as follows: -, no staining; +, weak staining; ++ distinct staining; +++, bright staining; ++++, very strong staining, /, some vessels in the same sample are in one category and some are in another category.
==== Refs
Nakhlis F Morrow M Ductal carcinoma in situ Surg Clin North Am 2003 83 821 839 12875598
Morrow M Schnitt SJ Treatment selection in ductal carcinoma in situ JAMA 2000 283 453 555 10659859 10.1001/jama.283.4.453
Weidner N Folkman J Pozza F Bevilacqua P Allred EN Moore DH Meli S Gasparini G Tumor angiogenesis: a new significant and independent prognostic indicator in early-stage breast carcinoma J Natl Cancer Inst 1992 84 1875 1887 1281237
Rahman MA Masakazu T Anti-angiogenic thereapy in breast cancer Biomed Pharmacother 2003 57 463 470 14637390 10.1016/j.biopha.2003.09.009
Kerbel RS Kamen BA The anti-angiogenic basis of metronomic chemotherapy Nat Rev Cancer 2004 4 423 436 15170445 10.1038/nrc1369
Browder T Butterfield CE Kraling BM Shi B Marshall B O'Reilly MS Folkman J Antiangiogenic scheduling of chemotherapy improves efficacy against experimental drug-resistant cancer Cancer Res 2000 60 1878 1886 10766175
Sivridis E Giatromanolaki A Koukourakis MI The vascular network of tumours: what is it not for? J Pathol 2003 201 173 180 14517833 10.1002/path.1355
Hasan J Byers R Jayson GC Intra-tumoural microvessel density in human solid tumours Br J Cancer 2002 86 1566 1577 12085206 10.1038/sj.bjc.6600315
Engels K Fox SB Whitehouse RM Gatter KC Harris AL Distinct angiogenic patterns are associated with high-grade in situ ductal carcinomas of the breast J Pathol 1997 181 207 212 9120727 10.1002/(SICI)1096-9896(199702)181:2<207::AID-PATH758>3.3.CO;2-W
Guelstein VI Tchypysheva TA Ermilova VD Ljubimov AV Myoepithelial and basement membrane antigens in benign and malignant human breast tumors Int J Cancer 1993 53 269 277 8425764
Engbring JA Kleinman HK The basement membrane matrix in malignancy J Pathol 2003 200 465 470 12845613 10.1002/path.1396
Kalluri R Basement membranes: structure, assembly and role in tumour angiogenesis Nat Rev Cancer 2003 3 422 433 12778132 10.1038/nrc1094
Duffy MJ Maguire TM Hill A McDermott E O'Higgins N Metalloproteinases: role in breast carcinogenesis, invasion and metastasis Breast Cancer Res 2000 2 252 257 11250717 10.1186/bcr65
Campo E Merino MJ Tavassoli FA Charonis AS Stetler-Stevenson WG Liotta LA Evaluation of basement membrane components and the 72 kDa type IV collagenase in serous tumors of the ovary Am J Surg Pathol 1992 16 500 507 1318001
Määttä M Virtanen I Burgeson R Autio-Harmainen H Comparative analysis of the distribution of laminin chains in the basement membranes in some malignant epithelial tumors: the α1 chain of laminin shows a selected expression pattern in human carcinomas J Histochem Cytochem 2001 49 711 726 11373318
Diaz LK Zhou X Welch K Sahin A Gilcrease MZ Chromogenic in situ hybridization for α6β4 integrin in breast cancer: correlation with protein expression J Mol Diagn 2004 6 10 15 14736821
Lipscomb EA Dugan AS Rabinovitz I Mercurio AM Use of RNA interference to inhibit integrin (α6β4)-mediated invasion and migration of breast carcinoma cells Clin Exp Metastasis 2003 20 569 576 14598892 10.1023/A:1025819521707
Schenk S Hintermann E Bilban M Koshikawa N Hojilla C Khokha R Quaranta V Binding to EGF receptor of a laminin-5 EGF-like fragment liberated during MMP-dependent mammary gland involution J Cell Biol 2003 161 197 209 12695504 10.1083/jcb.200208145
Fujiwara H Gu J Sekiguchi K Rac regulates integrin-mediated endothelial cell adhesion and migration on laminin-8 Exp Cell Res 2004 292 67 77 14720507 10.1016/j.yexcr.2003.08.010
Li J Zhang YP Kirsner RS Angiogenesis in wound repair: angiogenic growth factors and the extracellular matrix Microsc Res Tech 2003 60 107 114 12500267 10.1002/jemt.10249
Gonzales M Weksler B Tsuruta D Goldman RD Yoon KJ Hopkinson SB Flitney FW Jones JCR Structure and function of a vimentin-associated matrix adhesion in endothelial cells Mol Biol Cell 2001 12 85 100 11160825
Thyboll J Kortesmaa J Cao R Soininen R Wang L Iivanainen A Sorokin L Risling M Cao Y Tryggvason K Deletion of the laminin α4 chain leads to impaired microvessel maturation Mol Cell Biol 2002 22 1194 1202 11809810 10.1128/MCB.22.4.1194-1202.2002
Ljubimova JY Lakhter AJ Loksh A Yong WH Riedinger MS Miner JH Sorokin LM Ljubimov AV Black KL Overexpression of a4 chain-containing laminins in human glial tumors identified by gene microarray analysis Cancer Res 2001 61 5601 5610 11454714
Ljubimova JY Fujita M Khazenzon NM Das A Pikul BB Newman D Sekiguchi K Sorokin LM Sasaki T Black KL Association between laminin-8 and glial tumor grade, recurrence, and patient survival Cancer 2004 101 604 612 15274074 10.1002/cncr.20397
Khazenzon NM Ljubimov AV Lakhter AJ Fujita M Fujiwara H Sekiguchi K Sorokin LM Petajaniemi N Virtanen I Black KL Ljubimova JY Antisense inhibition of laminin-8 expression reduces invasion of human gliomas in vitro Mol Cancer Ther 2003 2 985 994 14578463
Ettner N Göhring W Sasaki T Mann K Timpl R The N-terminal globular domain of the laminin α1 chain binds to α1β1 and α2β1 integrins and to the heparan sulfate-containing domains of perlecan FEBS Lett 1998 430 217 221 9688542 10.1016/S0014-5793(98)00601-2
Engvall E Earwicker D Haaparanta T Ruoslahti E Sanes J Distribution and isolation of four laminin variants; tissue restricted distribution of heterotrimers assembled from five different subunits Cell Regul 1990 1 731 740 2099832
Sigle RO Gil SG Bhattacharya M Ryan MC Yang TM Brown TA Boutaud A Miyashita Y Olerud J Carter WG Globular domains 4/5 of the laminin α3 chain mediate deposition of precursor laminin 5 J Cell Sci 2004 117 4481 4494 15316072 10.1242/jcs.01310
Sasaki T Mann K Timpl R Modification of the laminin α4 chain by chondroitin sulfate attachment to its N-terminal domain FEBS Lett 2001 505 173 178 11557064 10.1016/S0014-5793(01)02812-5
Fujiwara H Kikkawa Y Sanzen N Sekiguchi K Purification and characterization of human laminin-8. Laminin-8 stimulates cell adhesion and migration through α3β1 and α6β1 integrins J Biol Chem 2001 276 17550 17558 11278628 10.1074/jbc.M010155200
Ljubimov AV Bartek J Couchman JR Kapuller LL Veselov VV Kovarik J Perevoshchikov AG Krutovskikh VA Distribution of individual components of basement membrane in human colon polyps and adenocarcinomas as revealed by monoclonal antibodies Int J Cancer 1992 50 562 566 1371500
Colognato H Yurchenco PD Form and function: the laminin family of heterotrimers Dev Dyn 2000 218 213 234 10842354 10.1002/(SICI)1097-0177(200006)218:2<213::AID-DVDY1>3.0.CO;2-R
Miner JH Yurchenco PD Laminin functions in tissue morphogenesis Annu Rev Cell Dev Biol 2004 20 255 284 15473841 10.1146/annurev.cellbio.20.010403.094555
Patarroyo M Tryggvason K Virtanen I Laminin isoforms in tumor invasion, angiogenesis and metastasis Semin Cancer Biol 2002 12 197 207 12083850 10.1016/S1044-579X(02)00023-8
Miner JH Patton BL Lentz SI Gilbert DJ Snider WD Jenkins NA Copeland NG Sanes JR The laminin alpha chains: expression, developmental transitions, and chromosomal locations of α1–5, identification of heterotrimeric laminins 8–11, and cloning of a novel α3 isoform J Cell Biol 1997 137 685 701 9151674 10.1083/jcb.137.3.685
Sorokin LM Pausch F Durbeej M Ekblom P Differential expression of five laminin α(1–5) chains in developing and adult mouse kidney Dev Dyn 1997 210 446 462 9415429 10.1002/(SICI)1097-0177(199712)210:4<446::AID-AJA8>3.0.CO;2-G
Tiger CF Champliaud MF Pedrosa-Domellof F Thornell LE Ekblom P Gullberg D Presence of laminin α5 chain and lack of laminin α1 chain during human muscle development and in muscular dystrophies J Biol Chem 1997 272 28590 28595 9353324 10.1074/jbc.272.45.28590
Seebacher T Medina JL Bade EG Laminin α5, a major transcript of normal and malignant rat liver epithelial cells, is differentially expressed in developing and adult liver Exp Cell Res 1997 237 70 76 9417868 10.1006/excr.1997.3758
St John PL Wang R Yin Y Miner JH Robert B Abrahamson DR Glomerular laminin isoform transitions: errors in metanephric culture are corrected by grafting Am J Physiol Renal Physiol 2001 280 F695 F705 11249861
Sollberg S Peltonen J Uitto J Differential expression of laminin isoforms and β4 integrin epitopes in the basement membrane zone of normal human skin and basal cell carcinomas J Invest Dermatol 1992 98 864 870 1375618 10.1111/1523-1747.ep12457080
Hsiao LL Engvall E Peltonen J Uitto J Expression of laminin isoforms by peripheral nerve-derived connective tissue cells in culture. Comparison with epitope distribution in normal human nerve and neural tumors in vivo Lab Invest 1993 68 100 108 7678659
Lekmine F Feracci H Milhaud G Treilhou-Lahille F Jeanne N Expression of laminin-2 by normal and neoplastic rat C cells during the development of medullary thyroid carcinoma Virchows Arch 1999 434 325 332 10335943 10.1007/s004280050348
Zetter BR Angiogenesis and tumor metastasis Annu Rev Med 1998 49 407 424 9509272 10.1146/annurev.med.49.1.407
Koshikawa N Giannelli G Cirulli V Miyazaki K Quaranta V Role of cell surface metalloprotease MT1-MMP in epithelial cell migration over laminin-5 J Cell Biol 2000 148 615 624 10662785 10.1083/jcb.148.3.615
Schug J Baxevanis AD, Davison DB, Page RDM, Petsko GA, Stein LD, Stormo GD Using TESS to predict transcription factor binding sites in DNA sequence Current Protocols in Bioinformatics 2003 New York: John Wiley & Sons, Inc unit 2.6.
Unoki M Nakamura Y EGR2 induces apoptosis in various cancer cell lines by direct transactivation of BNIP3L and BAK Oncogene 2003 22 2172 2185 12687019 10.1038/sj.onc.1206222
Nakahara Y Shiraishi T Okamoto H Mineta T Oishi T Sasaki K Tabuchi K Detrended fluctuation analysis of genome-wide copy number profiles of glioblastomas using array-based comparative genomic hybridization Neuro-oncol 2004 6 281 289 15494095 10.1215/S1152851703000632)
Matsushima-Nishiu M Unoki M Ono K Tsunoda T Minaguchi T Kuramoto H Nishida M Satoh T Tanaka T Nakamura Y Growth and gene expression profile analyses of endometrial cancer cells expressing exogenous PTEN Cancer Res 2001 61 3741 3749 11325847
Unoki M Nakamura Y Growth-suppressive effects of BPOZ and EGR2, two genes involved in the PTEN signaling pathway Oncogene 2001 20 4457 4465 11494141 10.1038/sj.onc.1204608
Garcia JM Silva JM Dominguez G Gonzalez R Navarro A Carretero L Provencio M Espana P Bonilla F Allelic loss of the PTEN region (10q23) in breast carcinomas of poor pathophenotype Breast Cancer Res Treat 1999 57 237 243 10617300 10.1023/A:1006273516976
Leighton X Srikantan V Pollard HB Sukumar S Srivastava M Significant allelic loss of ANX7region (10q21) in hormone receptor negative breast carcinomas Cancer Lett 2004 210 239 244 15183540 10.1016/j.canlet.2004.01.018
Chung MJ Jung SH Lee BJ Kang MJ Lee DG Inactivation of the PTEN gene protein product is associated with the invasiveness and metastasis, but not angiogenesis, of breast cancer Pathol Int 2004 54 10 15 14674989 10.1111/j.1440-1827.2004.01576.x
Slamon DJ Godolphin W Jones LA Holt JA Wong SG Keith DE Levin WJ Stuart SG Udove J Ullrich A Studies of the HER-2/neu proto-oncogene in human breast and ovarian cancer Science 1989 244 707 712 2470152
Pegram MD Pienkowski T Northfelt DW Eiermann W Patel R Fumoleau P Quan E Crown J Toppmeyer D Smylie M Results of two open-label, multicenter phase II studies of docetaxel, platinum salts, and trastuzumab in HER2-positive advanced breast cancer J Natl Cancer Inst 2004 96 759 769 15150304
Nahta R Hung MC Esteva FJ The HER-2-targeting antibodies trastuzumab and pertuzumab synergistically inhibit the survival of breast cancer cells Cancer Res 2004 64 2343 2346 15059883
| 15987446 | PMC1175051 | CC BY | 2021-01-04 16:04:33 | no | Breast Cancer Res. 2005 Apr 6; 7(4):R411-R421 | utf-8 | Breast Cancer Res | 2,005 | 10.1186/bcr1011 | oa_comm |
==== Front
Breast Cancer ResBreast Cancer Research1465-54111465-542XBioMed Central London bcr10151598744410.1186/bcr1015Research ArticleIncreased level of phosphorylated akt measured by chemiluminescence-linked immunosorbent assay is a predictor of poor prognosis in primary breast cancer overexpressing ErbB-2 Cicenas Jonas [email protected] Patrick [email protected] Vincent [email protected] Martin [email protected]üng Willy [email protected] Edward [email protected] Mark [email protected] Urs [email protected] Serenella [email protected] Stiftung Tumorbank Basel, Basel, Switzerland2 University Clinics, Department of Research, Molecular Tumor Biology, Basel, Switzerland3 OncoScore AG, Riehen, Switzerland4 University Clinics, Department of Gynecology, Basel, Switzerland5 University of Virginia, Department of Anatomy and Cell Biology, East Carolina School of Medicine, Charlottesville, Virginia, USA2005 24 3 2005 7 4 R394 R401 14 10 2004 16 12 2004 9 2 2005 28 2 2005 Copyright © 2005 Cicenas et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Introduction
Akt1, Akt2 and Akt3 kinases are downstream components of phosphoinositol 3-kinase derived signals from receptor tyrosine kinases, which influence cell growth, proliferation and survival. Akt2 overexpression and amplification have been described in breast, ovarian and pancreatic cancers. The present study was designed to investigate the prognostic significance of activated Akt in primary breast cancer and its association with other tumour biomarkers.
Methods
Using a two-site chemiluminescence-linked immunosorbent assay, we measured the quantitative expression levels of total phosphorylated (P-S473) Akt (Akt1/Akt2/Akt3) on cytosol fractions obtained from fresh frozen tissue samples of 156 primary breast cancer patients.
Results
Akt phosphorylation was not associated with nodal status or ErbB-2 protein expression levels. High levels of phosphorylated Akt correlated (P < 0.01) with poor prognosis, and the significance of this correlation increased (P < 0.001) in the subset of patients with ErbB-2 overexpressing tumours. In addition, phosphorylated Akt was found to be associated with mRNA expression levels of several proliferation markers (e.g. thymidylate synthase), measured using quantitative real-time RT-PCR.
Conclusion
Our findings demonstrate that, in breast cancer patients, Akt activation is associated with tumour proliferation and poor prognosis, particularly in the subset of patients with ErbB2-overexpressing tumours.
==== Body
Introduction
Akt/protein kinase B (PKB) is a serine/threonine kinase that is involved in mediating various biological responses, such as inhibition of apoptosis and stimulation of cell proliferation (for review [1,2]). Three mammalian isoforms are currently known [1]: Akt1/PKBα, Akt2/PKBβ and Akt3/PKBγ. Akt1 was first discovered as a cellular homologue of the viral oncogene v-Akt, which causes leukaemia in mice [3] and is the predominant isoform in most tissues. High expression of Akt2 has been observed in insulin-responsive tissues, whereas Akt3 has been shown to be predominantly expressed in brain and testis [2].
Phosphoinositol-3-phosphate (PIP3) is a product of phosphoinositol 3-kinase enzymatic activity and has been shown to be a prerequisite lipid modulator of Akt activity [4]. PIP3 has been described as a downstream component of a wide range of receptors, including the c-Met receptor [5], the epidermal growth factor receptor family [6], fibroblast growth factor receptor [7], insulin growth factor receptor [8] and platelet-derived growth factor receptor [9]. In addition, Akt activity can be regulated by the PTEN tumour suppressor gene, which negatively regulates PIP3 levels (for review [10]). After PIP3 binding, Akt1 is activated by phosphorylation on two critical residues, namely threonine 308 (T308) and serine 473 (S473); similar activation residues (S472 and S474, respectively) are highly conserved in Akt2 and Akt3 (for review [1,2]). Several studies have found Akt2 to be amplified or overexpressed at the mRNA level in various tumour cell lines [11-13] and in a number of human malignancies, such as colon, pancreatic and breast cancers [14-16]. However, activation of Akt1, Akt2 and Akt3 by phosphorylation appears to be more clinically relevant than detection of Akt2 amplification or overexpression.
To date, several groups have investigated the phosphorylation of active Akt in breast, prostate, colon and pancreatic tumours by immunohistochemistry [14,17-22]. Under such conditions, phosphorylation structures may be disturbed by formalin fixation, rendering specific antigen sites inaccessible. Moreover, immunohistochemistry gives only semiquantitative results, limiting statistical analysis. Alternatively, enzyme immunoassays (EIAs) have the advantage that they yield highly reproducible and sensitive results of quantitative values.
In the present study we detected phosphorylated Akt (P-Akt) by means of a novel two-site chemiluminescence-linked immunoassay (CLISA) in fresh frozen primary tissue samples from 156 primary breast cancer patients. Because it was shown in previous immunohistochemistry studies that S473 P-Akt has prognostic significance [17-19], the aim of the present study was to measure levels of P-Akt continuously using CLISA and correlate these with survival and factors that are involved in tumourigenesis. Given that the antibody used in the reported immunohistochemistry studies recognized all Akt isoforms, we have developed an assay that allows specific quantitative detection of active Akt1, Akt2 and Akt3 when phosphorylated on their corresponding residues, namely S473, S472 and S474, respectively.
Materials and methods
Tumor and patient characteristics
Fresh material obtained during surgery was kept on ice and examined by a pathologist. Representative specimens with more than 60% tumour cells were sent to the Stiftung Tumorbank Basel (STB), immediately shock frozen and cryopreserved (-80°C). All activities of the STB are in accordance with an official Swiss permit, which guarantees patient confidentiality and respects ethical issues. For the present study, 156 samples of primary breast tumours were selected. Those samples overexpressing ErbB-2 (>500 U/mg total protein) were selected, based on ErbB-2 protein expression levels routinely detected using EIAs at the time of surgery by the STB [23]. EIA ErbB-2 positive samples correlate strongly with DAKO 3+ and with ErbB-2 amplification detected by fluorescent in situ hybridization (FISH; data not shown).
All patients underwent primary surgery before January 1996. Sixty-seven patients (43%) experienced disease recurrence within the median follow-up time of 57 months (range 27–88 months). Sixty-six patients (42%) were node negative, and 90 (58%) were node positive. Forty tumours (26%) were oestrogen receptor (ER)-α negative. Ninety-five patients (61%) had ErbB-2-negative (<500 U/mg total protein) and 61 patients (39%) had ErbB-2 positive tumours [23]. None of the patients received neoadjuvant therapy. Patient and tumour characteristics are summarized in Table 1.
Cell lines and tissue culture
MCF-7 breast cancer cells were cultured in IMEM-ZO (improved minimal essential medium with zinc option) supplemented with 5% foetal bovine serum, l-glutamine and antibiotics (penicillin/streptomycin) at 37°C in a 5% carbon dioxide incubator. For the phospho-standard preparation, subconfluent MCF-7 cells were serum starved for 48 hours in serum-free media, and were treated with NaF and Na3VO4 for 1 hour, and then with 10% foetal bovine serum for 10 min. Cells were lysed for 5 min on ice in EB lysis buffer (20 mmol/l Tris-HCl [pH 7.4], 0.5 mol/l NaCl, 10 mmol/l EDTA, 1% Triton X100, 20 mmol/l NaF, 20 mmol/l glycerophosphate, 2 mmol/l Na3VO4, proteinase inhibitor cocktail [Roche, Indianapolis, IN, USA]), centrifuged at 20,000 g for 5 min and supernatant was stored at -80°C.
Measurement of oestrogen receptor, progesterone receptor and ErbB-2 protein levels in tumour extracts by enzyme immunoassay
Tissue homogenates were prepared in accordance with standard procedures for tumour marker measurement using EIAs, as previously described [23]. In brief, the frozen tissues were pulverized in liquid nitrogen using a Micro-Dismembrator U (B Braun Melsungen AG, Melsungen, Germany). The powder was homogenized using a tissue homogenizer (Ultra-Turrax; Janke & Kunkel, IKA-Werke, Staufen, Germany) for 20 s in three volumes of ice-cold extraction buffer. The homogenate was centrifuged at 800 g for 30 min at 2°C, and the resulting supernatant re-centrifuged in an ultracentrifuge (Beckman Instruments, Fullerton, CA, USA) at 100,000 g. The resulting supernatants (cytosols) were used for measurement of the hormone receptors (ER, progesterone receptor [PgR]), and the membrane fractions were used for EIA measurement of membrane-associated ErbB-2. ER and PgR concentrations were measured from tumour cytosolic extracts by commercial quantitative ER and PgR EIA kits (Abbott Laboratories, Abbott Park, IL, USA) using a Quantum II photometer (Abbott Laboratories, Abbott Park, IL, USA). Quality control of ER and PgR measurements was carried out in collaboration with the Receptor Biomarker Group of the European Organization for Research and Treatment of Cancer. ErbB-2 receptor levels were determined on the particulate membrane fractions of tumour extracts using a commercial monoclonal antibody EIA kit, described by Eppenberger-Castori and coworkers [23].
Immunoassay of phosphorylated Akt level
Neither antibody used in the CLISA discriminates between Akt isoforms. The catching antibody (anti-Akt/PKB, PH domain, clone SKB1; Upstate Biotechnology, Lake Placid, NY, USA) recognizes Akt1/PKBα, Akt2/PKBβ and Akt3/PKBγ (weak to none) based on immunoblot analysis using 100 ng recombinant fusion protein for each isoform, as reported by the manufacturer. The detecting phospho-specific (S473) Akt monoclonal antibody (4E2) detects endogenous levels of Akt1 only when phosphorylated at serine-473. This antibody also recognizes Akt2 (S472) and Akt3 (S474) if they are phosphorylated at the corresponding residues, according to the information obtained from the manufacturer (Cell Signaling Technology, Inc., Beverly, MA, USA). However, 4E2 does not recognize other Akt phosphorylation sites.
S473 phosphorylated Akt levels were measured using a novel two-site CLISA. Black 96-well microtitre plates (Nunc Black MaxiSorp Surface; Nalgen Nunc International, Rochester, NY, USA) were coated with coating antibody at a concentration of 3 mg/ml of coating buffer (phosphate-buffered saline with 0.6 mmol/l EDTA) in a volume of 100 μl/well and kept at 4°C overnight. To measure P-Akt, respective tumour extracts were prepared as described above in the presence of NaF and Na3VO4. Before sample applications, the coated microtitre plates were washed five times with 200 μl/well washing buffer (25 mmol/l HEPES [pH 7.4], 300 mmol/l NaCl, 0.05% Tween-20) and then blocked for 2 hours at room temperature with 250 μl blocking buffer (25 mmol/l HEPES [pH 7.4], 300 mmol/l NaCl, 0.05% Tween-20, 3% TopBlock [Juro AG, Lucerne, Switzerland]). Blocked wells were washed five times with 200 μl washing buffer, and then 100 μl diluted tumour membrane extracts or reference material was added to the wells and incubated overnight at 4°C.
As a reference for each assay, an extract of MCF-7 cells, prepared as described above, was used. For use in the assay, MCF-7 cell extracts were sequentially diluted with sample dilution buffer (blocking buffer, proteinase inhibitor cocktail, NaF and Na3VO4) at ratios of 1×, 0.75×, 0.5×, 0.25×, 0.125× and 0.025×, and then 100 μl aliquots were incubated on each microtitre plate, together with tumour tissue extracts and negative controls (containing only dilution buffer). After incubation of the samples and reference material, wells were washed five times with 200 μl washing buffer at room temperature to eliminate unbound particles. Biotinylated detection antibody was added, followed by incubation for 2 hours at room temperature. Complexes were detected with horseradish peroxidase-conjugated streptavidin, diluted in conjugate diluents for 1 hour at room temperature. Horseradish peroxidase activity was detected using SuperSignal WestPico substrate (Pierce, Rockford, IL, USA) in a glow luminometer. The response data for diluted reference material was fitted, and the respective curve was used for the quantification of tumour extracts. The value of undiluted MCF-7 extracts was denominated as 1 U/ml.
Quantitative real-time RT-PCR for the detection of proliferation markers
RNA was extracted using RNeasy kit (Qiagen, Hilden, Germany). Quality and quantity were checked using a Bioanalyzer 2100 (Agilent, Palo Alto, CA, USA). All genes were examined using SYBR Green I methods with Taqman 7000 (Applied-Biosystems, Foster City, CA, USA). Relative quantification (ΔΔCt) was obtained by normalization with ribosomal 18S and a standardization step with Human Universal Standard RNA (Stratagene, La Jolla, CA, USA). Quantitative real-time RT-PCR results were expressed in arbitrary units of reverse transcribed RNA (U/μg rt-RNA).
Statistical methods
The statistical significance of the association between P-Akt and other dichotomous variables (e.g. node status) was assessed using Mann–Whitney U-test. Spearman rank correlation (rs) was calculated to assess associations between continuous markers (e.g. ErbB-2 or tumour size and P-Akt protein expression levels). The continuous variable function of CLISA-determined P-Akt values was first tested for prognostic significance by univariate Cox regression. A cutoff or prognostic threshold value with respect to relapse-free survival was sought by means of classification and regression tree analysis [24,25]. Survival probabilities were calculated using the Kaplan–Meier method and compared by means of log-rank analysis [26]. The Cox proportional hazards regression model was also applied over multivariate analyses, with the associated likelihood ratio test used to assess test-of-trend differences. The results of multivariate Cox regression analysis were summarized in a table and expressed as relative risk for relapse.
Results
Distribution of phosphorylated Akt levels and its correlation with tumour characteristics
CLISA quantified P-Akt levels have a left-tailed distribution ranging from 0 to 1.08 U/mg total protein, with a median of 0.17 U/mg (mean 0.19 U/mg; Fig. 1) and could be transformed to normality by means of the 10th root. There was no correlation between P-Akt and ErbB-2 protein expression levels. In this set of primary breast cancer samples, we did not find any significant difference in P-Akt levels with respect to nodal status, tumour size, ER status or grading, nor any correlation between P-Akt levels and the continuous variables tumour size and ER level.
Prognostic significance of phosphorylated Akt levels
The prognostic value of P-Akt was investigated with respect to disease-free survival (DFS) in the patients overall (Fig. 2). Univariate Cox regression revealed a weak correlation between P-Akt levels and DFS (P < 0.05; likelihood ratio test). An optimal cutoff value for P-Akt (0.3 U/mg) was calculated using classification and regression tree analysis, dividing the patients into two subgroups: 21 patients (14%) patients expressed high levels of P-Akt (>0.31 U/mg total protein) and 135 patients (86%) expressed low levels of P-Akt. Subsequently, Kaplan–Meier survival curves stratified according to low and high P-Akt levels were plotted (Fig. 2). Sixty-seven per cent of patients (14 out of the 21) with high P-Akt levels relapsed, whereas only 36% (49 out of 135) with low P-Akt developed a relapse of disease within the period of observation (P < 0.01; log-rank test). The 5-year DFS was 33% in the high P-Akt group versus 60% in the low P-Akt group. The 5-year DFS in node-positive patients was 50% versus 68% in node-negative patients (P < 0.05; curves not shown).
Multivariate Cox analysis was performed including P-Akt and those additional variables that were found to have significant prognostic value in univariate Cox models (ER, ErbB-2 and node status, and tumour size and grading). In the tested multivariate model CLISA-determined elevated P-Akt level was an independent prognostic factor (P = 0.02), with a relative risk for breast cancer relapse of 2.09 (Table 2).
Prognostic significance of phosphorylated Akt in ErbB2-overexpressing tumours
Although no correlation was found between P-Akt and ErbB-2 expression, the prognostic impact of P-Akt was greater in ErbB2-overexpressing tumours than in the samples overall. As shown in the Kaplan–Meier curves in Fig. 3a, patient prognosis decreased significantly when tumours expressed P-Akt levels higher than the median value (P = 0.005). This effect was even more pronounced when P-Akt levels exceeded the third quartile value (P < 0.001), which, together with the multivariate Cox-analysis, indicates that P-Akt has independent and additive prognostic value in combination with ErbB-2 (Fig. 3b).
Correlation of phosphorylated Akt levels and mRNA expression of proliferation markers
Because involvement of P-Akt has been implicated in proliferation and apoptosis, we compared the quantitative P-Akt protein levels with the quantitative mRNA expression levels of genes involved in these biological processes. Using Spearman rank correlation, P-Akt levels were found to correlate with thymidylate synthase expression levels (rs = 0.38; P < 0.001) and, to a lesser extent, with expression levels of thymidine kinase 1, survivin, topoisomerase IIα and the E2F transcription factor (Fig. 4, Table 3).
Discussion
Correlations between elevated P-Akt and higher risk for relapsehas already been demonstrated by other investigators in certain subsets of patients, specifically patients who received adjuvant endocrine therapy [17], patients treated with radiotherapy [18] and patients with a node-negative disease [19]. Because ErbB-2 has been implicated in the activation of Akt [27], we investigated the association between P-Akt and ErbB-2 and its prognostic significance in tumours with known ErbB-2 expression levels. Our investigation re-confirmed the prognostic value of elevated P-Akt levels, and demonstrated that P-Akt expression levels are independent of other prognostic parameters, such as tumour size, grading, and node, ER and ErbB-2 status.
The lack of correlation between protein levels of ErbB-2 and P-Akt may be explained by the fact that Akt is also activated by various receptor tyrosine kinases [5-9], and by G-protein-coupled receptors [28]. Additionally, it was also observed that loss of PTEN activity is frequent in breast cancer and accompanied by increased activation of Akt [29], confirming that Akt can be activated by stimuli other than ErbB-2. The prognostic significance of P-Akt levels is increased if combined with ErbB-2 overexpression, suggesting that coactivation of Akt and ErbB-2 may have a synergistic clinical impact.
Our study is the first report on P-Akt assessed by EIA using a phospho-specific antibody in breast cancer cytosols of cryopreserved tumour samples; the technique allowed us to obtain precise and quantitative results (for review [30]). In contrast to semiquantitative immunohistochemistry data, tumour marker profiles assessed by quantitative EIA are more sensitive and reproducible. EIA tests conducted with fresh frozen tissue extracts avoid the potential antigen damage due to formalin fixation, paraffin embedding and uncontrolled storage. Furthermore, the two-site (sandwich) CLISA assay used in this investigation ensures increased specificity as compared with single-antibody assays, such as immunohistochemistry and western blotting. In addition, chemiluminometric detection guarantees high sensitivity in the detection of antigen–antibody complex.
We assayed for P-Akt in total breast tumour lysates, and not in tissue samples obtained from microdissection, both because we wished to correlate the protein expression levels of ErbB-2 and P-Akt levels directly in cells extracted from human tumour samples, and because it has been demonstrated that the activation status of Akt varies considerably in tumours of the same histotype, but not between different histotypes of the same tumour [31]. The CLISA assay used in the study was based on homogenized samples, which can include some stromal and normal tissue cells. The STB tissue samples contained at least 60% tumour cells, as observed by the pathologist. In addition, samples were previously analyzed for ErbB-2, ER and PgR using both EIA assays, as well as immunohistochemistry and/or fluorescence in situ hybridization. Importantly, good correlation between the assays was observed [23], suggesting that homogenization of samples does not play a crucial role in the final result. As in other assays that measure phosphorylation levels, the role played by phosphatases should not be ignored. We used phosphatase inhibitors in all steps of CLISA, as well as sample dilution. There could be some degradation before P-Akt testing, but all samples were treated identically, and the study compared relative P-Akt levels among all tumours. Reference units (U) were used in order to establish a standard curve, but not to measure absolute P-Akt levels in separate samples.
Also of interest is the positive correlation between P-Akt and mRNA expression levels of tumour proliferation markers shown in the present study. Akt is known to promote cell cycle progression by modulating the expression [32] and stabilization of cyclin D1 [33], which in turn activates the E2F transcription factor. Our results also reveal a significant correlation of P-Akt with E2F-1 transcription factor expression levels, as well as with genes regulated by E2F, such as thymidylate synthase, thymidine kinase 1, survivin and topoisomerase IIα.
Conclusion
Using a highly sensitive and specific CLISA assay, we demonstrated that elevated P-Akt is a marker of poor prognosis (decreased DFS). The prognostic value of Akt phosphorylation is independent of other characteristics, including tumour size and grade, and node, ErbB-2 and ER status. In a subset of patients with ErbB-2 overexpressing tumours, we demonstrated that P-Akt levels are of particular prognostic significance. In addition, Akt phosphorylation correlated with elevated mRNA expression levels of tumour proliferation factors. Based on these findings, we suggest that P-Akt could play a predictive role with respect to Herceptin, topoisomerase IIα inhibitors and combination therapies using Akt inhibitors, which are currently in clinical trials and should primarily be assessed in patients with ErbB-2-overexpressing tumours.
Abbreviations
CLISA = chemiluminescence-linked immunoassay; DFS = disease-free survival; EIA = enzyme immunoassay; ER = oestrogen receptor; P-Akt = phosphorylated Akt; PgR = progesterone receptor; PIP3 = phosphoinositol-3-phosphate; PKB = protein kinase B; RT-PCR = reverse transcription polymerase chain reaction; STB = Stiftung Tumorbank Basel.
Competing interests
The author(s) delcare that they have no competing interests.
Authors' contributions
JC carried out the development of CLISA assays, took measurements in breast cancer samples, participated in raw data analysis, participated in statistical analysis and drafted the manuscript. PU performed the statistical analysis. VV and ML carried out the RNA extraction and quantitative RT-PCR. WK participated in designing the study and participated in the raw data analysis. EW coordinated clinicians, providing tumour samples. MM helped in finalizing the manuscript. UE participated in designing the study and coordination, and helped to draft the manuscript. SE participated in coordinating the study and in statistical analysis, helped to draft the manuscript and coordinated the selection of samples from the STB. All authors read and approved the final manuscript.
Acknowledgements
This work was supported by a grant Nr. 31-059819.99/1 (U. Eppenberger) of the Swiss National Science Foundation and the Stiftung Tumorbank Basel (STB).
We thank Christine Wullschleger, Francoise David, Heidi Bodmer and Sabine Ehret for technical assistance, data management and tumour banking. We are indebted to A Almendral, M Anabitare, C Braschler, B von Castelberg, H Dieterich, D Fink, R Flury, R Gaudenz, K Lüscher, S Heinzl, M Mihatsch, H Moch, D Oertli, G Sauter, J Torhorst and M Zuber – clinicians and pathologists.
Figures and Tables
Figure 1 Chemiluminescence-linked immunoassay (CLISA)-quantified phosphorylated Akt (P-Akt) levels. (a) Histogram showing distribution of chemiluminescence-linked immunoassay (CLISA)-determined phosphorylated Akt (P-Akt) expression levels in 156 primary breast cancer samples. P-Akt levels ranged from 0 to 1.08 U/mg, with a median of 0.17 U/mg. (b) Scatter plot of P-Akt versus ErbB-2 expression levels. No correlation was found between the levels of P-Akt and ErbB-2.
Figure 2 Kaplan–Meier survival curves for patients overall. The curves are stratified by phosphorylated Akt (P-Akt) levels. Patients whose tumors express high levels of P-Akt exhibit a significantly worse outcome in terms of disease-free survival (DFS; P < 0.01).
Figure 3 Kaplan–Meier survival curves for the subset of patients with ErbB-2 overexpressing tumours. The curves stratified by (a) median and (b) last quartile values of phosphorylated Akt (P-Akt). Patients whose tumours express high levels of P-Akt exhibit a significantly worse outcome in terms of disease-free survival (DFS; P ≤ 0.005).
Figure 4 Scatter plot of phosphorylated Akt (P-Akt) versus thymidylate synthase (TS) mRNA expression. There is a good positive correlation (rs = 0.38; P < 0.001) between the two factors.
Table 1 Clinicopathological characteristics of the patients
Feature Number of patients (%)
Patients enrolled 156
Age (years):
<40 12 (8)
40–60 85 (54)
>60 59 (38)
Histology type:
Ductal 109 (70)
Lobular 17 (11)
Other 30 (19)
Tumour size:
T1 49 (31)
T2 90 (58)
T3-T4 17 (11)
Lymph-node status
Node negative 66 (42)
Node positive 90 (58)
Histopathological grade
I + II 57 (37)
III 86 (55)
Not analyzed 13 (8)
Oestrogen receptor
Positive (>20 fmol/mg) 116 (74)
Negative (≤ 20 fmol/mg) 40 (26)
Progesterone receptor
Positive (>20 fmol/mg) 85 (54)
Negative (≤ 20 fmol/mg) 71 (46)
Table 2 Univariate and multivariate Cox analysis of relapse-free survival in patients with primary breast cancer
Factor Univariate P Multivariate P Relative risk for relapse 95% CI
P-Akt 0.01 0.02 2.09 1.14–3.85
Node status 0.0003 0.09 1.33 0.95–1.85
ER status 0.03 0.17 0.67 0.38–1.19
ErbB-2 status 0.002 0.04 1.73 1.02–2.94
Grading 0.03 0.06 1.57 0.98–2.50
Tumour size 0.00005 0.02 1.51 1.07–2.13
CI, confidence interval; ER, oestrogen receptor; P-Akt, phosphorylated Akt.
Table 3 Spearman rank correlation of quantitative P-Akt levels and quantitative mRNA expression levels of proliferation markers
Proliferation marker rs P
Thymidylate synthase 0.38 <0.001
Thymidine kinase 1 0.23 <0.01
Survivin 0.22 <0.01
E2F 0.22 <0.01
Topoisomerase IIα 0.19 <0.05
rs, Spearman correlation coefficient.
==== Refs
Vivanco I Sawyers CL The phosphatidylinositol 3-kinase AKT pathway in human cancer Nat Rev Cancer 2002 2 489 501 12094235 10.1038/nrc839
Hanada M Feng J Hemmings BA Structure, regulation and function of PKB/AKT: a major therapeutic target Biochim Biophys Acta 2004 1697 3 16 15023346
Staal SP Molecular cloning of the akt oncogene and its human homologues AKT1 and AKT2: amplification of AKT1 in a primary human gastric adenocarcinoma Proc Natl Acad Sci USA 1987 84 5034 5037 3037531
Franke TF Kaplan DR Cantley LC PI3K: downstream AKTion blocks apoptosis Cell 1997 88 435 437 9038334 10.1016/S0092-8674(00)81883-8
Bowers DC Fan S Walter KA Abounader R Williams JA Rosen EM Laterra J Scatter factor/hepatocyte growth factor protects against cytotoxic death in human glioblastoma via phosphatidylinositol 3-kinase- and AKT-dependent pathways Cancer Res 2000 60 4277 4283 10945642
Hii CS Moghadammi N Dunbar A Ferrante A Activation of the phosphatidylinositol 3-kinase-Akt/protein kinase B signaling pathway in arachidonic acid-stimulated human myeloid and endothelial cells: involvement of the ErbB receptor family J Biol Chem 2001 276 27246 27255 11359783 10.1074/jbc.M103250200
Chen Y Li X Eswarakumar VP Seger R Lonai P Fibroblast growth factor (FGF) signaling through PI 3-kinase and Akt/PKB is required for embryoid body differentiation Oncogene 2000 19 3750 3756 10949929 10.1038/sj.onc.1203726
Duan C Liimatta MB Bottum OL Insulin-like growth factor (IGF)-I regulates IGF-binding protein-5 gene expression through the phosphatidylinositol 3-kinase, protein kinase B/Akt, and p70 S6 kinase signaling pathway J Biol Chem 1999 274 37147 37153 10601276 10.1074/jbc.274.52.37147
Franke TF Yang SI Chan TO Datta K Kazlauskas A Morrison DK Kaplan DR Tsichlis PN The protein kinase encoded by the Akt proto-oncogene is a target of the PDGF-activated phosphatidylinositol 3-kinase Cell 1995 81 727 736 7774014 10.1016/0092-8674(95)90534-0
Cantley LC Neel BG New insights into tumor suppression: PTEN suppresses tumor formation by restraining the phosphoinositide 3-kinase/AKT pathway Proc Natl Acad Sci USA 1999 96 4240 4245 10200246 10.1073/pnas.96.8.4240
Miwa W Yasuda J Murakami Y Yashima K Sugano K Sekine T Kono A Egawa S Yamaguchi K Hayashizaki Y Sekiya T Isolation of DNA sequences amplified at chromosome 19q13.1-q13.2 including the AKT2 locus in human pancreatic cancer Biochem Biophys Res Commun 1996 225 968 974 8780719 10.1006/bbrc.1996.1280
Cheng JQ Godwin AK Bellacosa A Taguchi T Franke TF Hamilton TC Tsichlis PN Testa JR AKT2, a putative oncogene encoding a member of a subfamily of protein-serine/threonine kinases, is amplified in human ovarian carcinomas Proc Natl Acad Sci USA 1992 89 9267 9271 1409633
Bellacosa A de Feo D Godwin AK Bell DW Cheng JQ Altomare DA Wan M Dubeau L Scambia G Masciullo V Molecular alterations of the AKT2 oncogene in ovarian and breast carcinomas Int J Cancer 1995 64 280 285 7657393
Roy HK Olusola BF Clemens DL Karolski WJ Ratashak A Lynch HT Smyrk TC AKT proto-oncogene overexpression is an early event during sporadic colon carcinogenesis Carcinogenesis 2002 23 201 205 11756242 10.1093/carcin/23.1.201
Ruggeri BA Huang L Wood M Cheng JQ Testa JR Amplification and overexpression of the AKT2 oncogene in a subset of human pancreatic ductal adenocarcinomas Mol Carcinog 1998 21 81 86 9496907
Bacus SS Altomare DA Lyass L Chin DM Farrell MP Gurova K Gudkov A Testa JR AKT2 is frequently upregulated in HER-2/neu-positive breast cancers and may contribute to tumor aggressiveness by enhancing cell survival Oncogene 2002 21 3532 3540 12032855 10.1038/sj.onc.1205438
Perez-Tenorio G Stal O Activation of AKT/PKB in breast cancer predicts a worse outcome among endocrine treated patients Br J Cancer 2002 86 540 545 11870534 10.1038/sj.bjc.6600126
Stal O Perez-Tenorio G Akerberg L Olsson B Nordenskjold B Skoog L Rutqvist LE Akt kinases in breast cancer and the results of adjuvant therapy Breast Cancer Res 2003 5 R37 R44 12631397 10.1186/bcr569
Schmitz KJ Otterbach F Callies R Levkau B Holscher M Hoffmann O Grabellus F Kimmig R Schmid KW Baba HA Prognostic relevance of activated Akt kinase in node-negative breast cancer: a clinicopathological study of 99 cases Mod Pathol 2004 17 15 21 14631376
Malik SN Brattain M Ghosh PM Troyer DA Prihoda T Bedolla R Kreisberg JI Immunohistochemical demonstration of phospho-Akt in high Gleason grade prostate cancer Clin Cancer Res 2002 8 1168 1171 11948129
Liao Y Grobholz R Abel U Trojan L Michel MS Angel P Mayer D Increase of AKT/PKB expression correlates with gleason pattern in human prostate cancer Int J Cancer 2003 107 676 680 14520710 10.1002/ijc.11471
Yamamoto S Tomita Y Hoshida Y Morooka T Nagano H Dono K Umeshita K Sakon M Ishikawa O Ohigashi H Nakamori S Monden M Aozasa K Prognostic significance of activated Akt expression in pancreatic ductal adenocarcinoma Clin Cancer Res 2004 10 2846 2850 15102693
Eppenberger-Castori S Kueng W Benz C Caduff R Varga Z Bannwart F Fink D Dieterich H Hohl M Muller H Paris K Schoumacher F Eppenberger U Prognostic and predictive significance of ErbB-2 breast tumor levels measured by enzyme immunoassay J Clin Oncol 2001 19 645 656 11157014
Chambers JM Hastie TJ Statistical Models in S 1971 London: Chapman & Hall 414
Breimann L Friedman JH Olsen RA Stone CJ Classification and Regression Trees 1984 Wadsworth: Belmont
Mantel N Evaluation of survival data and the two new rank order statistics arising its consideration Cancer Chemother Rep 1966 50 163 170 5910392
Hellyer NJ Kim MS Koland JG Heregulin-dependent activation of phosphoinositide 3-kinase and Akt via the ErbB2/ErbB3 co-receptor J Biol Chem 2001 276 42153 42161 11546794 10.1074/jbc.M102079200
Murga C Laguinge L Wetzker R Cuadrado A Gutkind JS Activation of Akt/protein kinase B by G protein-coupled receptors. A role for alpha and beta gamma subunits of heterotrimeric G proteins acting through phosphatidylinositol-3-OH kinase gamma J Biol Chem 1998 273 19080 19085 9668091 10.1074/jbc.273.30.19080
Mills GB Lu Y Fang X Wang H Eder A Mao M Swaby R Cheng KW Stokoe D Siminovitch K Jaffe R Gray J The role of genetic abnormalities of PTEN and the phosphatidylinositol 3-kinase pathway in breast and ovarian tumorigenesis, prognosis, and therapy Semin Oncol 2001 125 141 11706404 10.1053/sonc.2001.28554
Ross JS Fletcher JA HER-2/neu (c-erb-B2) gene and protein in breast cancer Am J Clin Pathol 1999 S53 S67 10396301
Wulfkuhle JD Aquino JA Calvert VS Fishman DA Coukos G Liotta LA Petricoin EF 3rd Signal pathway profiling of ovarian cancer from human tissue specimens using reverse-phase protein microarrays Proteomics 2003 3 2085 2090 14595806 10.1002/pmic.200300591
Muise-Helmericks RC Grimes HL Bellacosa A Malstrom SE Tsichlis PN Rosen N Cyclin D expression is controlled post-transcriptionally via a phosphatidylinositol 3-kinase/Akt-dependent pathway J Biol Chem 1998 273 29864 29872 9792703 10.1074/jbc.273.45.29864
Graff JR Konicek BW McNulty AM Wang Z Houck K Allen S Paul JD Hbaiu A Goode RG Sandusky GE Increased AKT activity contributes to prostate cancer progression by dramatically accelerating prostate tumor growth and diminishing p27Kip1 expression J Biol Chem 2000 275 24500 24505 10827191 10.1074/jbc.M003145200
| 15987444 | PMC1175052 | CC BY | 2021-01-04 16:04:33 | no | Breast Cancer Res. 2005 Mar 24; 7(4):R394-R401 | utf-8 | Breast Cancer Res | 2,005 | 10.1186/bcr1015 | oa_comm |
==== Front
Breast Cancer ResBreast Cancer Research1465-54111465-542XBioMed Central London bcr10191598744710.1186/bcr1019Research ArticleMechanisms underlying the growth inhibitory effects of the cyclo-oxygenase-2 inhibitor celecoxib in human breast cancer cells Basu Gargi D [email protected] Latha B [email protected] Teresa L [email protected] Sandra J [email protected] Pinku [email protected] Department of Biochemistry and Molecular Biology, Mayo Clinic College of Medicine, Scottsdale, Arizona, USA2005 4 4 2005 7 4 R422 R435 1 10 2004 14 12 2004 1 3 2005 4 3 2005 Copyright © 2005 Basu et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Introduction
Inhibitors of cyclo-oxygenase (COX)-2 are being extensively studied as anticancer agents. In the present study we evaluated the mechanisms by which a highly selective COX-2 inhibitor, celecoxib, affects tumor growth of two differentially invasive human breast cancer cell lines.
Methods
MDA-MB-231 (highly invasive) and MDA-MB-468 (moderately invasive) cell lines were treated with varying concentrations of celecoxib in vitro, and the effects of this agent on cell growth and angiogenesis were monitored by evaluating cell proliferation, apoptosis, cell cycle arrest, and vasculogenic mimicry. The in vitro results of MDA-MB-231 cell line were further confirmed in vivo in a mouse xenograft model.
Results
The highly invasive MDA-MB-231 cells express higher levels of COX-2 than do the less invasive MDA-MB-468 cells. Celecoxib treatment inhibited COX-2 activity, indicated by prostaglandin E2 secretion, and caused significant growth arrest in both breast cancer cell lines. In the highly invasive MDA-MB-231 cells, the mechanism of celecoxib-induced growth arrest was by induction of apoptosis, associated with reduced activation of protein kinase B/Akt, and subsequent activation of caspases 3 and 7. In the less invasive MDA-MB-468 cells, growth arrest was a consequence of cell cycle arrest at the G0/G1 checkpoint. Celecoxib-induced growth inhibition was reversed by addition of exogenous prostaglandin E2 in MDA-MB-468 cells but not in MDA-MB-231 cells. Furthermore, MDA-MB-468 cells formed significantly fewer extracellular matrix associated microvascular channels in vitro than did the high COX-2 expressing MDA-MB-231 cells. Celecoxib treatment not only inhibited cell growth and vascular channel formation but also reduced vascular endothelial growth factor levels. The in vitro findings corroborated in vivo data from a mouse xenograft model in which daily administration of celecoxib significantly reduced tumor growth of MDA-MB-231 cells, which was associated with reduced vascularization and increased necrosis in the tumor mass.
Conclusion
The disparate molecular mechanisms of celecoxib-induced growth inhibition in human breast cancer cells depends upon the level of COX-2 expression and the invasive potential of the cell lines examined. Data suggest a role for COX-2 not only in the growth of cancer cells but also in activating the angiogenic pathway through regulating levels of vascular endothelial growth factor.
==== Body
Introduction
The incidence of breast cancer – the second leading cause of cancer death in women in the USA – is increasing, and current therapy is unable to achieve clinical responses in patients with highly invasive metastatic disease. There is a consequent need for more effective approaches to prevention and treatment of breast cancer. Nonsteroidal anti-inflammatory drugs (NSAIDs) show great promise in this respect. Recent data on regular NSAID use for 5–9 years indicated a 21% reduction in the incidence of breast cancer, and regular NSAID use for 10 or more years produced a 28% reduction in the incidence of breast cancer [1]. Preclinical studies [2-4] have consistently shown that NSAIDs inhibit mammary carcinogenesis.
Various mechanisms may be responsible for the observed effects of NSAIDs against breast cancer. Inhibition of cyclo-oxygenase (COX), particularly the COX-2 isozyme, and blockade of the prostaglandin (PG) cascade may have impacts on neoplastic growth and development by inhibiting several key features of mammary carcinogenesis – namely proliferation, angiogenesis and metastasis. Inhibition of COX also causes induction of apoptosis in malignant cells and enhances antineoplastic activity of cytotoxic T lymphocytes [5-8]. Our study conducted in newly diagnosed stage I and stage II breast cancer patients [9] showed impaired functionality of T cells and dendritic cells, which correlated with COX-2 overexpression in the tumors and increased levels of PGE2 in the serum and tumor milieu. Therefore, a convincing case has been made for COX-2 being an important target for the antineoplastic action of NSAIDs. Unlike NSAIDs, COX-2 selective inhibitors such as celecoxib and rofecoxib do not inhibit COX-1 and thus show promise as drugs that spare the gastrointestinal system.
COX-2 is overexpressed in breast cancer tissues, and greater extent of its expression is associated with poorer prognosis [10]. Various environmental and nutritional risk factors induce COX-2 expression in animal models of breast cancer [11,12]. Moreover, COX-2 selective inhibitors significantly delayed the incidence of mammary tumors in transgenic mice expressing the Her2/Neu, and polyoma-middle T oncogenes [13,14]. Recently, a transgenic mouse model was developed in which the human COX-2 gene was expressed in the mammary gland under the control of the murine mammary tumor virus promoter [15]. That study demonstrated that enhanced COX-2 expression strongly predisposes to transformation of the mammary gland in multiparous animals. These data strongly suggest that local expression of COX-2 is sufficient for in situ tumor initiation and/or progression. Another transgenic overexpression study with COX-2 targeted to the epidermis also supports the concept that COX-2 is a critical regulator of tumor progression [16]. Transfections of the breast cancer cell line Hs578T with cDNA for COX-2 led to an increase in expression and activity of matrix metalloproteinase-2, resulting in increasingly invasive behavior of the cells [17]. COX-2 specific inhibitors have the ability to block cell growth, and induce apoptosis and cell cycle arrest in murine mammary tumor cell lines [18]. However, the molecular mechanisms involved are not well understood. If COX-2 inhibitors act only by modulating COX-2 expression, then that would imply that this therapy would be limited to COX-2 overexpressing tumors; hence, this question is of considerable clinical importance.
In the present study we established that the level of COX-2 expression and the invasive property of breast cancer cells determines the mechanism of celecoxib-induced growth inhibition; that COX-2 is involved in extracellular matrix associated microvascular channel formation by breast cancer cells; and that COX-2 inhibits angiogenesis in vivo. The study should further our understanding of the cellular and molecular mechanisms underlying the chemopreventive effect of a COX-2 selective inhibitor in breast cancer. To the best of our knowledge, this is the first study demonstrating the diverse mode of action of celecoxib on human breast cancer cells, which may be dependent upon the cells' invasive properties and levels of COX-2 expression. This is also the first report suggesting a direct role for COX-2 in matrix associated microvascular channel formation by breast cancer cells.
Methods
Cell culture
The human breast cancer cell lines MDA-MB-231 and MDA-MB-468 were obtained from the American Type Culture Collection (ATCC; Rockville, MD, USA) and cultured following instructions from the ATCC. Briefly, cells were grown in Dulbecco's modified eagle medium (DMEM; GIBCO-BRL, Rockville, MD, USA) supplemented with 5% fetal calf serum (FCS), 100 U penicillin, 0.1 μg streptomycin and 2 mmol/l L-glutamax. Cells were maintained in log phase in 37°C incubator with 10% carbon dioxide. For each experiment cells were plated in FCS-containing media in 58 cm2 culture dishes at a cell density of approximately 1 × 106 cells/dish and incubated for another 48 hours. Cell cultures were treated with increasing concentrations of celecoxib (20–60 μmol/l; Pfizer, New York, NY, USA) and with dimethyl sulfoxide (DMSO; the vehicle in which celecoxib was dissolved). The concentration of celecoxib used in our experiments is clinically relevant because the serum concentrations of COX-2 inhibitors in patients range from 20 to 100 μmol/l [19]. The concentrations used in the study are based on our titrations with celecoxib for the two cell lines and from several published references on other cell lines [20-22]. In both the cell lines tested there was no evidence of apoptosis or cell cycle arrest at concentrations below 20 μmol/l.
SDS-PAGE immunoblotting
Following harvesting of adherent cells by scraping, cell lysates were prepared and quantified by BCA assay. Lysates (100 μg) were resolved on a 10–15% acrylamide gel and electroblotted onto immobilon-P polyvinylidene diflouride membranes (Sigma, St. Louis, MO, USA). These were probed with primary antibodies for COX-2 (p66), BAX (p23), Bcl-2 (p26), and vascular endothelial growth factor (VEGF; p20), all from Santa Cruz Biotechnology Inc. (Santa Cruz, CA, USA), and phosphorylated Akt (pAkt; p60; Cell Signaling, Beverly MA, USA), and then probed with the appropriate secondary antibodies. Bound antibodies were detected using an enhanced chemiluminescence detection kit (SuperSignal West Dura, Pierce, Rockford, IL, USA), and developed on high performance chemiluminescence films (Amersham Pharmacia Biotech, Piscataway, NJ, USA).
Proliferation assay
Cell proliferation was determined by using [3H]thymidine incorporation, in which 1 μCi of [3H]thymidine was added to the drug or vehicle treated cultures 16 hours before harvesting using a Packard Cell Harvester (Packard Biosciences, Shelton, CT, USA). Incorporated thymidine was evaluated using the Topcount micro-scintillation counter (Packard Biosciences). Results were expressed as [3H]thymidine uptake. All determinations were performed in triplicate. Proliferation is directly correlated to radioactive counts/min. In order to determine whether added PGE2 could counteract the growth inhibitory effect of celecoxib, we treated cells with celecoxib (40 μmol/l) and 12.5–200 pg/ml PGE2 and incubated them for 96 hours before determining [3H]thymidine incorporation, as mentioned above.
Assay for apoptosis
Following treatment of cells with celecoxib for 48 hours, apoptosis was determined by staining the cells with annexin V and propidium iodide (PI), in accordance with the manufacturer's instructions for use of the BD Pharmingen (San Diego, CA, USA) apoptosis kit. Briefly, an aliquot of 105 cells was incubated with annexin V–fluorescein isothiocyanate and PI for 15 min at room temperature in the dark. Cells were immediately analyzed by flow cytometry. Viable cells exclude PI and are negative for annexin V staining, whereas early apoptotic cells are annexin V positive and PI negative. Cells that are not viable due to apoptotic cell death stain positive for annexin V and PI. The percentage of stained cells in each quadrant was quantified using CellQuest software (BD Biosciences, San Jose, CA, USA) and the total number of apoptotic cells (both early and late apoptosis) was quantified.
Confocal microscopy for detection of apoptotic bodies
Cells were grown with celecoxib (60 μmol/l) for 48 hours and then trypsinized. Cells were resuspended in phosphate-buffered saline (PBS) with 0.1% bovine serumn albumin at a final concentration of 5 × 107 cells/ml and 2 μl of 5 mmol/l carboxyfluoroscein succinimidyl ester (CFSE)/ml (Molecular Probes, Eugene, OR, USA) was added. After 10 min of incubation at 37°C the staining was quenched by adding five times the volume of ice-cold PBS and excess stain was washed off by repeated washes in PBS. Cells were fixed in 95% ethanol for 1 hour on ice and resuspended in PBS containing 20 μg/ml PI (Sigma) and 15 μg/ml RNase A (Sigma). Images were captured on the LSM510 confocal microscope (Carl Zeiss Inc., Gottingen, Germany) using excitation wavelengths of 488 nm (for CFSE) and 543 nm (for PI).
Assay for caspases 3 and 7
To evaluate whether celecoxib treatment can induce activation of caspases 3 and 7, we detected levels of active forms of caspases 3 and 7 in cell lysates from treated and untreated cells using the EnzChek Caspase-3/7 Assay Kit (Molecular Probes), in accordance with the manufacturer's protocol. In principle, active caspase 3 or 7 cleaves a fluorogenic substrate; this releases the fluorochrome, which is detected using a spectrofluorometer.
Cell cycle analysis
Cells were treated with increasing concentrations (20–60 μmol/l) of celecoxib or DMSO (vehicle) in medium supplemented with 5% FCS for 48 hours. The adherent and the nonadherent cell fractions were harvested and cell pellets were fixed and permeabilized in 95% cold ethanol, and resuspended in PBS containing 20 μg/ml PI (Sigma) and 15 μg/ml RNase A (Sigma). Samples were incubated in the dark at 37°C for 30 min and analyzed by flow cytometry (Becton Dickinson, San Diego, CA, USA). For each sample, 50,000 fluorescent cells were counted. Data were analyzed using the ModFit software (Verity Software House Inc., Topsham, ME, USA) to determine DNA content and cell cycle phase (G0/G1–S–G2/M phase). Cell doublets and clumps were eliminated from the analyses by gating.
Prostaglandin E2 production
Cells were treated with increasing concentrations (20–60 μmol/l) celecoxib or DMSO (vehicle) in medium supplemented with 5% FCS for 48 hours. Levels of PGE2 released in media were measured using a PGE2 enzyme immunoassay kit from Cayman Chemical Co. (Ann Arbor, MI, USA). Medium was sampled, centrifuged to remove floating cells and frozen immediately at -70°C until assay. The PGE2 assay was performed in accordance with the manufacturer's instructions, following dilution to ensure that readings were within the limits of accurate detection by the assay. The results are expressed as pg PGE2/ml ± standard deviation.
Assay for vasculogenic mimicry
This assay was performed as described [23]. Cells were grown until they were about 80% confluent. The growth medium was replaced with serum-free DMEM supplemented with 100 μg/ml heparin (Elkins-Sinn, Inc. Cherry Hill, NJ, USA) and antibiotics, and cells were incubated for 24 additional hours. The cells were trypsinized, counted, and resuspended in media (at a concentration of 4 × 104 cells/ml) containing 40 and 60 μmol/l concentrations of celecoxib or vehicle. The wells of a 24-well tissue culture plate were evenly coated with 0.1 ml/well growth factor reduced Matrigel (BD Biosciences), which was allowed to solidify at 37°C for 30 min, in accordance with the manufacturer's instructions, before the cells were plated. The cell suspension was plated (1 ml/well) onto the surface of Matrigel and incubated at 37°C for 48 hours and photographed using a Nikon inverted phase contrast photomicroscope (Nikon USA, Garden City, NY, USA). Channel formation was quantified as percentage of channels formed by counting the number of connected cells in five randomly selected fields, using 200× magnification, and dividing the number by the total number of cells in the same field.
Xenografts
Male athymic nude mice (age 6–8 weeks) were obtained from NxGen Biosciences Inc. (San Diego, CA, USA) and animals were housed under specific pathogen-free conditions. Five mice/group were prophylactically treated with either celecoxib (25 mg/kg body weight) or vehicle DMSO for 7 days before the tumor cells were inoculated. MDA-MB-231 cells were harvested by centrifugation and 5 × 106 cells were suspended in 150 μl of serum free DMEM with an equal volume of cold liquid Matrigel (10 mg/ml). The suspension was injected subcutaneously in the mice. In order to determine the optimal cell number to be injected, titration with varying cell numbers was done on nude mice and the tumorigenicity of the cell line determined. The growth of these tumors was monitored by weekly examination, and growth rates were determined using caliper measurements. Tumor weight was calculated according to the following equation [24]: tumor weight (g) = (length (cm) × width (cm)2) × 0.5. Experiments were terminated 45 days after tumor cell injection. It was necessary to kill some of the mice earlier because of the aggressive nature of the tumor.
Histologic studies
All solid tumors resulting were excised and fixed in formaldehyde, and paraffin-embedded blocks was sectioned at a thickness of 7 μm. Histologic evaluation of vascularity was determined by Masson's trichrome staining [25]. This method stains fibrous tissue and stroma green. Blood vessels containing red blood cells stain bright red. Immunohistochemical localization of factor VIII related antigen on endothelial cells was determined using the polyclonal rabbit antihuman von Willebrand factor purchased from Dako Cytomation (Glostrup, Denmark), using the manufacturer's recommended staining protocol.
Statistical analysis
The celecoxib experiments were run in triplicate; the mean as well as standard deviations were computed. The means were then compared using one-way analysis of variance with Dunnett adjustment.
Results
Cyclo-oxygenase-2 protein is differentially expressed in breast cancer cell lines
We studied two human breast cancer cell lines, MDA-MB-231 and MDA-MB-468, for COX-2 expression by western blotting. Both cell lines expressed COX-2, although MDA-MB-468 cells exhibited lower protein expression than did MDA-MB-231 cells. Western blot analysis for COX-2 protein in the MDA-MB-231 cell line showed little change in protein expression after treatment with 20–40 μmol/l celecoxib. At the level of 60 μmol/l there was a slight increase in COX-2 expression. However, in the MDA-MB-468 cell line there was significant downregulation of COX-2 expression upon drug treatment (Fig. 1a).
Celecoxib inhibits growth and proliferation of breast cancer cell lines
Celecoxib at concentrations of 20, 40, and 60 μmol/l was used to treat the two cell lines for 48 hours. Under the phase contrast microscope, both cell lines exhibited a dramatic morphologic change as well as growth arrest after 48 hours of drug treatment (data not shown). The rate of proliferation in response to celecoxib treatment was assayed by measuring incorporation of [3H]thymidine uptake. Significant inhibition of proliferation was observed in both cell lines in a dose-dependent manner, in response to varying concentrations of celecoxib at 96 hours after treatment (P < 0.001; Fig. 1b). Similar growth inhibition was observed at earlier time points (48 and 72 hours after treatment (data not shown).
Celecoxib induces apoptosis in MDA-MB-231 but not in MDA-MB-468 cells
Because COX inhibitors have been reported to mediate apoptosis in many cells [26,27], we investigated whether the observed growth inhibition mediated by celecoxib was associated with induction of programmed cell death. Flow cytometric analysis of annexin V/PI staining in celecoxib-treated and vehicle-treated cells was used to analyze apoptosis. Following 48 hours of drug treatment, induction of apoptosis was observed in the MDA-MB-231 cells in a dose-dependent manner (Fig. 2a). Celecoxib at concentrations of 40 and 60 μmol/l caused significant increases in the percentage apoptotic cells (P = 0.01 and P < 0.001, respectively). In the MDA-MB-468 cell line apoptosis was not induced with celecoxib treatment (Fig. 2a). In spite of the lack of evidence of increased apoptosis, MDA-MB-468 cells had significantly lower proliferation after drug treatment (Fig. 1b). Treated cells appeared rounded up and exhibited atypical morphology (data not shown), which suggested that alterations in the adhesive properties of these cells might have occurred and other pathways may be involved in the growth inhibition observed in MDA-MB-468 cells.
Celecoxib induces formation of apoptotic bodies and loss of nuclear envelope integrity in MDA-MB-231 cells
To follow up on the celecoxib-induced apoptosis of the MDA-MB-231 cells, we analyzed morphological changes in MDA-MB-231 cells after celecoxib treatment using confocal microscopy. Celecoxib at concentrations of 40 and 60 μmol/l caused loss of integrity of nuclear envelope and induced formation of peripheral, sharply delineated masses of condensed chromatin or apoptotic bodies, which are characteristic structural features of apoptosis (Fig. 2b–e). Membrane blebbing was also observed, along with loss of plasma membrane integrity in some cells (data not shown). These results indicate that celecoxib treatment caused architectural changes in membrane and cell nucleus within 48 hours of treatment. No such changes were observed in MDA-MB-468 cells (data not shown), which correlated with our observation that there was no significant induction of apoptosis in these cells after celecoxib treatment (Fig. 2a).
Celecoxib inhibits activation of protein kinase B/Akt kinase in MDA-MB-231 cells
Protein kinase B, Akt, is a serine/threonine protein kinase that is involved in promoting cell survival signals through the phosphoinositide 3-kinase pathway, leading to inactivation of a series of proapoptotic proteins. Akt also represents a key signaling component in cell survival by mediating the activation of downstream effectors such as BAD [28,29] and procaspase-9 [30]. Celecoxib was recently shown to induce apoptosis of cancer cells by blocking Akt activation in rat cholangiocarcinoma and human prostate cancer cells in vitro [21,22]. To explore whether inhibition of Akt activation may be the mechanism responsible for induction of apoptosis in MDA-MB-231 cells, we determined the effect of celecoxib on phosphorylation of Akt on breast cancer cell lines. Breast cancer cells were exposed to varying doses of celecoxib for 48 hours, and Akt and pAkt in cell lysates were determined by western blot analysis. At a concentration of 20 μmol/l, celecoxib caused slight increase in pAkt in MDA-MB-231 cells. At a concentration of 60 μmol/l, celecoxib treatment significantly (P = 0.002) downregulated the level of phosphorylation of Akt in MDA-MB-231 cells but not in MDA-MB-468 cells (Fig. 3a,b), suggesting that the mechanism of apoptosis induction in MDA-MB-231 cells was, in part, dependent upon decreased phosphorylation of Akt protein. Because Akt represents a key signaling component in cell survival by activating downstream apoptotic proteins [28-31], we evaluated the levels of Bax and Bcl-2 by western blot analysis of lysates derived from both cell lines after celecoxib treatment. Treatment with celecoxib at concentrations of 40 and 60 μmol/l induced increased expression of Bax in the MDA-MB-231 cells (Fig. 3c), but no significant decrease in Bcl-2 was observed (data not shown). In MDA-MB-468 cells, in which apoptosis was not evident, levels of pAkt and Bax remained unchanged with treatment (Fig. 3a,b).
Celecoxib induces caspase-3/7 activation in MDA-MB-231 cells
Caspases are responsible for many of the biochemical and morphological changes that occur during apoptosis. Most apoptotic signals induce intracellular cleavage of caspases 3 and 7 from an inactive precursor (p32–p35) to the active forms (p17 and p12); hence, these proteins are the most extensively studied apoptotic proteins. The effector caspases 3 and 7 proteolytically cleave and activate several other caspases as well as several other apoptotic proteins, including the DNA fragmentation protein poly-ADP-ribose polymerase (PARP), which is one of the primary activators of DNA fragmentation and cell death [32-34].
We investigated whether celecoxib induced the activation of caspase 3 and caspase 7 in MDA-MB-231 cells in which apoptosis was induced. Caspase activity is presented as fluorescence emission, which is directly proportional to activities of caspases 3 and 7. Treatment with celecoxib (40 and 60 μmol/l) for 48 hours caused significant increases in activation of caspases 3 and 7 (fivefold increase at the 40 μmol/l concentration [P = 0.008] and sixfold increase at the 60 μmol/l concentration [P = 0.002]; Fig. 3d). Caspase activation was completely blocked by incubation with the caspase inhibitor Ac-DEVD-CHO (data not shown). These results suggest that celecoxib-induced apoptosis in MDA-MB-231 cells is due to activation of caspases 3 and 7, which is corroborated by studies indicating that the blockade or absence of caspase activation is sufficient to inhibit effective apoptosis [35]. In contrast, caspase activation was not observed in celecoxib-treated MDA-MB-468 cells, which correlated with no significant increase in apoptosis with celecoxib treatment (Fig. 2).
Celecoxib induces cell cycle arrest at the G0/G1 checkpoint in MDA-MB-468 but not in MDA-MB-231 cells
To determine whether celecoxib-induced growth inhibition was due to changes in cell cycle progression, flow cytometric analysis was performed on cells treated with increasing concentrations of celecoxib (20–60 μmol/l) for 48 hours. In MDA-MB-468 cells, in which celecoxib did not induce apoptosis, there was induction of cell cycle arrest. At 40 and 60 μmol/l concentrations of celecoxib, significant increases (P = 0.02 and P < 0.001, respectively) in the proportion of cells that were arrested at the G0/G1 checkpoint of the cell cycle were observed. Subsequently, significant inhibition of transition to the G2/M phase (P = 0.02 at 60 μmol/l) and S phase (P = 0.006 and P < 0.001 at 40 and 60 μmol/l) was observed (Fig. 4b). Thus, growth inhibitory activity of celecoxib on these MDA-MB-468 cells was due to cell cycle arrest at G0/G1 phase and not due to induction of apoptosis. The cell cycle arrest persisted at 72 hours after drug treatment (data not shown). In MDA-MB-231 cells there was no significant difference in cell cycle progression with celecoxib treatment for 48 hours (Fig. 4a).
Celecoxib inhibits cyclo-oxygenase-2 induced prostaglandin E2 production in both cell lines
COX-2 converts arachidonic acid to bioactive prostanoids. It has been demonstrated that COX-2 derived PGE2 is the major prostaglandin produced by breast cancer cells [36]. To determine whether COX-2 activity was affected by celecoxib treatment, PGE2 production using a PGE2-specific enzyme-linked immunosorbent assay was measured in conditioned medium collected from the breast cancer cell lines after celecoxib-treatment (20–60 μmol/l) for 48 hours. All doses of celecoxib significantly reduced PGE2 secretion by both cell lines (P < 0.01 for MDA-MB-231 and P = 0.03, 0.02 and 0.01 for MDA-MB-468 cells; Table 1), indicating that celecoxib is a potent inhibitor of COX-2 induced PGE2 production.
Celecoxib-induced growth inhibition is reversed by exogenous prostaglandin E2 only in MDA-MB-468 cells
Because celecoxib caused growth inhibition in the two breast cancer cell lines and inhibited PGE2 secretion, we hypothesized that this growth inhibition was PGE2 dependent. To determine whether celecoxib-induced growth inhibition could be reversed by exogenous PGE2, PGE2 was added to cultures of MDA-MB-231 and MDA-MB-468 cells treated with constant dose (40 μmol/l) of celecoxib. Varying amounts of PGE2 (12.5–200 pg/ml) were added to the medium in order to take into account the fact that some of the PGE2 may degrade or be internalized into cells. In MDA-MB-231 cells, growth inhibition induced by 40 μmol/l celecoxib could not be restored by addition of exogenous PGE2 (Fig. 5a), thereby suggesting that celecoxib-induced growth inhibition in MDA-MB-231 cells may be independent of PGE2 levels. However, addition of 200 pg/ml PGE2 completely reversed the growth inhibition induced by 40 μmol/l celecoxib in the less invasive MDA-MB-468 cells (Fig. 5b), suggesting that celecoxib-induced growth regulation of these cell lines may be dependent on the levels of PGE2.
Celecoxib inhibits in vitro matrix-associated vascular channel formation
Recent findings demonstrate the unusual ability of aggressive human breast cancer cells to form tubular structures in three-dimensional Matrigel cultures. The generation of these channels by epithelial tumor cells is called vascular mimicry [37-39]. One study [40] suggested a connection between angiogenesis and formation of these channels. Because celecoxib is known to act as an inhibitor of angiogenesis, we investigated the ability of MDA-MB-231 and MDA-MB-468 cells to form the microvascular channels with and without celecoxib treatment. MDA-MB-231 cells, which express elevated levels of COX-2 and are highly invasive, begin to form tubular structures in under 16 hours when plated on Matrigel (data not shown) and form very characterized microvascular channels by 48 hours. In contrast, MDA-MB-468 cells, which have lower COX-2 and are less invasive, start tubule formation much later, at approximately 30 hours, and exhibit significantly fewer microvascular channels at 48 hours than do MDA-MB-231 cells. These observations were specific for the high or moderately invasive cells, because the noninvasive breast cancer cells (ZR-75-1) did not form channels in vitro under identical culture conditions (data not shown).
We found that celecoxib treatment at concentrations of 40 and 60 μmol/l was able to reduce significantly the formation of channels in both breast cancer cell lines in a dose-dependent manner, as compared with vehicle treated cells (P < 0.001 for MDA-MB-231 cells and P < 0.001 for MDA-MB-468 cells; Fig. 6a), suggesting a role for COX-2 in channel formation. The effect of celecoxib on channel formation was only quantified on live adherent cells in Matrigel as the apoptosed and dead cells float into the media. Thus, we believe that the negative effect of celecoxib on channel formation was not due to cell death, which was also measured by trypan blue exclusion (data not shown).
Celecoxib inhibits expression of vascular endothelial growth factor protein in MDA-MB-231 cells
Recent reports have shown that a nonspecific COX inhibitor (indomethacin) suppresses the expression of VEGF gene expression in vitro in mammary tumor cells [41]. We evaluated the levels of VEGF protein from tumor lysate of cells treated with vehicle or increasing doses of celecoxib. Compared with control, celecoxib (20–60 μmol/l) treatment reduced expression of VEGF in the MDA-MB-231 cells in a dose-dependent manner (Fig. 6b). No such reduction was observed in the MDA-MB-468 cells treated with celecoxib (Fig. 6b), suggesting that in the highly aggressive MDA-MB-231 cells the COX-2/PGE2 pathway may play a critical role in channel formation and angiogenesis, in part by activating proangiogenic proteins such as VEGF. Future studies will evaluate other proteins associated with the angiogenic pathway.
In vivo tumor growth was reduced with celecoxib treatment
Nude mice were prophylactically treated with celecoxib or vehicle for 1 week before tumor challenge with MDA-MB-231 cells in Matrigel. Celecoxib treatment was continued for 45 days after tumor challenge. Mice treated with celecoxib (25 mg/kg body weight) exhibited significant (P = 0.01) reduction in tumor growth as compared with vehicle-treated mice without evidence of systemic toxicity (Fig. 7a). A representative mouse from each treatment group is shown in Fig. 7b; the treated mouse has reduced tumor mass compared with the control mouse.
In vivo inhibition of angiogenesis and increase in necrosis with celecoxib treatment
Vascularity of tumor implants was histologically evaluated using Masson's trichrome and factor VIII-related antigen staining. Tumors from celecoxib-treated mice showed reduced blood vessels as compared with tumors excised from vehicle-treated mice (Fig. 8). Furthermore, there was evidence of necrosis in the celecoxib-treated tumors relative to those obtained from vehicle-treated animals (Fig. 8a,b).
Discussion
The results presented here clearly show that celecoxib strongly suppresses cell growth and proliferation in both human breast cancer cell lines (Fig. 1b). However, the mechanism of antitumor effect is dependent upon COX-2 expression and the invasive properties of the cancer cell. The highly invasive MDA-MB-231 cells undergo induction of apoptosis (Fig. 2) and the less invasive MDA-MB-468 cells undergo cell cycle arrest (Fig. 4) after treatment with celecoxib. The two cell lines exhibit different levels of COX-2 protein expression, with MDA-MB-231 cells expressing much higher levels than MDA-MB-468 cells (Fig. 1a), which directly correlated with the amount of PGE2 production by the cells (Table 1) and their invasive properties. Our data are in good agreement with the postulate that elevated production of COX-2-induced prostanoids is a hallmark of highly metastasizing breast cancer cells [41,42]. The two cell lines regulate COX-2 protein differently after celecoxib treatment, with downregulation of the protein observed in MDA-MB-468 cells but not in MDA-MB-231 cells (Fig. 1a). In fact there was an increase in COX-2 expression in MDA-MB-231 cells at the 60 μmol/l level of celecoxib, the mechanism for which is not known. However, one or more COX-produced products may repress COX expression in a negative feedback loop. Removal of negative feedback by celecoxib treatment would result in COX-2 induction. There are similar reports on celecoxib treatment leading to strong upregulation of COX-2 protein expression in 184htert breast cancer cells [43].
Regardless of COX-2 expression and regulation patterns, celecoxib treatment reduced PGE2 secretion by both cell lines (Table 1), but provision of exogenous PGE2 reversed celecoxib-induced growth inhibition in the MDA-MB-468 cells only, and not in the MDA-MB-231 cells (Fig. 5). This suggests that celecoxib-induced growth inhibition of the highly aggressive MDA-MB-231 cells is independent of PGE2. Corroborating our findings are previous reports that growth inhibition induced by COX-2 inhibitors in some carcinoma cell lines can be completely abrogated by exogenous addition of PGE2 [44], whereas in other studies addition of PGE2 had no effect [45,46]. One possible PGE2-independent mechanism by which celecoxib may have caused apoptosis in MDA-MB-231 cell lines could be through the accumulation of the prostaglandin precursor arachidonic acid. Arachidonic acid is known to be converted to an intermediate, apoptosis-signaling compound, namely ceramide, which causes NSAID-induced apoptosis in cancer cells [47]. This phenomenon of ceramide-induced apoptosis has been proven in a murine mammary tumor cell line treated with celecoxib [18]. Because PGE2 is the major prostanoid released from breast cancer cells [41], we focused our studies on PGE2 levels. However, a possible role of other prostanoids such as PGD2, PGI, PGF2α and thromboxane2 cannot be ruled out, and future studies will include analyses of other prostanoids.
Thus, we observed that the mechanisms driving celecoxib-induced growth inhibition are very diverse in the two cells lines, depending upon COX-2 expression levels, invasive properties, and dependence on PGE2. At the cellular level, celecoxib induced the characteristic features of apoptosis in the MDA-MB-231 cells (Fig. 2). At the molecular level, activation of protein kinase B/Akt was significantly reduced at 60 μmol/l concentration of celecoxib, with increased activation of proapoptotic protein Bax and caspases 3 and 7 (Fig. 3). These results are in agreement with those of other studies in which it was suggested that activation of effector caspases 3 and 7 and Bax proteins, downstream of phosphoinositide 3-kinase/Akt inactivation, was the mechanism of celecoxib-induced tumor cell apoptosis [22,48]. Mechanisms leading to the downregulation of Akt activation are not clear. It has been suggested that inhibition of the tumor suppressor PTEN, a phosphatase that targets phosphoinositol triphosphate, or inhibition of 3-phosphoinositide-dependent kinase 1 activity may be involved [48-50].
In contrast to MDA-MB-231 cells, growth of MDA-MB-468 cells was inhibited by induction of cell cycle arrest at the G0/G1 phase of the cell cycle (Fig. 4). Similar cell cycle arrest has been reported using a murine mammary tumor cell line derived from a spontaneously occurring tumor [18], human pancreatic cancer cell lines [51], and human ovarian cancer cell lines [52]. It is not clear from our studies that celecoxib directly affects cell cycle distribution by regulating cyclin D1 levels, which is one of the major cyclins known to be upregulated during cancer. Preliminary data evaluating cyclin D1 levels in MDA-MB-468 cells after celecoxib treatment were inconclusive (data not shown) and more thorough analysis is needed. The question remains whether COX-2 induced PGE2 can directly regulate cyclin D1 or other network of cyclins, cyclin-dependent kinases (CDKs) or CDK inhibitors. For other cell types, including colon, lung and squamous cell carcinomas, it has been reported that treatment with NSAIDs results in upregulation of CDK inhibitors that regulate accumulation of cells in G0/G1 [53-55]. In breast cancer cells, this remains to be examined.
Angiogenesis plays a crucial role in tumor development and progression. COX-2 dependent PGE2 production represents a likely candidate for the angiogenic response observed in several tumors, including mammary tumors [36,56-58]. To explore the role played by COX-2 inhibitors in angiogenesis, we used both in vitro and in vivo model systems. Aggressive breast epithelial cells are known to differentiate into tubules when cultured on growth factor reduced Matrigel. This phenomenon is known as vasculogenic mimicry. Its presence has been reported in inflammatory breast cancer patients and is associated with reduced 5-year survival and higher percentage of recurrence [59]. Shirakawa and coworkers [40] suggested a connection between vascular mimicry and angiogenesis, based on the existence of blood flow in the vascular channels. When plated on growth factor reduced Matrigel, human breast cancer cell lines have the unique ability to form tubular channels. We showed that the more aggressive MDA-MB-231 cells generate channels more efficiently and in higher numbers than do the less aggressive MDA-MB-468 cell line (Fig. 6a). Similarly, it was shown that highly aggressive melanoma cells, when seeded on three-dimensional matrices of collagen I, form extracellular matrix-rich patterned networks that surround clusters of tumor cells; however, under the same culture conditions, poorly aggressive melanoma cells did not form the patterned networks [38]. When treated with increasing concentrations of celecoxib (40–60 μmol/l) we observed a dose-dependent decrease in the ability of both cell lines to differentiate into channels (Fig. 6a). Our findings are in accordance with those of other reports, in which capillary-like tube formation by human umbilical vein endothelial cells cocultured with COX-2 overexpressing Caco-2 cells was inhibited by a COX-2 selective inhibitor, NS-398, in a dose-dependent manner [60].
COX-2 inhibitors have already been reported to inhibit angiogenesis, and our study shows for the first time that COX-2 regulates vascular channel formation in human breast cancer cells. The mechanism of action of celecoxib in inhibiting channel formation is not known. Our data suggest that treatment with celecoxib caused a dose-dependent downregulation of VEGF in the MDA-MB-231 cells but not in the MDA-MB-468 cells (Fig. 6b). Although additional mechanisms are involved in mediating the angiogenic effects of COX-2, our data imply that COX-2 inhibitors influence angiogenesis at least in part by decreasing the release of VEGF. It was recently reported that COX-2 induced PGE2 stimulated the expression of angiogenic regulatory genes, including VEGF, in mammary tumor cells isolated from COX-2 transgenic mice, and that treatment with indomethacin (a nonspecific COX inhibitor) suppressed the expression of these genes in vitro [36]. To confirm the in vitro data, the antiangiogenic effects of celecoxib were evaluated in an in vivo xenograft model using MDA-MB-231 cell containing Matrigel implants. Results showed that celecoxib dramatically reduced the vascularity within the tumor tissue (Fig. 8). In addition, the treatment caused increased necrosis and reduced viable tissue mass within the tumor (Figs 7 and 8). Therefore, the reduced tumor burden in the treated mice can be explained in part by the inhibition of angiogenesis and confirms our in vitro data. Previous studies have reported similar effects of COX-2 inhibitors in an in vivo angiogenesis assay using the highly metastatic murine mammary tumor cell line C3L5 [45]. Additional studies are needed to fully elucidate the complex events involved in COX-2 mediated angiogenesis in human mammary tumors.
To our knowledge, this is the first study to identify some key mechanisms of action of celecoxib in vitro and in vivo in human breast cancer cells. More cell lines must be evaluated to characterize fully the antitumor actions of celecoxib, including identification of its primary targets, the precise molecular mechanism of cell damage, and the basis for its preferential effect on tumor cells. Although COX-2 inhibitor treatment alone is unlikely to eliminate an existing tumor, it is likely that it can confer significant benefit as part of a carefully chosen regimen involving other drugs. The strategy to target multiple pathways simultaneously may be critical to improving the efficacy of therapy in the treatment of breast cancer, especially for metastatic breast cancer. Moore and coworkers [61] reported that celecoxib, in combination with 5-fluorouracil or cyclophosphamide, greatly enhanced the antitumor effects of chemotherapy in a colon cancer model. In another tumor model, COX-2 selective inhibitors showed promise in combination with radiation therapy, enhancing tumor radiation responses [62]. Celecoxib was recently shown to have chemopreventive effects against the development of chemically induced mammary tumors in the rat [12]. Finally, recent evidence that combined treatment with a nonselective NSAID and EGFR tyrosine kinase inhibitor significantly decreased polyp formation in Min APC+/- mice supports the notion that combination therapy may be more effective [63]. These studies, combined with the present study and the reports of aberrant COX-2 expression in human breast cancer [9,64], suggest that selective COX-2 inhibitors have an important role to play in chemoprevention, chemo-intervention, and therapy of human breast cancer.
Conclusion
We showed that the mechanisms driving celecoxib-induced growth inhibition of human breast cancer cells are dependent upon COX-2 expression levels, invasive properties, and dependence on PGE2. At the cellular level, celecoxib induced apoptosis in highly invasive cells, but it caused cell cycle arrest at the G0/G1 phase of the cell cycle without causing apoptosis in the less invasive cells. At the molecular level, pAkt was inactivated with increased activation of proapoptotic protein Bax and caspases 3 and 7. Furthermore, we showed for the first time that celecoxib inhibited microvascular channel formation in a dose-dependent manner, associated with downregulation of VEGF in the highly invasive cells. An in vivo xenograft model confirmed the in vitro data and showed dramatic reduction in tumor mass accompanied by reduced vascularity and increased necrosis within the tumor, suggesting that the reduced tumor burden in the treated mice may in part be due to reduced angiogenesis.
Abbreviations
CDK = cyclin-dependent kinase; COX = cyclo-oxygenase; DMEM = Dulbecco's modified eagle medium; DMSO = dimethyl sulfoxide; FCS = fetal calf serum; NSAID = nonsteroidal anti-inflammatory drug; pAkt = phosphorylated Akt; PBS = phosphate-buffered saline; PG = prostaglandin; PI = propidium iodide; VEGF = vascular endothelial growth factor.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
GDB conducted the mouse studies including daily gavaging, palpating tumors and monitoring tumor growth, as well as end-point assays such as apoptosis, and caspase assays. LBP and TLT performed the western blotting and PGE2 assays. SJG provided expert scientific advice with regard to the MTag transgenic mice and mammary carcinogenesis. PM is the PI of the laboratory in which all experiments were conducted and is the recipient of the grant that funded the project. She was instrumental in writing the manuscript.
Acknowledgements
This study was supported by a grant from the Susan G Komen Breast Cancer Foundation (BCTR0202089) to PM. We thank Dr Ronald J Marler for his help in evaluating the histologic specimens, Leslie Dixon for helping with histology, Jim Tarara for helping with the confocal studies and flow cytometry, Marvin H Ruona for help with the graphic production, Scott Dulla for helping with photography, and Carol Williams for preparation of the manuscript.
Figures and Tables
Figure 1 Celecoxib regulates COX-2 levels and causes growth arrest in human breast cancer cells. (a) Cyclo-oxygenase (COX)-2 is expressed in both MDA-MB-231 and MDA-MB-468 cell lines. Western blot analysis of vehicle and celecoxib (20–60 μmol/l) treated cells. SDS-PAGE electrophoresis was performed using a 10% resolving gel. Protein was loaded at 100 μg per lane. Lipopolysaccharide/phorbol 12-myristate 13-acetate treated whole cell lysate from RAW264.7 cell line was used as positive control. Gels were blotted and probed with COX-2 monoclonal antibody. Both cell lines expressed COX-2. MDA-MB-231 cells expressed higher levels of COX-2 than did MDA-MB-468 cells. With drug treatment, COX-2 protein level did not change in the MDA-MB-231 cells, but there was reduction in the level of COX-2 protein in the MDA-MB-468 cells after treatment. β-Actin blot is included to confirm equal loading. These experiments were repeated three times with similar results. (b) Celecoxib induced dose-dependent inhibition of proliferation of breast cancer cell lines. Cells were incubated for 4 days with vehicle or celecoxib, and [3H]thymidine was added 24 hours before harvest. After washing off excess thymidine, cells were lysed with 5% Triton X-100, and incorporated thymidine was evaluated. Celecoxib treatment caused significant dose-dependent growth inhibition in both human breast cancer cell lines. Mean values of three experiments ± standard deviation is shown. P values represent significant differences between vehicle control and celecoxib treatment.
Figure 2 Celecoxib induces apoptosis in MDA-MB-231 cells. (a) Flow cytometric analysis of vehicle-treated and celecoxib-treated cells stained with annexin V and propidium iodide (PI) was done 48 hours after treatment. The population shown in the figure is total apoptotic cells, which includes early and late apoptosis. Significant induction of apoptosis was observed in the MDA-MB-231 cells at 40 and 60 μmol/l concentrations of celecoxib. Apoptosis was not induced in MDA-MB-468 cells. Mean values of three experiments ± standard deviation is shown. P values represent significant differences between vehicle control and celecoxib treatment. (b–e) Celecoxib induces formation of apoptotic bodies in MDA-MB-231 cells. Shown are confocal images of MDA-MB-231 cells subjected to 48 hours of celecoxib (60 μmol/l) treatment. Cells were stained with CFSE (panel d) and then fixed in 95% ethanol and stained with PI (panel c). Cells were visualized in a confocal microscope (Carl Zeiss Inc.) using excitation wavelengths of 488 nm (for CFSE) and 543 nm (for PI). Loss of integrity of nuclear envelope and formation of peripheral, sharply delineated masses of condensed chromatin or apoptotic bodies are visualized. Panel b represents phase contrast images of the cells and panel e represents colocalization of CFSE and PI. Images were taken 200×.
Figure 3 Celecoxib induced down-regulation of pAkt, increase in Bax, and caspase 3/7 in MDA-MB-231 cells. (a) Total Akt and phosphorylated Akt (pAkt). Western blot analysis of cell lysates prepared from vehicle and celecoxib (20–60 μmol/l) treated cells. SDS-PAGE electrophoresis was performed using 10% resolving gel. Protein was loaded at 100 μg per lane and the protein of interest was detected using specific antibodies. Celecoxib treatment at 40 and 60 μmol/l caused decreases in the levels of pAkt in MDA-MB-231 cells, with no change in MDA-MB-468 cells. Numbers below each lane represents percentage of protein expression compared with vehicle-treated cell lysate, which was set to equivalent to 100%, as determined by densitometric analysis. Control cell extracts from Jurkats were used as positive control for Akt and pAkt. (b) Average densitometric values of three separate experiments showing reduction in pAkt with celecoxib treatment. There was a significant decrease (P = 0.002) in the levels of pAkt with 60 μmol/l celecoxib treatment. (c) Western blot analysis of BAX. Increased expression of BAX protein was observed with increasing concentrations of celecoxib in MDA-MB-231 cells but not in MDA-MB-468 cells. The experiment was repeated three times with similar results. A β-actin blot is included to show equal loading. (d) Spectrofluorometric analysis of lysates prepared from vehicle and celecoxib (20–60 μmol/l) treated cells at 48 hours. Activity of caspases 3 and 7 was monitored by enzymatic cleavage using a fluorescence microplate reader with excitation at 485 ± 10 nm and emission detection at 530 ± 12.5 nm. In MDA-MB-231 cells, activities of caspases 3 and 7 were increased significantly at 40 μmol/l and 60 μmol/l drug concentrations. No increase in caspase activity was evident in the MDA-MB-468 cells. Mean values from three experiments ± standard deviation is shown. P values represent significant differences between vehicle control and celecoxib treatment.
Figure 4 Celecoxib causes cell cycle arrest in MDA-MB-468 cells. (a,b) Flow cytometric analysis of cells subjected to treatment with vehicle or celecoxib (20–60 μmol/l) for 48 hours. Cells were fixed and permeabilized with 95% ethanol, stained with propidium iodide, and analyzed by flow cytometry. Celecoxib induced growth arrest at the G0/G1 cell cycle checkpoint in MDA-MB-468 cells (panel b) with no cell cycle arrest in the MDA-MB-231 cells (panel a). Mean values for three experiments ± standard deviation of the mean is shown. P values represent significant difference between vehicle control and celecoxib treatment. Experiments were repeated three times, with similar results.
Figure 5 Growth inhibition of MDA-MB-468 cells was abrogated by exogenous prostaglandin (PG)E2 addition. [3H]thymidine uptake assay was done to determine proliferation of (a) MDA-MB-231 and (b) MDA-MB-468 cells treated with 40 μmol/l celecoxib with or without varying amounts of exogenous PGE2 (12.5–200 pg/ml). Cells were harvested after 96 hours in culture. In MDA-MB-231 cells, growth inhibition induced by 40 μmol/l celecoxib could not be restored by addition of exogenous PGE2; however, addition of 200 pg/ml PGE2 completely reversed the growth inhibition induced by 40 μmol/l celecoxib in the less invasive MDA-MB-468 cells. Average values of three experiments ± standard deviation is shown. P values represent significant differences between vehicle control and celecoxib treatment.
Figure 6 Celecoxib treatment causes reduction in microvascular channel formation by regulating VEGF levels. (a) The percentage of cells forming channels was much greater in MDA-MB-231 cells than in MDA-MB-468 cells. In both cells, treatment with 40 and 60 μmol/l celecoxib caused significant reduction in the number of channels. P values represent significant differences between vehicle control and celecoxib treatment. (b) Western blot analysis of cell lysates prepared from vehicle and celecoxib (20–60 μmol/l) treated cells. SDS-PAGE electrophoresis was performed using 15% resolving gel. Protein was loaded at 100 μg per lane and the protein was detected using vascular endothelial growth factor (VEGF) antibody. Celecoxib treatment decreased VEGF levels in MDA-MB-231 cells in a dose-dependent manner.
Figure 7 Celecoxib treatment reduced MDA-MB-231 tumor growth in nude mice. (a) Five mice per group were treated with either celecoxib (25 mg/kg body weight) or vehicle (dimethyl sulfoxide) and the mice were killed 45 days after the tumor cells were inoculated. Tumor growth was monitored by weekly examination using digital calipers, and tumor weight was calculated using to the following equation [23]: tumor weight (g) = (length (cm) × width (cm)2) × 0.5. Three mice from the vehicle-treated group had to be killed early because of the aggressive nature of the tumor. The other two mice in the vehicle-treated group had significantly greater tumor burden (P = 0.01) than did the five mice in celecoxib-treated group. (b) A representative mouse from each treatment group is illustrated; lower tumor mass Is evident in the treated animal as compared with the vehicle control.
Figure 8 In vivo inhibition of angiogenesis and increase in necrosis with celecoxib treatment. Vascularity of tumor implants was histologically evaluated by Masson's trichrome and factor VIII related antigen staining. Shown is evidence of central necrosis and decreased number of blood vessels in (b) a section of celecoxib-treated tumors relative to (a) a section obtained from a vehicle-treated animal (magnification 50×). Greater magnification (100×) of (c) panel a and (d) panel b are shown in the next two panels. Arrows in panel c point to blood vessels. Endothelial cells lining the blood vessels stained positively for factor VIII related antigen and showed larger blood vessels in the (e) vehicle-treated than in the (f) celecoxib-treated samples (magnification 100×).
Table 1 Celecoxib inhibited prostaglandin E2 secretion by breast cancer cells
Treatment MDA-MB-231 P value MDA-MB-468 P value
Vehicle 430.0 ± 178.2 76.6 ± 15.2
Celecoxib 20 μmol/l 30.1 ± 2.9 <0.01 39.0 ± 16.5 0.03
Celecoxib 40 μmol/l 30.7 ± 4.7 <0.01 37.3 ± 11.0 0.02
Celecoxib 60 μmol/l 37.7 ± 13.6 <0.01 30.3 ± 13.6 0.01
Forty-eight hours post treatment, conditioned medium from vehicle and celecoxib (20–60 μmol/l) treated cells were harvested and prostaglandin (PG)E2 levels (pg/ml) determined by enzyme-linked immunosorbent assay. In all cell lines, PGE2 levels were significantly reduced at all doses of celecoxib. The experiment was repeated three times and in triplicate. P values represent significant differences between vehicle control and celecoxib treatment.
==== Refs
Harris RE Chlebowski RT Jackson RD Frid DJ Ascenseo JL Anderson G Loar A Rodabough RJ White E McTiernan A Breast cancer and nonsteroidal anti-inflammatory drugs: prospective results from the Women's Health Initiative Cancer Res 2003 63 6096 6101 14522941
Lee PP Ip MM Regulation of proliferation of rat mammary tumor cells by inhibitors of cyclooxygenase and lipoxygenase Prostaglandins Leukot Essent Fatty Acids 1992 45 21 31 1546063 10.1016/0952-3278(92)90098-4
Robertson FM Parrett ML Joarder FS Ross M Abou-Issa HM Alshafie G Harris RE Ibuprofen-induced inhibition of cyclooxygenase isoform gene expression and regression of rat mammary carcinomas Cancer Lett 1998 122 165 175 9464506 10.1016/S0304-3835(97)00387-X
Alshafie GA Harris RE Robertson FM Parrett ML Ross M Abou-Issa H Comparative chemopreventive activity of ibuprofen and N-(4-hydroxyphenyl) retinamide against the development and growth of rat mammary adenocarcinomas Anticancer Res 1999 19 3031 3036 10652588
Zhao Y Agarwal VR Mendelson CR Simpson ER Estrogen biosynthesis proximal to a breast tumor is stimulated by PGE2 via cyclic AMP, leading to activation of promoter II of the CYP19 (aromatase) gene Endocrinology 1996 137 5739 5742 8940410 10.1210/en.137.12.5739
Brueggemeier RW Quinn AL Parrett ML Joarder FS Harris RE Robertson FM Correlation of aromatase and cyclooxygenase gene expression in human breast cancer specimens Cancer Lett 1999 140 27 35 10403538 10.1016/S0304-3835(99)00050-6
Richards JA Brueggemeier RW Prostaglandin E2 regulates aromatase activity and expression in human adipose stromal cells via two distinct receptor subtypes J Clin Endocrinol Metab 2003 88 2810 2816 12788892 10.1210/jc.2002-021475
Bandyopadhyay GK Imagawa W Wallace D Nandi S Linoleate metabolites enhance the in vitro proliferative response of mouse mammary epithelial cells to epidermal growth factor J Biol Chem 1987 262 2750 2756 3102474
Pockaj B Basu GD Pathangey LB Gray RJ Hernandez JL Gendler SJ Mukherjee P Reduced T cell and dendritic cell function is related to COX-2 over-expression and prostaglandin E2 secretion in patients with breast cancer Ann Surg Oncol 2004 11 328 339 14993030 10.1245/ASO.2004.05.027
Ristimaki A Sivula A Lundin J Lundin M Salminen T Haglund C Joensuu H Isola J Prognostic significance of elevated cyclooxygenase-2 expression in breast cancer Cancer Res 2002 62 632 635 11830510
Harris RE Namboodiri KK Farrar WB Nonsteroidal antiinflammatory drugs and breast cancer Epidemiology 1996 7 203 205 8834563
Harris RE Alshafie GA Abou-Issa H Seibert K Chemoprevention of breast cancer in rats by celecoxib, a cyclooxygenase 2 inhibitor Cancer Res 2000 60 2101 2103 10786667
Howe LR Subbaramaiah K Patel J Masferrer JL Deora A Hudis C Thaler HT Muller WJ Du B Brown AM Celecoxib, a selective cyclooxygenase 2 inhibitor, protects against human epidermal growth factor receptor 2 (HER-2)/neu-induced breast cancer Cancer Res 2002 62 5405 5407 12359744
Basu GD Pathangey LB Tinder TL LaGioia M Gendler SJ Mukherjee P Cyclooxygenase-2 inhibitor induces apoptosis in breast cancer cells in an in vivo model of spontaneous metastatic breast cancer Proc Natl Acad Sci USA 2002 2 632 642
Liu CH Chang SH Narko K Trifan OC Wu MT Smith E Haudenschild C Lane TF Hla T Overexpression of cyclooxygenase-2 is sufficient to induce tumorigenesis in transgenic mice J Biol Chem 2001 276 18563 18569 11278747 10.1074/jbc.M010787200
Muller-Decker K Neufang G Berger I Neumann M Marks F Furstenberger G Transgenic cyclooxygenase-2 overexpression sensitizes mouse skin for carcinogenesis Proc Natl Acad Sci USA 2002 99 12483 12488 12221288 10.1073/pnas.192323799
Takahashi Y Kawahara F Noguchi M Miwa K Sato H Seiki M Inoue H Tanabe T Yoshimoto T Activation of matrix metalloproteinase-2 in human breast cancer cells overexpressing cyclooxygenase-1 or -2 FEBS Lett 1999 460 145 148 10571077 10.1016/S0014-5793(99)01328-9
Kundu N Smyth MJ Samsel L Fulton AM Cyclooxygenase inhibitors block cell growth, increase ceramide and inhibit cell cycle Breast Cancer Res Treat 2002 76 57 64 12408376 10.1023/A:1020224503335
Munkarah AR Genhai Z Morris R Baker VV Deppe G Diamond MP Saed GM Inhibition of paclitaxel-induced apoptosis by the specific COX-2 inhibitor, NS398, in epithelial ovarian cancer cells Gynecol Oncol 2003 88 429 433 12648598 10.1016/S0090-8258(03)00084-2
Williams CS Watson AJ Sheng H Helou R Shao J DuBois RN Celecoxib prevents tumor growth in vivo without toxicity to normal gut: lack of correlation between in vitro and in vivo models Cancer Res 2000 60 6045 6051 11085526
Hsu AL Ching TT Wang DS Song X Rangnekar VM Chen CS The cyclooxygenase-2 inhibitor celecoxib induces apoptosis by blocking Akt activation in human prostate cancer cells independently of Bcl-2 J Biol Chem 2000 275 11397 11403 10753955 10.1074/jbc.275.15.11397
Lai GH Zhang Z Sirica AE Celecoxib acts in a cyclooxygenase-2-independent manner and in synergy with emodin to suppress rat cholangiocarcinoma growth in vitro through a mechanism involving enhanced Akt inactivation and increased activation of caspases-9 and -3 Mol Cancer Ther 2003 2 265 271 12657721
Jones MK Wang H Peskar BM Levin E Itani RM Sarfeh IJ Tarnawski AS Inhibition of angiogenesis by nonsteroidal anti-inflammatory drugs: insight into mechanisms and implications for cancer growth and ulcer healing Nat Med 1999 5 1418 1423 10581086 10.1038/70995
Simpson-Herrerns L Lloyd HH Kinetic parameters and growth curves for experimental tumor systems Cancer Chemother Rep 1970 54 143 174 5527016
Jadeski LC Lala PK Nitric oxide synthase inhibition by N(G)-nitro-L-arginine methyl ester inhibits tumor-induced angiogenesis in mammary tumors Am J Pathol 1999 155 1381 1390 10514420
Tsujii M DuBois RN Alterations in cellular adhesion and apoptosis in epithelial cells overexpressing prostaglandin endoperoxide synthase 2 Cell 1995 83 493 501 8521479 10.1016/0092-8674(95)90127-2
Liu XH Yao S Kirschenbaum A Levine AC NS398, a selective cyclooxygenase-2 inhibitor, induces apoptosis and down-regulates bcl-2 expression in LNCaP cells Cancer Res 1998 58 4245 4249 9766645
Datta SR Dudek H Tao X Masters S Fu H Gotoh Y Greenberg ME Akt phosphorylation of BAD couples survival signals to the cell-intrinsic death machinery Cell 1997 91 231 241 9346240 10.1016/S0092-8674(00)80405-5
del Peso L Gonzalez-Garcia M Page C Herrera R Nunez G Interleukin-3-induced phosphorylation of BAD through the protein kinase Akt Science 1997 278 687 689 9381178 10.1126/science.278.5338.687
Cardone MH Roy N Stennicke HR Salvesen GS Franke TF Stanbridge E Frisch S Reed JC Regulation of cell death protease caspase-9 by phosphorylation Science 1998 282 1318 1321 9812896 10.1126/science.282.5392.1318
Scheid MP Woodgett JR Unravelling the activation mechanisms of protein kinase B/Akt FEBS Lett 2003 546 108 112 12829245 10.1016/S0014-5793(03)00562-3
Susin SA Zamzami N Castedo M Daugas E Wang HG Geley S Fassy F Reed JC Kroemer G The central executioner of apoptosis: multiple connections between protease activation and mitochondria in Fas/APO-1/CD95- and ceramide-induced apoptosis J Exp Med 1997 186 25 37 9206994 10.1084/jem.186.1.25
Keane RW Srinivasan A Foster LM Testa MP Ord T Nonner D Wang HG Reed JC Bredesen DE Kayalar C Activation of CPP32 during apoptosis of neurons and astrocytes J Neurosci Res 1997 48 168 180 9130145 10.1002/(SICI)1097-4547(19970415)48:2<168::AID-JNR9>3.0.CO;2-A
Krajewski S Gascoyne RD Zapata JM Krajewska M Kitada S Chhanabhai M Horsman D Berean K Piro LD Fugier-Vivier I Immunolocalization of the ICE/Ced-3-family protease, CPP32 (Caspase-3), in non-Hodgkin's lymphomas, chronic lymphocytic leukemias, and reactive lymph nodes Blood 1997 89 3817 3825 9160689
Hirpara JL Seyed MA Loh KW Dong H Kini RM Pervaiz S Induction of mitochondrial permeability transition and cytochrome C release in the absence of caspase activation is insufficient for effective apoptosis in human leukemia cells Blood 2000 95 1773 1780 10688837
Chang SH Liu CH Conway R Han DK Nithipatikom K Trifan OC Lane TF Hla T Role of prostaglandin E2-dependent angiogenic switch in cyclooxygenase 2-induced breast cancer progression Proc Natl Acad Sci USA 2004 101 591 596 14688410 10.1073/pnas.2535911100
Shirakawa K Tsuda H Heike Y Kato K Asada R Inomata M Sasaki H Kasumi F Yoshimoto M Iwanaga T Absence of endothelial cells, central necrosis, and fibrosis are associated with aggressive inflammatory breast cancer Cancer Res 2001 61 445 451 11212228
Maniotis AJ Folberg R Hess A Seftor EA Gardner LM Pe'er J Trent JM Meltzer PS Hendrix MJ Vascular channel formation by human melanoma cells in vivo and in vitro: vasculogenic mimicry Am J Pathol 1999 155 739 752 10487832
Folberg R Hendrix MJ Maniotis AJ Vasculogenic mimicry and tumor angiogenesis Am J Pathol 2000 156 361 381 10666364
Shirakawa K Kobayashi H Heike Y Kawamoto S Brechbiel MW Kasumi F Iwanaga T Konishi F Terada M Wakasugi H Hemodynamics in vasculogenic mimicry and angiogenesis of inflammatory breast cancer xenograft Cancer Res 2002 62 560 566 11809710
Timoshenko AV Xu G Chakrabarti S Lala PK Chakraborty C Role of prostaglandin E2 receptors in migration of murine and human breast cancer cells Exp Cell Res 2003 289 265 274 14499627 10.1016/S0014-4827(03)00269-6
Liu XH Rose DP Differential expression and regulation of cyclooxygenase-1 and -2 in two human breast cancer cell lines Cancer Res 1996 56 5125 5127 8912844
Levitt RJ Buckley J Blouin MJ Schaub B Triche TJ Pollak M Growth inhibition of breast epithelial cells by celecoxib is associated with upregulation of insulin-like growth factor binding protein-3 expression Biochem Biophys Res Commun 2004 316 421 428 15020235 10.1016/j.bbrc.2004.02.062
Minter HA Eveson JW Huntley S Elder DJ Hague A The cyclooxygenase 2-selective inhibitor NS398 inhibits proliferation of oral carcinoma cell lines by mechanisms dependent and independent of reduced prostaglandin E2 synthesis Clin Cancer Res 2003 9 1885 1897 12738747
Rozic JG Chakraborty C Lala PK Cyclooxygenase inhibitors retard murine mammary tumor progression by reducing tumor cell migration, invasiveness and angiogenesis Int J Cancer 2001 93 497 506 11477553 10.1002/ijc.1376
Elder DJ Halton DE Crew TE Paraskeva C Apoptosis induction and cyclooxygenase-2 regulation in human colorectal adenoma and carcinoma cell lines by the cyclooxygenase-2-selective non-steroidal anti-inflammatory drug NS-398 Int J Cancer 2000 86 553 560 10797271 10.1002/(SICI)1097-0215(20000515)86:4<553::AID-IJC18>3.0.CO;2-9
Chan TA Morin PJ Vogelstein B Kinzler KW Mechanisms underlying nonsteroidal antiinflammatory drug-mediated apoptosis Proc Natl Acad Sci USA 1998 95 681 686 9435252 10.1073/pnas.95.2.681
Soengas MS Lowe SW Apoptosis and melanoma chemoresistance Oncogene 2003 22 3138 3151 12789290 10.1038/sj.onc.1206454
Maehama T Taylor GS Dixon JE PTEN and myotubularin: novel phosphoinositide phosphatases Annu Rev Biochem 2001 70 247 279 11395408 10.1146/annurev.biochem.70.1.247
Arico S Pattingre S Bauvy C Gane P Barbat A Codogno P Ogier-Denis E Celecoxib induces apoptosis by inhibiting 3-phosphoinositide-dependent protein kinase-1 activity in the human colon cancer HT-29 cell line J Biol Chem 2002 277 27613 27621 12000750 10.1074/jbc.M201119200
Yip-Schneider MT Sweeney CJ Jung SH Crowell PL Marshall MS Cell cycle effects of nonsteroidal anti-inflammatory drugs and enhanced growth inhibition in combination with gemcitabine in pancreatic carcinoma cells J Pharmacol Exp Ther 2001 298 976 985 11504793
Denkert C Furstenberg A Daniel PT Koch I Kobel M Weichert W Siegert A Hauptmann S Induction of G0/G1 cell cycle arrest in ovarian carcinoma cells by the anti-inflammatory drug NS-398, but not by COX-2-specific RNA interference Oncogene 2003 22 8653 8661 14647459 10.1038/sj.onc.1206920
Goldberg Y Nassif II Pittas A Tsai LL Dynlacht BD Rigas B Shiff SJ The anti-proliferative effect of sulindac and sulindac sulfide on HT-29 colon cancer cells: alterations in tumor suppressor and cell cycle-regulatory proteins Oncogene 1996 12 893 901 8632912
Toyoshima T Kamijo R Takizawa K Sumitani K Ito D Nagumo M Inhibitor of cyclooxygenase-2 induces cell-cycle arrest in the epithelial cancer cell line via up-regulation of cyclin dependent kinase inhibitor p21 Br J Cancer 2002 86 1150 1156 11953864 10.1038/sj.bjc.6600183
Hung WC Chang HC Pan MR Lee TH Chuang LY Induction of p27(KIP1) as a mechanism underlying NS398-induced growth inhibition in human lung cancer cells Mol Pharmacol 2000 58 1398 1403 11093779
Ben-Av P Crofford LJ Wilder RL Hla T Induction of vascular endothelial growth factor expression in synovial fibroblasts by prostaglandin E and interleukin-1: a potential mechanism for inflammatory angiogenesis FEBS Lett 1995 372 83 87 7556649 10.1016/0014-5793(95)00956-A
Seno H Oshima M Ishikawa TO Oshima H Takaku K Chiba T Narumiya S Taketo MM Cyclooxygenase 2- and prostaglandin E(2) receptor EP(2)-dependent angiogenesis in Apc(Delta716) mouse intestinal polyps Cancer Res 2002 62 506 511 11809702
Chu J Lloyd FL Trifan OC Knapp B Rizzo MT Potential involvement of the cyclooxygenase-2 pathway in the regulation of tumor-associated angiogenesis and growth in pancreatic cancer Mol Cancer Ther 2003 2 1 7 12533667 10.1186/1476-4598-2-1
Shirakawa K Wakasugi H Heike Y Watanabe I Yamada S Saito K Konishi F Vasculogenic mimicry and pseudo-comedo formation in breast cancer Int J Cancer 2002 99 821 828 12115483 10.1002/ijc.10423
Tsujii M Kawano S Tsuji S Sawaoka H Hori M DuBois RN Cyclooxygenase regulates angiogenesis induced by colon cancer cells Cell 1998 93 705 716 9630216 10.1016/S0092-8674(00)81433-6
Moore RJ Zweifel BS Heuvelman DM Leahy KM Edwards DA Woerner BM Ornberg RL Seibert K Koki AT Masferrer JL Enhanced antitumor activity co-administration of celecoxib and the chemotherapeutic agents cyclophosphamide and 5-fu Proceedings of the American Association for Cancer Research: 1–5 April 2000 Cadmus Journal Services, Linthicum, MD USA Abstract 2600 409
Kishi K Petersen S Petersen C Hunter N Mason K Masferrer JL Tofilon PJ Milas L Preferential enhancement of tumor radioresponse by a cyclooxygenase-2 inhibitor Cancer Res 2000 60 1326 1331 10728694
Torrance CJ Jackson PE Montgomery E Kinzler KW Vogelstein B Wissner A Nunes M Frost P Discafani CM Combinatorial chemoprevention of intestinal neoplasia Nat Med 2000 6 1024 1028 10973323 10.1038/79534
Soslow RA Dannenberg AJ Rush D Woerner BM Khan KN Masferrer J Koki AT COX-2 is expressed in human pulmonary, colonic, and mammary tumors Cancer 2000 89 2637 2645 11135226 10.1002/1097-0142(20001215)89:12<2637::AID-CNCR17>3.0.CO;2-B
| 15987447 | PMC1175053 | CC BY | 2021-01-04 16:04:33 | no | Breast Cancer Res. 2005 Apr 4; 7(4):R422-R435 | utf-8 | Breast Cancer Res | 2,005 | 10.1186/bcr1019 | oa_comm |
==== Front
Breast Cancer ResBreast Cancer Research1465-54111465-542XBioMed Central London bcr10201598744810.1186/bcr1020Research ArticleClinical utility of serum HER2/neu in monitoring and prediction of progression-free survival in metastatic breast cancer patients treated with trastuzumab-based therapies Esteva Francisco J [email protected] Carol D [email protected] Herbert [email protected] Monica [email protected] Dennis [email protected] Robert P [email protected] Diana [email protected] Farooq [email protected] The University of Texas, MD Anderson Cancer Center, Houston, TX, USA2 Bayer HealthCare, LLC, Diagnostics Division, Tarrytown, NY, USA3 Memorial Sloan Kettering Cancer Center, New York, NY, USA4 University of California, Los Angeles, Department of Medicine, Los Angeles, CA, USA5 Thiel Statistical Consultants, Oxford, CT, USA6 Charité Hospital, Universitätsmedizin Berlin, Berlin, Germany2005 8 4 2005 7 4 R436 R443 28 10 2004 18 1 2005 21 2 2005 4 3 2005 Copyright © 2005 Esteva et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Introduction
The purpose of this retrospective study was to determine the clinical utility of serum HER2/neu in monitoring metastatic breast cancer patients undergoing trastuzumab-based therapy and to compare these results with those obtained using cancer antigen (CA) 15-3. We also sought to determine whether early changes in serum HER2/neu concentrations could be a predictor of progression-free survival.
Methods
Sera were obtained retrospectively from 103 women at four medical institutions. Patients eligible for participation were women with metastatic breast cancer who had HER2/neu tissue overexpression and were scheduled to be treated with trastuzumab with or without additional therapies as per the established practices of the treating physicians. A baseline serum sample for each patient was taken before trastuzumab-based therapy was started. Patients were subsequently monitored over 12 to 20 months and serum samples were taken at the time of clinical assessment and tested with Bayer's HER2/neu and CA15-3 assays.
Results
Concordance between clinical status in patients undergoing trastuzumab-based treatment and HER2/neu and CA15-3 used as single tests was 0.793 and 0.627, respectively, and increased to 0.829 when the tests were used in combination. Progression-free survival times did not differ significantly in patients with elevated baseline HER2/neu concentrations (≥ 15 ng/mL) and those with normal concentrations (<15 ng/mL). However, progression-free survival differed significantly (P = 0.043) according to whether the patient's HER2/neu concentration at 2 to 4 weeks after the start of therapy was >77% or ≤ 77% of her baseline concentration. The median progression-free survival times for these two groups were 217 and 587 days, respectively. A similar trend was observed for a subcohort of patients treated specifically with a combination of trastuzumab and taxane.
Conclusion
These findings indicate that serum HER2/neu testing is clinically valuable in monitoring metastatic breast cancer patients undergoing trastuzumab-based treatment and provides additional value over the commonly used CA15-3 test. The percentage of baseline HER2/neu concentrations in the early weeks after the start of therapy may be an early predictor of progression-free-survival.
==== Body
Introduction
The human-epidermal-growth-factor receptor 2 (HER2, also known as neu, ErbB-2, and p185HER2) is a transmembrane glycoprotein with an intracellular tyrosine kinase activity and an extracellular domain very similar to those of the epidermal-growth-factor-binding domain of the epidermal-growth-factor receptor [1]. The HER2/neu proto-oncogene is amplified and/or overexpressed in approximately 20 to 25% of invasive breast cancers [2,3]. HER2/neu overexpression has been associated with a poor rate of disease-free survival [4]. The role of HER2/neu as a predictive marker of response to hormone therapy and chemotherapy is controversial [5-8]. The extracellular domain (ECD) of the HER2/neu protein is frequently cleaved and released into the circulation, where it can be detected by ELISA in up to 45% of patients with metastatic breast cancer [9]. Rising serum HER2/neu concentrations have been associated with progressive metastatic disease and poor response to chemotherapy and hormonal therapy [10]. Lipton and colleagues [11] showed that patients with an elevated HER2/neu ECD before therapy were less likely to respond to second-line endocrine therapy. The same team evaluated the predictive role of serum HER2/neu ECD in a randomized clinical trial of tamoxifen versus letrozole for patients with metastatic breast cancer. In that trial, patients with low concentrations of circulating HER2/neu ECD had improved response rates and time to progression of disease if they had been treated with letrozole. In patients with an elevated HER2/neu ECD, however, there were no significant differences in outcome between patients treated with tamoxifen and those treated with letrozole [12].
Trastuzumab (Herceptin™, Genentech, South San Francisco, CA, USA) is the only HER2/neu-directed therapy approved by the Food and Drug Administration (FDA) for the treatment of patients with metastatic breast cancer. Trastuzumab is a humanized monoclonal antibody directed against the HER2/neu ECD. Single-agent response rates range from 12 to 30%, depending on the HER2/neu status of the tumor and the patient's prior treatment [13,14]. Response rates and time to disease progression in patients with metastatic breast cancer are better when trastuzumab is combined with chemotherapy than when treatment is with chemotherapy alone [15]. Trastuzumab has been shown to be synergistic with a variety of commonly used chemotherapies such as paclitaxel, docetaxel, platinum salts, and vinorelbine [15-17].
The most commonly used methods of selecting patients for trastuzumab monoclonal antibody therapy are immunohistochemistry and fluorescence in situ hybridization (FISH) [18]. Retrospective studies have shown that HER2/neu gene amplification, measured using FISH, is the best predictive marker of response to trastuzumab-based therapy [14]. The role of circulating HER2/neu ECD as a predictive marker of response to such therapy and its role for monitoring therapy in metastatic breast cancer are not well defined. For the current retrospective, multicenter evaluation, we sought to determine the clinical utility of serum HER2/neu in monitoring metastatic breast cancer patients undergoing trastuzumab-based therapies and to compare these results with those obtained using CA (cancer antigen) 15-3. Additionally, we sought to determine whether early changes in serum HER2/neu concentrations can be a predictor of progression-free survival.
Materials and methods
Patient population
Sera were obtained from 103 subjects meeting the study entry criteria at four medical institutions: the University of Texas MD Anderson Cancer Center, Memorial Sloan Kettering Cancer Center, the Charité Hospital, and the University of California. The study was reviewed and approved at each site by the local institutional review board and written informed consent was obtained from each enrolled subject. Patients eligible for participation were women with metastatic breast cancer who had HER2/neu overexpression as determined by immunohistochemistry (scores of 2+ or 3+ on the Dako Herceptest, Dako Corp, Carpinteria, CA, USA, or CB11 antibody Pathway, Ventana Medical Systems Inc, Tucson, AZ, USA) or by FISH (PathVysion HER2/neu method, Vysis Inc, Downers Grove, IL, USA) or the INFORM HER2/neu gene detection system, Ventana Medical Systems) and who were scheduled to be treated with trastuzumab with or without additional therapies as per the established practices of the treating physicians. Patients were trastuzumab-naive at entry to the study. The first (baseline) serum sample for each patient was taken before trastuzumab-based therapy was started. Patients were subsequently monitored over 12 to 20 months during trastuzumab-based therapy, with at least four subsequent serial serum samples being collected at least 4 weeks apart, at the discretion of the physician at the time of clinical assessment.
Evaluation of tumor response
Clinical status was determined by available clinical data in the form of physical examination and imaging techniques (ultrasound, x-ray, computed tomography, magnetic resonance imaging) at each patient visit. All patients had documented, measurable metastatic breast cancer at entry to the study. Response to treatment was assessed according to the criteria of the World Health Organization [19] or RECIST (response evaluation criteria in solid tumors) [20]. Clinical status – whether response was complete or partial and whether disease was stable or progressive – was determined at each patient visit. Complete or partial responses were confirmed at a minimum of 4 weeks after complete or partial response was observed.
Sample collection and serum HER2/neu antigen testing
Serum samples were processed immediately upon collection of blood and kept at -20°C or colder. Serum samples were either tested at the site or were subsequently shipped on dry ice to Bayer Diagnostics (Tarrytown, NY, USA) for batch testing. Frozen samples were tested immediately upon initial thawing.
The HER2/neu antigen testing was performed using either the Bayer Immuno1 or ADVIA Centaur™ automated assays. Both methods are currently cleared by the FDA with an indication for follow-up and monitoring of patients with metastatic breast cancer. Previous studies have shown similar diagnostic performance of the automated methods, with very high correlation between the methods (r2 = 0.99) (product method sheet), because the antibodies used for capture and detection of the circulating HER2/neu antigen are identical for these methods. A serum HER2/neu concentration of 15 ng/ml has been defined as the upper limit of normal [21].
The Immuno1 assay technology has been previously described [21]. Briefly, the ADVIA Centaur HER2/neu assay is a two-site sandwich immunoassay using direct, chemiluminescent technology. The Lite Reagent is composed of the monoclonal mouse antibody TA-1 labelled with acridinium ester. The Fluorescein Conjugate Reagent is composed of the monoclonal mouse antibody NB-3 labelled with fluorescein. These two monoclonal antibodies are specific for unique epitopes on the extracellular domain of HER2/neu. The Solid Phase is composed of purified monoclonal mouse capture antibody (antifluorescein) covalently coupled to paramagnetic particles, which binds the immunocomplex. The reaction is initiated and the measured chemiluminescence is directly proportional to the quantity of HER2/neu antigen in the sample.
Samples were also tested for CA15-3 concentrations using the ADVIA Centaur automated assay in accordance with the manufacturer's instructions. Single determinations of HER2/neu and CA15-3 were obtained on each specimen tested.
Statistical analysis
Monitoring of metastatic breast cancer patients with serum HER2/neu
To determine the association between HER2/neu and disease status throughout the monitoring period, patients with stable partial or complete response were grouped together and considered as having no disease progression. This group was compared with the patients whose disease was progressing. For each pair of serial measurements, changes in HER2/neu of ≥ 15% or < 15% and in CA15-3 of ≥ 21% or <21% were considered to indicate progression or lack of progression, respectively. These values (i.e., the 15% or 21% change) were based on the variability of each marker when tested longitudinally over time in a population of apparently healthy woman (product method sheets). Concordance with clinical status was determined for both markers with 95% confidence intervals (CI) as single tests and, when the two markers were used in combination, as a series test. For a series test, change was considered positive if it was ≥ 15% for HER2/neu and ≥ 21% for CA15-3. All other changes were considered no change. Concordance estimates and their CIs were obtained using Windows Excel with the Resampling Stat add-in.
Association between baseline HER2/neu concentrations and progression-free survival
To determine whether the baseline concentration of serum HER2/neu is a predictor of progression-free survival in patients with metastatic breast cancer treated with trastuzumab-based therapies, time to progression was measured from baseline to the date of first documented disease progression for the serum-positive patient group (pretreatment HER2/neu ≥ 15 ng/ml) as compared with the serum-negative patient group (pretreatment HER2/neu <15 ng/ml), using the Kaplan–Meier method. A log-rank test was used to determine if significant difference in progression-free-survival existed between the two groups.
Association between percentage of baseline HER2/neu concentrations after 2 to 4 weeks of therapy and progression-free survival
Where possible based on serum sample availability, the percentage of baseline HER2/neu concentrations at 2 to 4 weeks of therapy was determined for a patient. ROC (receiver operating characteristic) curve analysis as a function of the percentage of change in the HER2/neu concentrations between baseline and 2 to 4 weeks after initiation of therapy was used to determine the optimal cutoff point. Progression-free survival was compared between the groups of patients above and below the cutoff point using the Kaplan–Meier method. A log-rank test was used to determine significant difference in progression-free-survival between the two groups.
Kaplan–Meier analysis and graphs were generated using SPSS (Statistical Package for the Social Sciences) for Windows (version 11.0).
Results
This retrospective study included a total of 103 women with HER2/neu-overexpressing metastatic breast cancer treated with trastuzumab-based therapies. The mean serum HER2/neu concentration at baseline was 170.3 ng/mL, with a median of 24.0 ng/mL. The mean HER2/neu value for patients with an immunohistochemistry score of ≤ 2+ was 182.8 ng/mL. For patients with an immunohistochemistry score of 3+, the mean Her-2/neu value was 201.8 ng/mL; a two-sample t-test indicated that the difference was not significant (t = 0.118, df = 74, P = 0.9). The mean serum CA15-3 concentration at baseline was 224.4 ng/mL, with a median of 42.6 ng/mL. The average age of the patients at entry to the study was 54 years (range 26 to 97 years). Table 1 shows the characteristics of the patients with respect to grade of HER2/neu overexpression and treatments.
To assess the clinical value of serum HER2/neu in monitoring patients longitudinally over the time course of trastuzumab-based treatment, a total of 99 of the evaluable patients with both HER2/neu and CA15-3 measurements were evaluated. These patients generated a total of 362 visit pairs for analysis. The median of time points was 4. Table 2 shows the estimates of concordance with clinical status for each method alone and in combination as a series test. HER2/neu showed considerably greater concordance than CA15-3 in patients monitored longitudinally, and when the two tests were used in combination, the concordance with clinical status was slightly better than when HER2/neu was used alone.
Baseline concentrations of serum HER2/neu were elevated (that is, ≥ 15 ng/mL) in 70% (72) of the 99 evaluable patients. Of the 27 patients with a normal HER2/neu baseline value, disease progressed in 21, and Her-2/neu values were available for 20 of these at the time when progression was found: 7 (35%) had elevated HER2/neu values at that time. Of the 27 patients with normal baseline values of Her-2/neu, 10 who were receiving the trastuzumab–taxane combination had progression of their disease and 3 of these 10 had elevated HER2/neu values at the time of the progression.
Kaplan–Meier analysis in patients with elevated baseline concentrations of HER2/neu (≥ 15 ng/mL) versus those with normal baseline concentrations (<15 ng/mL) showed no significant differences in progression-free survival between these two groups (log-rank statistic = 2.38, df = 1, P > 0.05). The median follow-up was 237 days. The median progression-free survival times were 324 days for patients with normal baseline values of HER2/neu and 462 days for those with elevated baseline values.
We determined the association between percentage of baseline HER2/neu concentrations at 2 to 4 weeks after the start of therapy and progression-free survival in a subcohort of 26 evaluable patients. Nineteen of these patients were treated with trastuzumab (Herceptin) in combination with a taxane; one received Herceptin in combination with vinorelbine; and six received single-agent Herceptin, without chemotherapy. Figure 1 represents the ROC (receiver operating characteristic) curve for disease progression as a function of the percentage of change in HER2/neu from baseline to 2 to 4 weeks after initiation of therapy. The area under the curve is 0.72 (95% CI 0.52 to 0.92). This value of the area under the curve indicates that the variable (percentage of change in HER2/neu from baseline) discriminates between those patients whose disease progressed and those whose disease did not progress.
Examination of the coordinates of the curve indicates that an optimal cutoff point for the variable is 77%. At this value, the true-positive fraction is 72.7%, with a true-negative fraction of 53.3%. In addition, the odds ratio for progression is 1.56, which is a relative maximum at the cutoff.
Figure 2 shows the Kaplan–Meier curves for progression-free survival in patients with >77% of baseline concentrations of HER2/neu and ≤ 77% of baseline HER2/neu concentrations. The difference between the median progression-free survival times for these two groups, 217 and 587 days, respectively, was significant (log-rank statistic 4.04, df = 1, P = 0.043).
A subcohort of 52 patients treated specifically with a combination of trastuzumab and taxane was stratified according to the patients' baseline HER2/neu concentrations; 35 had elevated baseline concentrations (≥ 15 ng/mL) and 17 had baseline concentrations <15 ng/mL. Kaplan–Meier analysis for progression-free survival in these two groups showed no significant differences (log-rank statistic = 0.00, df = 1, P = 0.12).
Additionally, we evaluated, in a subpopulation of 15 patients, the association between percentage of baseline HER2/neu concentrations found 2 to 4 weeks after the start of trastuzumab–taxane therapy and progression-free survival. The median progression-free survival in the group of patients whose HER2/neu value fell to ≤ 77% of baseline (median decrease for the cohort) at 2 to 4 weeks of therapy was 587 days, whereas in the group whose HER2/neu values were >77% of baseline values, the median progression-free survival time was 119 days (Fig. 3). Although there was no significant difference observed in progression-free survival for these two groups of patients (log-rank statistic = 3.40, df = 1, P = 0.065), there was a trend that was similar to that observed for the larger cohort of patients who were undergoing any of the trastuzumab–based therapies included in our study.
Discussion
In this cohort, the HER2/neu ECD concentration after 2 to 4 weeks of treatment relative to that at baseline predicted progression-free survival in patients undergoing trastuzumab-based therapy for metastatic breast cancer. Since most patients would not usually undergo imaging studies for assessment of their response to therapy after 2 to 3 months, the serum HER2/neu ECD data may be clinically useful to determine if an individual patient is responding, or whether earlier imaging studies are warranted. Furthermore, in our patients, the serum HER2/neu ECD provided additional value over tumor marker CA15-3.
Currently, patients with metastatic breast cancer are selected for trastuzumab-based therapy if the HER2/neu protein is overexpressed (immunohistochemistry score 3+) in the primary tumor, or if FISH provides evidence of HER2/neu gene amplification [14,15,22]. Although the HER2/neu ECD assay is approved by the FDA for monitoring patients undergoing systemic therapy for metastatic breast cancer, the clinical utility of this marker in patients undergoing trastuzumab-based therapy is not well defined. In 2000, the American Society of Clinical Oncology Tumor Markers Expert Panel did not recommend the use of CEA, CA15-3, CA27-29, or HER2/neu ECD to monitor patients with metastatic breast cancer [23]. Although a rising tumor marker may be detected sooner than radiologically visible progression or than symptoms, available data do not indicate that this approach would lead to improved survival in patients with metastatic breast cancer. However, as novel targeted therapies are developed, specific assays such as the HER2/neu ECD may gain clinical utility [24].
At the time of the report of the expert panel of the American Society of Clinical Oncology, there were no published data regarding circulating HER2/neu ECD and response to trastuzumab-based therapy. Since then, several preliminary studies have found that serum HER2/neu followed the course of disease while patients were receiving trastuzumab monoclonal antibody therapy. Esteva and colleagues [25] evaluated serum HER2/neu in 30 women with tissue HER2/neu overexpressing metastatic breast cancer who were undergoing weekly docetaxel and trastuzumab therapy. They found that serum HER2/neu concentrations decreased in 87% of the responding patients, indicating that changes in serum HER2/neu concentrations correlate well with the clinical course of disease. Furthermore, they showed that patients with elevated pretreatment serum concentrations (≥ 15 ng/mL) of HER2/neu had a high (76%) overall response rate to therapy, compared with a 33% response rate in patients with baseline concentrations <15 ng/mL. Dnistrian and collaborators [26] evaluated serum HER2/neu in 54 patients who had metastatic breast cancer with tissue overexpression of HER2/neu and who were undergoing trastuzumab therapy either alone or in combination with paclitaxel. Similar to the results of Esteva and colleagues, they found that among patients with an abnormal pretreatment serum HER2/neu, 83% responded favorably to trastuzumab therapy. Futhermore, Kostler and coworkers [27] evaluated the clinical utility of serum HER2/neu in the early prediction of response to trastuzumab-based therapy in 55 patients with tissue HER2/neu-overexpressing metastatic breast cancer. They found that the ratio of the relative values of HER2/neu from each serial sample during trastuzumab treatment to the baseline HER2/neu concentration before initiation of trastuzumab therapy were significantly decreased in patients responsive to treatment in the early weeks after initiation of therapy. These authors showed that serial HER2/neu concentrations could predict the risk for disease progression as early as day 15 of treatment.
Burstein and collaborators [28] measured HER2/neu ECD concentrations at baseline and during treatment with vinorelbine and trastuzumab. In that study, HER2/neu ECD concentrations were not predictive of response to therapy, but a lack of decline in that concentration was a predictor for tumor progression after cycle 1. In our retrospective study, baseline HER2/neu ECD concentrations were not predictive of response to therapy or time to progression. However, patients who had ≤ 77% of baseline in serum HER2/neu ECD concentrations 2 to 4 weeks after initiation of therapy had a longer time to progression (median 587 days) than patients with >77% of baseline concentrations of HER2/neu ECD (median 217 days). A similar trend in the prediction of progression-free survival was observed in the subcohort of patients treated with a combination of trastuzumab and taxane.
One of the limitations of this study is that we evaluated a selected population, since all patients have some degree of HER2 expression in their primary tumors. However, approximately two-thirds of HER2/neu tissue-positive patients do not respond to single-agent trastuzumab [15], and approximately a third of HER2/neu tissue-positive patients do not respond to trastuzumab in combination with chemotherapy [15,25]. Serum HER2/neu allows for real-time assessment of HER2 status, which could be quantified, and because it is a noninvasive procedure repeated testing is possible to determine the change in HER2/neu concentrations. Therefore, we believe that serum HER2 ECD concentrations in patients with metastatic disease complements the data regarding tissue expression at the time of initial diagnosis (generally, HER2/neu status is assessed in primary breast cancer specimens, not in metastatic deposits).
Conclusion
This article does not advocate using baseline serum HER2/neu to predict the efficacy of trastuzumab (Herceptin) therapy. However, it does show value in using the delta change in HER2/neu ECD from baseline to 2 to 4 weeks after the start of therapy. Monitoring efficacy during the course of therapy would allow clinicians to make necessary adjustments to drug combination or even prompt them to follow up patients more frequently. Our hypotheses are currently being tested in a prospective, multicenter clinical trial.
Abbreviations
CA = cancer antigen; ELISA = enzyme-linked immunosorbent assay; ECD = extracellular domain; FISH = fluorescence in situ hybridization; FDA = Food and Drug Administration; HER = human-epidermal-growth-factor receptor.
Competing interests
CC and FG are employees at Bayer Diagnostics. FJE, MF, DS, and DL received research funds from Bayer Diagnostics.
Authors' contributions
All authors read and approved the final manuscript. FJE, CDC, and FG conceived of the study, participated in its design and coordination, and helped to draft the manuscript. HB, MF, DS, and DL participated in the study design and coordination and helped to draft the manuscript. RPT participated in the design of the study and performed the statistical analysis.
Figures and Tables
Figure 1 ROC curve analysis for progression of metastatic breast cancer in 26 patients. Nineteen of these patients were treated with trastuzumab (Herceptin) in combination with a taxane, one with trastuzumab (Herceptin) in combination with vinorelbine, and six with single-agent Herceptin without chemotherapy. The curve shows disease progression as a function of the percentage of change in the HER2/neu concentration from baseline to 2 to 4 weeks after the start of therapy. HER, human-epidermal-growth-factor receptor; ROC, receiver operating characteristic.
Figure 2 Kaplan–Meier curves for progression-free survival in patients with metastatic breast cancer given trastuzumab-based treatment. Curves were plotted for two groups of patients, those whose HER2/neu concentrations seen at 2 to 4 weeks of trastuzumab-based therapy were ≤ 77% of the baseline concentrations and those for whom it was >77% (median progression-free survival time 587 and 217 days, respectively) df = 1. HER, human-epidermal-growth-factor receptor.
Figure 3 Kaplan–Meier curves for progression-free survival in patients with metastatic breast cancer receiving trastuzumab–taxane. Differences between the survival times (days) for patients whose HER2/neu concentrations at 2 to 2 weeks of treatment had fallen to ≤ 77% of baseline values and those whose values had remained at >77% of baseline values (respective survival times 587 and 119 days).
Table 1 Characteristics of 103 women with HER2/neu-overexpressing metastatic breast cancer
Patient characteristic n %
HER2/neu overexpressiona
Score 2+ 31 30.1
Score 3+ 72 69.9
Treatment
Single-agent trastuzumab 13 12.6
Trastuzumab + taxane 52 50.5
Trastuzumab + vinorelbine 11 10.7
Trastuzumab + taxane + megestrol acetate 6 5.8
Trastuzumab + taxane + tamoxifen 5 4.8
Trastuzumab + bevacizumab 5 4.8
Trastuzumab + taxane + carboplatin 4 3.9
Trastuzumab + taxane + CMF 3 2.9
Trastuzumab + erlotinib 2 1.9
Trastuzumab + doxorubicin + cyclophosphamide 2 1.9
aScores indicating HER2/neu overexpression as determined immunohistochemically using the Dako Herceptest. CMF, cyclophosphamide, methotrexate, fluorouracil; HER, human-epidermal-growth-factor receptor.
Table 2 Estimates of concordance of serum values with clinical status of patients treated with trastuzumab
Concordance HER2/neu CA15-3 Series testa
Cb 0.793 0.627 0.829
aChange was considered positive if it was ≥ 15% for HER2/neu and ≥ 21% for CA15-3. bDifferences in concordance for HER2/neu and CA15-3 with 95%CI = 0.166 (lower limit 0.102, upper limit 0.227).
==== Refs
Yamamoto T Ikawa S Akiyama T Semba K Nomura N Miyajima N Saito T Toyoshima K Similarity of protein encoded by the human c-erbB-2 gene to epidermal growth factor receptor Nature 1986 319 230 234 3003577 10.1038/319230a0
Slamon DJ Clark GM Wong SG Levin WJ Ullrich A McGuire WL Human breast cancer: correlation of relapse and survival with amplification of the HER-2/neu oncogene Science 1987 235 177 182 3798106
Pauletti G Dandekar S Rong H Ramos L Peng H Seshadri R Slamon DJ Assessment of methods for tissue-based detection of the HER-2/neu alteration in human breast cancer: a direct comparison of fluorescence in situ hybridization and immunohistochemistry J Clin Oncol 2000 18 3651 3664 11054438
Slamon DJ Godolphin W Jones LA Holt JA Wong SG Keith DE Levin WJ Stuart SG Udove J Ullrich A Studies of the HER-2/neu proto-oncogene in human breast and ovarian cancer Science 1989 244 707 712 2470152
Elledge RM Green S Ciocca D Pugh R Allred DC Clark GM Hill J Ravdin P O'Sullivan J Martino S HER-2 expression and response to tamoxifen in estrogen receptor-positive breast cancer: a Southwest Oncology Group Study Clin Cancer Res 1998 4 7 12 9516946
Menard S Valagussa P Pilotti S Gianni L Biganzoli E Boracchi P Tomasic G Casalini P Marubini E Colnaghi MI Response to cyclophosphamide, methotrexate, and fluorouracil in lymph node-positive breast cancer according to HER2 overexpression and other tumor biologic variables J Clin Oncol 2001 19 329 335 11208823
Paik S Bryant J Tan-Chiu E Yothers G Park C Wickerham DL Wolmark N HER2 and choice of adjuvant chemotherapy for invasive breast cancer: National Surgical Adjuvant Breast and Bowel Project Protocol B-15 J Natl Cancer Inst 2000 92 1991 1998 11121461 10.1093/jnci/92.24.1991
Gonzalez-Angulo AM Krishnamurthy S Yamamura Y Broglio KR Pusztai L Buzdar AU Hortobagi GN Esteva FJ Lack of association between amplification of her-2 and response to preoperative taxanes in patients with breast carcinoma Cancer 2004 101 258 263 15241821 10.1002/cncr.20348
Rosenberg SA The immunotherapy of solid cancers based on cloning the genes encoding tumor-rejection antigens Annu Rev Med 1996 47 481 491 8712798 10.1146/annurev.med.47.1.481
Baselga J Is circulating HER-2 more than just a tumor marker? Clin Cancer Res 2001 7 2605 2607 11555569
Lipton A Ali SM Leitzel K Demers L Chinchilli V Engle L Harvey HA Brady C Nalin CM Dugan M Elevated serum HER-2/neu level predicts decreased response to hormone therapy in metastatic breast cancer J Clin Oncol 2002 20 1467 1472 11896093 10.1200/JCO.20.6.1467
Lipton A Ali SM Leitzel K Demers L Harvey HA Chaudri-Ross HA Brady C Wyld P Carney W Serum HER-2/neu and response to the aromatase inhibitor letrozole versus tamoxifen J Clin Oncol 2003 21 1967 1972 12743150 10.1200/JCO.2003.09.098
Baselga J Tripathy D Mendelsohn J Baughman S Benz CC Dantis L Sklarin NT Seidman AD Hudis CA Moore J Phase II study of weekly intravenous recombinant humanized anti-p185HER2 monoclonal antibody in patients with HER2/neu-overexpressing metastatic breast cancer J Clin Oncol 1996 14 737 744 8622019
Vogel CL Cobleigh MA Tripathy D Gutheil JC Harris LN Fehrenbacher L Slamon DJ Murphy M Novotny WF Burchmore M Efficacy and safety of trastuzumab as a single agent in first-line treatment of HER2-overexpressing metastatic breast cancer J Clin Oncol 2002 20 719 726 11821453 10.1200/JCO.20.3.719
Slamon DJ Leyland-Jones B Shak S Fuchs H Paton V Bajamonde A Fleming T Eiermann W Wolter J Pegram M Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2 N Engl J Med 2001 344 783 792 11248153 10.1056/NEJM200103153441101
Pegram MD Finn RS Arzoo K Beryt M Pietras RJ Slamon DJ The effect of her-2/neu overexpression on chemotherapeutic drug sensitivity in human breast and ovarian cancer cells Oncogene 1997 15 537 547 9247307 10.1038/sj.onc.1201222
Nahta R Esteva FJ In vitro effects of trastuzumab and vinorelbine in trastuzumab-resistant breast cancer cells Cancer Chemother Pharmacol 2004 53 186 190 14605867 10.1007/s00280-003-0728-3
Nahta R Esteva FJ HER-2-targeted therapy: lessons learned and future directions Clin Cancer Res 2003 9 5078 5084 14613984
World Health Organization Handbook for Reporting Results of Cancer Treatment 1979 Geneva: World Health Organization Offset publication 48.
Therasse P Arbuck SG Eisenhauer EA Wanders J Kaplan RS Rubinstein L Verweij J Van Glabbeke M van Oosterom AT Christian MC New guidelines to evaluate the response to treatment in solid tumors. European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada J Natl Cancer Inst 2000 92 205 216 10655437 10.1093/jnci/92.3.205
Cook GB Neaman IE Goldblatt JL Cambetas DR Hussain M Luftner D Yeung KK Chan DW Schwartz MK Allard WJ Clinical utility of serum HER-2/neu testing on the Bayer Immuno 1 automated system in breast cancer Anticancer Res 2001 21 1465 1470 11396233
Cobleigh MA Vogel CL Tripathy D Robert NJ Scholl S Fehrenbacher L Wolter JM Paton V Shak S Lieberman G Multinational study of the efficacy and safety of humanized anti-HER2 monoclonal antibody in women who have HER-2 overexpressing metastatic breast cancer that has progressed after chemotherapy for metastatic disease J Clin Oncol 1999 17 2639 2648 10561337
Bast RC Ravdin P Hayes DF Bates S Fritsche H Jessup JM Kemeny N Locker GY Mennel RG Somerfield MR 2000 update of recommendations for the use of tumor markers in breast and colorectal cancer: clinical practice guidelines of the American Society of Clinical Oncology J Clin Oncol 2001 19 1865 1878 11251019
Luftner D Cheli C Mickelson K Sampson E Possinger K ADVIA Centaur HER-2/neu shows value in monitoring patients with metastatic breast cancer Int J Biol Markers 2004 19 175 182 15503818
Esteva FJ Valero V Booser D Guerra LT Murray JL Pusztai L Cristofanilli M Arun B Esmaeli B Fritsche HA Phase II study of weekly docetaxel and trastuzumab for patients with HER-2-overexpressing metastatic breast cancer J Clin Oncol 2002 20 1800 1808 11919237 10.1200/JCO.2002.07.058
Dnistrian AM Schwartz MK Schwartz DC Ghani F Kish L Significance of serum her-2/neu oncoprotein, CA 15-3, and CEA in the clinical evaluation of metastatic breast cancer J Clin Ligand Assay 2002 25 215 220
Kostler WJ Schwab B Singer CF Neumann R Rucklinger E Brodowicz T Tomek S Niedermayr M Hejna M Steger GG Monitoring of serum Her-2/neu predicts response and progression-free survival to trastuzumab-based treatment in patients with metastatic breast cancer Clin Cancer Res 2004 10 1618 1624 15014012
Burstein HJ Harris LN Marcom PK Lambert-Falls R Havlin K Overmoyer B Friedlander RJ Gargiulo J Strenger R Vogel CL Trastuzumab and vinorelbine as first-line therapy for HER2-overexpressing metastatic breast cancer: multicenter phase II trial with clinical outcomes, analysis of serum tumor markers as predictive factors, and cardiac surveillance algorithm J Clin Oncol 2003 21 2889 2895 12885806 10.1200/JCO.2003.02.018
| 15987448 | PMC1175054 | CC BY | 2021-01-04 16:54:50 | no | Breast Cancer Res. 2005 Apr 8; 7(4):R436-R443 | utf-8 | Breast Cancer Res | 2,005 | 10.1186/bcr1020 | oa_comm |
==== Front
Breast Cancer ResBreast Cancer Research1465-54111465-542XBioMed Central London bcr10221598744510.1186/bcr1022Research ArticleStromal cell derived factor-1: its influence on invasiveness and migration of breast cancer cells in vitro, and its association with prognosis and survival in human breast cancer Kang Hua 12Watkins Gareth 1Parr Christian 1Douglas-Jones Anthony 3Mansel Robert E 1Jiang Wen G 1Jiang, [email protected] Metastasis and Angiogenesis Research Group, Wales College of Medicine, Cardiff University, Cardiff, UK2 Currently working in the Department of Surgery, Xuanwu Hospital, Beijing, China3 Department of Pathology, Wales College of Medicine, Cardiff University, Cardiff, UK2005 4 4 2005 7 4 R402 R410 6 11 2004 26 1 2005 1 2 2005 8 3 2005 Copyright © 2005 Kang et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Introduction
Stromal cell-derived factor (SDF)-1 (CXC chemokine ligand-12) is a member of the CXC subfamily of chemokines, which, through its cognate receptor (CXC chemokine receptor [CXCR]4), plays an important role in chemotaxis of cancer cells and in tumour metastasis. We conducted the present study to evaluate the effect of SDF-1 on the invasiveness and migration of breast cancer cells, and we analyzed the expression of SDF-1 and its relation to clinicopathological features and clinical outcomes in human breast cancer.
Method
Expression of SDF-1 mRNA in breast cancer, endothelial (HECV) and fibroblast (MRC5) cell lines and in human breast tissues were studied using RT-PCR. MDA-MB-231 cells were transfected with a SDF-1 expression vector, and their invasiveness and migration was tested in vitro. In addition, the expression of SDF-1 was investigated using immunohistochemistry and quantitative RT-PCR in samples of normal human mammary tissue (n = 32) and mammary tumour (n = 120).
Results
SDF-1 expression was identified in MRC5, MDA-MB-435s and MDA-MB-436 cell lines, but CXCR4 expression was detected in all cell lines and breast tissues. An autocrine loop was created following transfection of MDA-MB-231 (which was CXCR4 positive and SDF-1 negative) with a mammalian expression cassette encoding SDF-1 (MDA-MB-231SDF1+/+) or with control plasmid pcDNA4/GFP (MDA-MB-231+/-). MDA-MB-231SDF1+/+ cells exhibited significantly greater invasion and migration potential (in transfected cells versus in wild type and empty MDA-MB-231+/-; P < 0.01). In mammary tissues SDF-1 staining was primarily seen in stromal cells and weakly in mammary epithelial cells. Significantly higher levels of SDF-1 were seen in node-positive than in node-negative tumours (P = 0.05), in tumours that metastasized (P = 0.05), and tumours from patients who died (P = 0.03) than in tumours from patients who were disease free. It was most notable that levels of SDF-1 correlated significantly with overall survival (P = 0.001) and incidence-free survival (P = 0.035).
Conclusion
SDF-1 can increase the invasiveness and migration of breast cancer cells. Its levels correlated with node involvement and long-term survival in patients with breast cancer. SDF-1 may therefore have potential value in assessing clinical outcomes of patients with breast cancer.
==== Body
Introduction
Breast cancer is the most common female cancer in the UK and USA. One in ten women will develop breast cancer in their lifetime in Western countries [1,2]. The poor prognosis of patients with breast cancer is related to tumour recurrence and metastasis [3,4]. Breast cancer is characterized by metastasis to regional lymph nodes, bone marrow, lungs and the liver [5]. Previous studies [6,7] demonstrated that sites of metastasis are determined not only by the characteristics of neoplastic cells but also by the microenvironment of the specific organs. Organ specific attractant molecules can promote homing of tumour cells to particular sites [5,7].
Stromal cell-derived factor (SDF)-1 (CXC chemokine ligand-12) is a member of CXC chemokine family, which was initially cloned from murine bone marrow and characterized as a pre-B-cell growth stimulating factor [8-10]. SDF-1 exerts effects through its cognate receptor CXC chemokine receptor (CXCR4), which is the only physiological receptor for SDF-1 and is known to play roles in chemotaxis [11,12], haematopoiesis [13,14], vasculogenesis [15-17] and tumour spread and metastasis [6,18,19]. It was recently shown that CXCR4 is involved in homing of tumour cells to specific organs and in tumour progression [6,18-20]. Muller and coworkers [19] found that SDF-1/CXCR4 plays a critical role in determining the metastatic destination of breast cancer cells. Moreover, they demonstrated that neutralization with a specific monoclonal antibody against CXCR4 effectively inhibited the metastasis of breast cancer cells to the lung or lymph nodes in mice [19].
However, despite the accumulated information on CXCR4, few studies have been conducted to evaluate SDF-1 expression and its prognostic value in patients with breast cancer. In the present study we evaluated the effect of the SDF-1 gene in breast cancer cells on their invasive and migration properties, using a SDF-1 transfection technique. Furthermore, we analyzed SDF-1 expression by real-time quantitative RT-PCR and immunohistochemical staining, and its relation with clinicopathological features and clinical outcomes in human breast cancer.
Materials and method
Materials
The RNA extraction kit and reverse transcription kit were obtained from AbGene Ltd (Epsom, Surrey, UK). PCR primers were designed using Beacon Designer (Palo Alto, CA, USA) and synthesized by Invitrogen Ltd (Paisley, UK). Molecular biology grade agarose and DNA ladder were obtained from Invitrogen. The master mix for routine PCR and quantitative PCR was from AbGene Ltd. Goat anti-human SDF-1 polyclonal antibodies and rabbit anti-human CXCR4 polyclonal antibody were purchased from Santa Cruz Biotechnology Ltd (Santa Cruz, CA, USA). Peroxidase conjugated anti-goat and anti-rabbit antibodies were obtained from Sigma (Poole, Dorset, England, UK) and a biotin universal staining kit was from Vector Laboratories (Nottingham, England, UK). Matrigel (reconstituted basement membrane) was purchased from Collaborative Research Products (Bedford, MA, USA). A transwell plate equipped with a porous insert (pore size 8 μm] was obtained from Becton Dickinson Labware (Oxford, UK).
Cell lines and culture conditions
The following human breast cancer cell lines were used: MDA-MB-157, MDA-MB-231, MDA-MB-435s, MDA-MB-436, MDA-MB-453, MCF7, BT549 and ZR751 (purchased from the European Collection of Animal Cell Cultures, Salisbury, UK). Human foetal lung fibroblast cell line MRC5 (from the European Collection of Animal Cell Cultures) and human vascular endothelial cell line HECV (from the Biology and Cellular and Molecular Pathology Department, Naples, Italy) were also used. The cell lines were maintained in Dulbecco's modified Eagle's medium with 10% foetal calf serum, 100 units/ml penicillin and 100 μg/ml streptomycin, and at 37°C in a humid atmosphere of 5% carbon dioxide/95% air.
Construction of SDF-1 expression cassette
Full-length human SDF-1 cDNA was obtained by amplifying the mRNA from normal human fibroblasts, using RT-PCR with the following primers: sdf1exf1 (5'-atgaacgccaaggtcgtg-3'] and SDF1ExR1 (5'-tcacatcttgaacctcttgtt-3'). The discrete SDF-1 product was subsequently TA cloned into pcDNA4/GFP-NT vector (Invitrogen Ltd), followed by transformation using One-Shot E. coli (Invitrogen Ltd), verification, and amplification. Purified plasmid, or control plasmid, was used to transfect MDA-MB-231 cells by electroporation using an electroporator, EasyJet Plus (Flowgen, Boughton, Kent, England, UK), followed by selection with G418 (Sigma). Stable SDF-1 transfectant (MDA-MB-231SDF1+/+), or stable control plasmid transfectant (MDA-MB-231+/-), was subsequently established and verified.
In vitro invasion analysis
This technique was previously reported and modified in our laboratory [21]. Briefly, transwell inserts with 8 μm pore size were coated with 50 μg Matrigel and dried, before being rehydrated. Breast cancer cells (20 × 103) were added to each well. After 96 hours cells that had migrated through the matrix and stuck to the other side of the insert were fixed (4% formalin), stained with 0.5% (weight/volume) crystal violet and counted under a microscope.
Migration assay
The migration assay was based on a method established in our laboratories [22]. Confluent cells were first overlaid with light mineral oil and then placed on a stage heated to 37°C. The cell monolayer was scratched using a fine plastic pipette, creating wounds of approximate 250 μm in width. These wounds were then continuously monitored using a digital camera and time-lapse video recorder. Images were subsequently obtained at 10-min intervals and analyzed using a motion analysis package (Optimas 6) (Optimas Corporation, Bothell, Washington, USA). The accumulated distance that cells travelled over a period of 10 min was analyzed. More than 20 cells were analyzed in each setting, and data were automatically processed using Excel software.
Tissue samples
Tissue samples were collected from patients with breast cancer who had undergone mastectomy. Breast cancer tissue samples (n = 120) and normal mammary tissue samples (from the same patients but away from tumours, and free from tumour cells, as confirmed by subsequent histological analysis; n = 32) were collected immediately after surgery and stored at -80°C until use. Patients were routinely followed clinically after surgery and details were stored in a database. The median follow-up period was 72 months. Details of histology were obtained from pathology reports (Table 1).
RT-PCR and real-time Quantitative PCR
Frozen sections of tissues were cut at a thickness of 5–10 μm and kept for immunohistochemistry and routine histology. An additional 15–20 sections were mixed and homogenized using a hand-held homogenizer, in ice-cold RNA extraction solution. Total RNA extraction from frozen tissues and culture cells was performed using standard RNA isolation kit. The concentration of RNA was determined using an ultraviolet spectrophotometer. Reverse transcription was conducted using a reverse transcription kit with an anchored oligo [dT] primer supplied by AbGene Ltd, using 1 μg total RNA in a 96-well plate. The quality of cDNA was verified using β-actin primers (5'-caggaggttgaaggactaaa-3' and 5'-gggatcagttttctttgtca-3').
Conventional PCR was performed with specific primers for SDF-1 and CXCR4. Amplication conditions were as follows: 94°C for 5 min, followed by 40 cycles of 94°C for 30 s, 55°C for 1 min and 72°C for 1 min. This was followed by a final extension for 5 min at 72°C. The products were visualized on 2% agarose gel after stain with ethidium brominde.
The level of SDF-1 and CXCR4 transcripts from the prepared cDNA was determined using a real-time quantitative PCR, based on Amplifluor technology, modified from a method reported previously [23]. (TCS Biologicals Oxford, England, UK) Briefly, pairs of PCR primers were similarly designed using Beacon Designer software, version 2 (Biosoft International, Palo Alto, California, USA) (primer sequence: sense SDF-1 5'-ttcaggagtacctggagaaa-3', CXCR4 5'-cttcttaactggcattgtgg-3'; antisense SDF-1 5'-actgaacctgaccgtacacctaacactggt-3', CXCR4 5'-actgaacctgaccgtacagtgatgacaaag-3'), but an additional sequence was added to one of the primers [24]. This is known as the Z sequence (5'-actgaacctgaccgtaca-3') which is complementary to the universal Z probe (Intergen Inc, Oxford, UK). The primers used for quantitation of oestrogen receptor (ER) and ER-β were as we reported previously [23] (ER; 5'-cctactacctggagaacgag-3' and 5'-ctcttcggtcttttcgtatg-3'; and ER-β: 5'-aaaagaatcattcaatgaca-3' and 5'-attaacacctccatccaaca-3'). Primers used to quantify CK19 were as previously reported (5'-caggtccgaggttactgac-3' and 5'-actgaacctgaccgtacacactttctgc cagtgtgtcttc-3', respectively) [23,25].
The reaction was carried out using the following: Hot-start Q-master mix (AbGene Ltd); 10 pmol of specific forward primer; 1 pmol reverse primer, which has the Z sequence; 10 pmol of FAMtagged probe (Intergen Inc), and cDNA from ~50 ng of RNA. The reaction was conducted using IcyclerIQ (Bio-Rad, Hemel Hempstead, Herts, England, UK), which is equipped with an optic unit that allows real-time detection of 96 reactions, under the following conditions: 94°C for 12 min and 50 cycles of 94°C for 15 s, 55°C for 40 s, and 72°C for 20 s [15]. The levels of transcripts were generated from a standard that was simultaneously amplified with the samples.
Immunohistochemical staining of SDF-1 proteins
In the present study, normal breast tissue samples (n = 32) and their respective matched breast tumour samples (n = 32) were used for immunohistochmemical analysis. Tissues were frozen and sectioned at a thickness of 6 μm using a cryostat. The sections were mounted on SuperFrostPlus microscope slides (Ramond A Lamb, London, England, UK) and were air-dried and then fixed in a mixture of 50% acetone and 50% methanol. The sections were then placed in Optimax wash buffer (San Ramon, California, USA) for 5–10 min to rehydrate. Sections were incubated for 20 min in a 0.6% bovine serum albumin blocking solution and were then probed with the primary antibody for 1 hour. After extensive washings in buffer, sections were incubated for 30 min in the secondary biotinylated antibody (Multilink Swine anti-goat and anti-rabbit immunoglobulin; Dako Inc., Angel Drove, Ely, Cambridgeshire, England, UK). After washing, avidin biotin complex (Vector Laboratories) was then applied to the sections followed once more by extensive washings. Diaminobenzidine chromogen (Vector Laboratories) was then added to the sections, which were then incubated in the dark for 5 min. Sections were then counterstained in Gill's haematoxylin and were dehydrated in ascending grades of methanol before clearing in xylene and mounting under a coverslip.
Statistical analysis
Statistical analysis was carried out using the Mann–Whitney U-test and the Kruskal–Wallis test, survival analysis was using Kaplan–Meier survival analysis and Cox hazardous proportion analysis, using the SPSS version 11 program (SPSS Inc., Chicago, IL, USA). P < 0.05 was considered statistically significant.
Results
Expression of SDF-1/CXCR4 mRNA in cell lines and in human breast cancer tissures
SDF-1 mRNA was identified in MRC5, MDA-MB-435s, MDA-MB-436 and breast cancer tissues, but not in other breast cancer cell lines and HECV cells. It has been suggested that the MDA-MB-435 cell line is of melanocyte origin, and MDA-MB-436 was the only SDF-1 positive breast cancer cell line of all the lines tested in the present study. In contrast, CXCR4 mRNA expression was detected in all eight breast cancer cell lines, in MRC5 and HECV cells (Fig. 1), and in breast cancer tissue (data not shown). Quantitative analysis of the SDF-1 transcript revealed that breast tumour tissues had high levels of SDF-1 transcript (mean ± standard deviation: 195 ± 103 copies) as compared with normal mammary tissues (85.6 ± 54), but the difference was not statistically significant (P = 0.35). To take into account the contribution made by cellularity in mammary tissues, levels of SDF1 were normalized to the level of CK19. Dispite a higher SDF1:CK19 ratio in tumour tissues (39.3 ± 13.6) than in normal breast tissue (30.7 ± 3.97), the difference was not significant (P = 0.84). With respect to ER, those tumours negative for ER had higher levels of SDF-1 (246 ± 138) than did ER-positive tumours (57.9 ± 45.4; P = 0.20). A similar, insignificant trend was seen with ER-β (248.0 ± 131 for ER-β- tumours and 1.3 ± 0.72 for ER-β+ tumours). The SDF-1:CK19 ratio for ER-negative tumours was 52.7 ± 41.6 and that for ER-positive tumours was 30.8 ± 14.4 (P = 0.62). The ratio was 41.2 ± 17.3 for ER-β-negative and 8.3 ± 5.1 for ER-β-positive tumours (P = 0.072).
SDF-1 has the potential to promote invasion and migration
MDA-MB-231SDF1+/+ cells, which stably expressed SDF-1 (Fig. 2a), and MDA-MB-231+/- (stable control plasmid transfectant) and wild-type MDA-MB-231 cells, which were SDF-1 negative, were tested for their invasiveness and migration. MDA-MB-231SDF1+/+ cells exhibited greater invasiveness through Matrigel than did wild-type and MDA-MB-231+/- cells (P < 0.01; Fig. 2b). In addition, the migration speed of MDA-MB-231SDF1+/+ cells was markedly increased compared with the respective controls (Fig. 2c).
SDF-1/CXCR4 immunohistochemical staining in human breast cancer
Immunohistochemical staining confirmed expression of SDF-1 at the protein level in breast cancer tissue samples. In contrast to the adjacent nonmalignant tissue, we were able to demonstrate heterogeneous but consistent expression of SDF-1 antigen in tumour tissue. Immunohistochemical staining of SDF-1 appeared in most tumour cells and in stromal cells (Fig. 3a,b). As expected, staining of CXCR4 were seen in both normal and tumour cells (Fig. 3c,d), with staining in tumour cells being markedly stronger.
SDF-1 expression and lymphatic nodal status, histological types, grades and staging
We analyzed the levels of SDF-1 in relation to nodal status (Fig. 4a). Node-positive tumours had significantly higher levels of SDF-1 than did node-negative ones. The expression level of SDF-1 tended to be higher in the node-positive group, although there was no statistically significant difference between node-positive and node-negative groups (P = 0.05). The data were further analyzed by dividing node-positive and node-negative tumours into ER-positive and ER-negative groups. For SDF-1 no significant differences between subgroups were observed (ER-/node- versus ER-/node+, P = 0.25; ER-/node- versus ER+/node-, P = 0.57; P values for SDF1:CK19 were 0.27 and 0.32, respectively). No significant difference in SDF-1 was seen between ER-positive/node-negative and ER-positive/node-positive subgroups (P = 0.24; for SDF1:CK19 P = 0.27). Similarly, when node-positive and node-negative tumours were subdivided into ER-β-positive and ER-β-negative subgroups, no significant difference was seen.
We examined expression of SDF-1 relative to tumor types, grade and staging (Table 2). There was a trend in the differences in SDF1 expression between tumour grades, in that grade 3 and grade 2 tumours tended to have higher SDF1 levels than did grade 1 tumours, but this was not statistically significant. There were no significant relations between expression level of SDF-1 and tumor type and stage.
SDF-1 expression correlated with prognosis and long term survival
The expression level of SDF-1 correlated with clinical outcome; patients with local recurrence (P = 0.05) and those who died from breast cancer (P = 0.03) had signfiantly higher levels of SDF-1 transcript (Fig. 4b). Those patients with metastasis and local recurrence, and who died from breast cancer had significantly higher levels of SDF-1 than did the disease-free group (P = 0.01; Fig. 4c).
To determine whether SDF-1 transcript levels were associated with long-term survival, we divided patients into those with high levels (n = 79) and those with low levels (n = 41) of SDF-1. The cutoff point was determined using the Nottingham Prognostic Index, and was set at the level at which patients had moderate prognoses (Nottingham Prognostic Index 3.4–5.4). As shown in the Kaplan–Meier survival curve (Fig. 5), high levels of SDF-1 significantly correlated with shorter overall survival (mean survival 94.1 months [95% confidence interval 65.4–122.9 months] versus 143.6 months [95% confidence interval 135.2–152.0 months] months for those with low levels of SDF-1; P = 0.001; Fig. 5a). Further analysis taking tumour grade into account was not possible because the sample number in each subgroup was too small. Similarly, high SDF-1 levels were associated with reduced incidence-free survival (P = 0.035 by Cox proportion analysis; Fig. 5b).
Discussion
Chemokines are a family of small molecular weight proteins (8–10 kDa) that are classified into four distinct groups, depending on the positioning of the cysteine motif at the NH2 terminus. The family members include CXC, CC, C and CXXXC chemokines [26,27]. The specific effects of chemokines on their target cells are mediated by members of a family of seven-transmembrane-spanning, G-protein-coupled receptors [14,28].
SDF-1 is a member of the CXC subfamily of chemokines and its receptor is CXCR4. SDF-1 is constitutively expressed in various organs including bone, lung, liver, brain, thymus and lymph nodes [10,14,19], but SDF-1 is mainly produced by stromal cells, such as osteoblasts, fibroblasts and endothelial cells in the bone marrow [29,30]. Despite numerous studies on CXCR4 in breast cancer, reports on SDF-1 in human breast cancer are limited.
In the present study the expression of CXCR4 was detected in various cell lines and in malignant and nonmalignant breast tissues, but SDF-1 expression was only observed in two out of the eight breast cancer cell lines and in the fibroblast cell line MRC5. These results indicate that certain breast cancer cells co-express SDF-1 and CXCR4, which may act as a potential autocrine mechanism in breast cancer. We have reported that the fibroblast cell line, MRC5, strongly expressed SDF-1. Furthermore, in the present study immunohistochemical staining of SDF-1 was apparent in most tumour cells and in stromal cells. Collectively, from the results, we suggest that SDF-1 in breast cancer is produced by both tumour cells and stromal cells. The other potential source is the infiltrated immune cells, which frequently express CXCR4 and SDF1. The present study did not examine the proportion of these cells that produced SDF1 or the degree of expression, which would be an interesting focus for future studies.
The present study provides strong evidence that, when the SDF-1/CXCR4 complex existed (i.e. in MDA-MB-231SDF1+/+ cells, which expressed both SDF-1 and CXCR4), breast cancer cells exhibited significant increases in invasiveness and faster migration. These findings suggest that breast cancer cells that co-express SDF-1 and CXCR4 may be more aggressive. In the present study we were unable to transfect fibroblasts with the current bacterial vector because no fibroblasts subsequently survived the electroporation and genetic marker selection process. It will be useful to develop viral expression for the purpose for future work. In addition, high levels of SDF-1 expression tended to be present in grade 3 and grade 2 tumors as compared with grade 1 tumours, further supporting the contention that breast cancer cells that express high levels of SDF-1 are more invasive.
Recently, studies implicated CXCR4 in chemotaxis, invasiveness and metastasis of tumours, particularly in metastasis of breast cancer, in an organ-specific manner. Muller and coworkers [19] found CXCR4 to be highly expressed in breast cancer cells, malignant breast tumours and metastases. On the other hand, peak levels of CXC chemokine ligand (CXCL)12 occurred in those organs that represent the initial destinations of breast cancer metastasis (i.e. lymph nodes, lung, liver and bone marrow). Furthermore, neutralizing the interaction between CXCL12 and CXCR4 significantly impaired metastasis to regional lymph nodes and lung in mice. Other reports have also shown that the SDF-1/CXCR4 biological axis is involved in regulating metastasis of tumours [6,18,31,32]. In the present study we found that that node-positive tumours had significantly higher levels of SDF-1 than did node-negative tumors, suggesting that SDF-1 may be involved in the lymph node metastatic process. Given that lymph node metastasis directly affects the prognosis of patients with breast cancer [4], we propose that SDF-1, via the CXCR4 pathway, is potentially a marker of nodal involvement. It was recently reported that SDF-1 can act as a direct target for ER-α in breast cancer cells (e.g. MCF-7 cells) [33,34]. In the present study it is noteworthy that EF-negative and ER-β-negative tumours tended to have higher levels of SDF-1. Although differences between these subgroups were not statistically significant, the trend, together with the in vitro studies, indicate that this link warrants further investigation. It is also noteworthy that SDF-1 expression in mammary tissues was primarily confined to stromal cells and, to some degree, cancer cells. We did not observe SDF-1 staining in vascular endothelial cells, HECV, and in vascular endothelial cells in the tissues – observations echoed by other studies [35,36]. This finding indicates that paracrine regulation may be the main pathway in breast cancer but that autocrine pathways may also exist. Secretion and production of SDF-1 are regulated by other factors. For example, expression of SDF-1 is decreased by IL-1, tumour necrosis factor and inflammation [37], whereas oestradiol can induce the production and secretion of SDF-1 in breast cancer cells [38]. On the other hand, tumour cells exposed to high concentrations of SDF-1 induce reduction in CXCR4 expression [18]. Furthermore, vascular edothelial grwoth factor can also induce CXCR4 expression in breast cancer cells [39,40]. Factors contributing to over-expression of SDF-1 in breast cancer thus warrant further investigation.
Finally, we demonstrated a significant correlation between SDF-1 expression and overall and disease-free survival in patients with breast cancer. The high level of SDF-1 expression suggests that there is a high likelihood of node metastasis, local recurrence and death from breast cancer in these patients. We and others found the expression pattern of CXCR4 to be significantly correlated with the degree of lymph node metastases but not with haematogenous metastases [41-43]. Therefore, SDF-1, together with its receptor CXCR4, may have potential value when assessing long-term clinical outcome in breast cancer.
Conclusion
The present study demonstrated that breast cancer cells that express SDF-1, and therefore that have an active SDF-1/CXCR4 pathway, are more invasive and motile, thus have a more aggressive phenotype. In clinical breast cancers, and supported by data from cell lines, we found that SDF-1 appears to exist primarily in stromal cells and, to some degree, in breast cancer cells. That levels of SDF-1 are significantly correlated with nodal status, recurrence and, most notably, both overall and disease-free survival indicates that SDF-1 – and indeed the SDF-1 receptor complex – have strong predictive value in assessing long-term clinical outcome.
Abbreviations
CXCL = CXC chemokine ligand; CXCR = CXC chemokine receptor; ER = oestrogen receptor; RT-PCR = reverse transcription polymerase chain reaction; SDF = Stromal cell derived factor.
Authors' contributions
HK carried out in vitro testing and data analysis, and prepared the manuscript. GW conducted the immunohistochemistry study. CP contributed to the screening and ribozyme work. ADJ contributed to histological analysis. REM contributed to clinical follow ups. WGJ contributed to the study design, design of ribozymes, quantitative analysis of SDF1 transcript and statistical analysis.
Acknowledgements
We thank Breast Cancer Campaign for supporting WGJ and CP. Dr Kang is an international fellow of the Overseas Scholar Scheme.
Figures and Tables
Figure 1 SDF-1/CXCR4 expression in various cell lines. 1: MDA-MB-157; 2: MDA-MB-231; 3: MDA-MB-435s; 4: MDA-MB-436; 5: MDA-MB-453; 6: MCF7; 7: BT549; 8: ZR751; 9: MRC5; 10: HECV; 11: negative control. CXCR, CXC chemokine receptor; SDF, stromal cell-derived factor.
Figure 2 Manipulation of expression of SDF-1 in breast cancer cells. (a) The efficiency of stromal cell-derived factor (SDF)-1 transfected in MDA-MB-231 cells was confirmed by PCR. M: marker; 1: negative control; 2: MDA-MB-231 wild-type; 3: empty vector control MDA-MB-231+/-; 4: SDF-1-transfected MDA-MB-231SDF1+/+. (b) Invasiveness of transfected cells. *P < 0.01 versus control and wild-type. (c) Cellular migration.
Figure 3 Immunohistochemical analysis of SDF1 and its receptor. Imunohistochemical staining of (a,b) SDF-1 and (c,d) the SDF-1 receptor CXCR4 in mammary tissues. The left panels show normal tissues, and the right panels show breast tumour tissues. CXCR, CXC chemokine receptor; SDF, stromal cell-derived factor.
Figure 4 Levels of SDF-1 transcript in human breast tumours. (a) Stromal cell-derived factor (SDF)-1 expression level and lymph node metastasis, showing SDF-1 expression in node-negative and node-positive samples (0.89 ± 0.47 versus 399 ± 210; P = 0.05). (b) Significantly raised SDF-1 transcript in patients with local recurrence and with mortality. (c) Expression level of SDF-1 and clinical outcome (disease-free versus poor out come: 0.83 ± 0.35 versus 670 ± 346; P = 0.01).
Figure 5 Kaplan–Meier survival curves. (a) Overall survival (P = 0.01). (b) Disease-free survival (P = 0.035). Median follow up: 72.2 months. Stromal cell-derived factor (SDF)-1 (H), patients with high levels of SDF-1 transcript (n = 79); SDF-1 (L), patients with low levels of SDF-1 transcript (n = 41).
Table 1 Clinical features of patients included in the study
Clinical feature n
Node status
Node negative 65
Node positive 55
Grade
1 23
2 41
3 56
Histology
Ductal 88
Lobular 14
Others 8
TNM staging
1 69
2 40
3 7
4 4
Clinical outcome
Disease free 87
With metastasis 6
With local recurrence 5
Died from breast cancer 16
Died of unrelated disease 6
Table 2 SDF-1 expression and correlation with clinical pathology
Clinical pathology SDF-1 level (mean ± SD) P
Type
Ductal 237 ± 131
Lobular 88.9 ± 88.8 0.67
Others 0.89 ± 0.9
Grade
Grade 1 12.8 ± 12.5
Grade 2 27 ± 27 0.07
Grade 3 371 ± 206 0.08
Staging
TNM1 5.6 ± 4.5
TNM2 290 ± 165 0.09
TNM3 1628 ± 1530 0.34
TNM4 1.1 ± 1.1
SDF, stromal cell-derived factor.
==== Refs
Parkin DM Pisani P Ferlay J Global cancer statistics Ca Cancer J Clin 1999 49 33 64 10200776
Lacey JV JrDevesa SS Brinton LA Recent trends in breast cancer incidence and mortality Environ Mol Mutagen 2002 39 82 88 11921173 10.1002/em.10062
Figueroa JA Yee D McGuire WL Prognostic indicators in early breast cancer Am J Med Sci 1993 305 176 182 7680528
Jatoi I Hilsenbeck SG Clark GM Osborne CK Significance of axillary lymph node metastasis in primary breast cancer J Clin Oncol 1999 17 2334 2340 10561295
Moore MAS The role of chemoattraction in cancer metastases BioEssays 2001 23 674 676 11494314 10.1002/bies.1095
Phillips RJ Burdick MD Lutz M Belperio JA Keane MP Strieter RM The stromal derived factor-1/CXCL12-CXC chemokine receptor4 biological axis in non–small cell lung cancer metastases Am J Respir Crit Care Med 2003 167 1676 1686 12626353 10.1164/rccm.200301-071OC
Hart IR Fidler IJ Role of organ selectivity in the determination of metastatic patterns of B16 melanoma Cancer Res 1980 40 2281 2287 7388794
Nagasawa T Kikutani H Kishimoto T Molecular cloning and structure of a pre-B-cell growth-stimulating factor Proc Natl Acad Sci USA 1994 91 2305 2309 8134392
Tashiro K Tada H Heilker R Shirozu M Nakano T Honjo T Signal sequence trap: a cloning strategy for secreted proteins and type I membrane proteins Science 1993 261 600 603 8342023
Shirozu M Nakano T Inazawa J Tashiro K Tada H Shinohara T Honjo T Structure and chromosomal localization of the human stromal cell-derived factor 1 (SDF1) gene Genomics 1995 28 495 500 7490086 10.1006/geno.1995.1180
Fernandis AZ Prasad A Band H Klosel R Ganju RK Regulation of CXCR4 mediated chemotaxis and chemoinvasion of breast cancer cells Oncogene 2004 23 157 167 14712221 10.1038/sj.onc.1206910
Libura J Drukala J Majka M Tomescu O Navenot JM Kucia M Marquez L Peiper SC Barr FG Janowska-Wieczorek A Ratajczak MZ CXCR4-SDF-1 signaling is active in rhabdomyosarcoma cells and regulates locomotion, chemotaxis, and adhesion Blood 2002 100 2597 2606 12239174 10.1182/blood-2002-01-0031
Zou YR Kottmann AH Kuroda M Taniuchi I Littman DR Function of the chemokine receptor CXCR4 in haematopoiesis and in cerebellar development Nature 1998 393 595 599 9634238 10.1038/31269
Nagasawa UT Tachibana UK Kishimoto T A novel CXC chemokine PBSF/SDF-1 and its receptor CXCR4: their functions in development, hematopoiesis and HIV infection Semin Immunol 1998 10 179 185 9653044 10.1006/smim.1998.0128
Tachibana K Hirota S Iizasa H Yoshida H Kawabata K Kataoka Y Kitamura Y Matsushima K Yoshida N Nishikawa S The chemokine receptor CXCR4 is essential for vascularization of the gastrointestinal tract Nature 1998 393 591 594 9634237 10.1038/31261
Salcedo R Wasserman K Young HA Grimm MC Howard OM Anver MR Kleinman HK Murphy WJ Oppenheim JJ Vascular endothelial growth factor and basic fibroblast growth factor induce expression of CXCR4 on human endothelial cells: in vivo neovascularization induced by stromal derived factor-1α Am J Pathol 1999 154 1125 1135 10233851
Yamaguchi J Kusano KF Masuo O Kawamoto A Silver M Murasawa S Bosch-Marce M Masuda H Losordo DW Isner JM Stromal cell-derived factor-1 effects on ex vivo expanded endothelial progenitor cell recruitment for ischemic neovascularization Circulation 2003 107 1322 1328 12628955 10.1161/01.CIR.0000055313.77510.22
Geminder H Sagi-Assif O Goldberg L Meshel T Rechavi G Witz IP Ben-Baruch A A possible role for CXCR4 and its ligand, the CXC chemokine stromal cell-derived factor-1, in the development of bone marrow metastases in neuroblastoma J Immunol 2001 167 4747 4757 11591806
Muller A Homey B Soto H Ge N Catron D Buchanan ME McClanahan T Murphy E Yuan W Wagner SN Involvement of chemokine receptors in breast cancer metastasis Nature 2001 410 50 56 11242036 10.1038/35065016
Zeelenberg IS Ruuls-Van Stalle L Roos E The chemokine receptor CXCR4 is required for outgrowth of colon carcinoma micrometastases Cancer Res 2003 63 3833 3839 12839981
Jiang WG Hiscox S Hallett MB Horrobin DF Mansel RE Puntis MCA Regulation of the expression of E-cadherin on human cancer cells by gamma linolenic acid Cancer Res 1995 55 5043 5048 7585549
Jiang WG Hiscox S Cai J Martin T Matsumoto K Nakamura T Mansel RE Antagonistic effects of NK4, an novel HGF variant, on the in vitro angiogenesis of human vascular endothelial cells Clin Cancer Res 1999 5 3695 3703 10589789
Jiang WG Watkins G Fodstad O Douglas-Jones A Mokbel K Mansel RE Differential expression of the CCN family members Cyr61 from CTGF and Nov in human breast cancer Endocr Relat Cancer 2004 11 781 791 15613452 10.1677/erc.1.00825
King JAC Ofori-Acquah AF Stevens T Al-Mehdi AB Fodstad O Jiang WG Prognostic value of ALCAM in human breast cancer Breast Cancer Res 2004 6 R478 R487 15318930 10.1186/bcr815
Jiang WG Douglas-Jones A Mansel RE Level of expression of PPAR-gamma and its co-activator (PPAR-GCA) in human breast cancer Int J Cancer 2003 106 752 757 12866036 10.1002/ijc.11302
Rossi D Zlotnik A The biology of chemokines and their receptors Annu Rev Immunol 2000 18 217 242 10837058 10.1146/annurev.immunol.18.1.217
Murdoch C CXCR4: chemokine receptor extraordinaire Immunol Rev 2000 177 175 184 11138774 10.1034/j.1600-065X.2000.17715.x
Feng Y Broder CC Kennedy PE Berger EA HIV-1 entry cofactor: functional cDNA cloning of a seven-transmembrane, G protein-coupled receptor Science 1996 272 872 877 8629022
Yun HJ Jo DY Production of stromal cell-derived factor-1 (SDF-1) and expression of CXCR4 in human bone marrow endothelial cells J Korean Med Sci 2003 18 679 685 14555820
Aust G Steinert M Kiessling S Kamprad M Simchen C. Reduced expression of stromal-derived factor 1 in autonomous thyroid adenomas and its regulation in thyroid-derived cells J Clin Endocrinol Metab 2001 86 3368 3376 11443213 10.1210/jc.86.7.3368
Koshiba T Hosotani R Miyamoto Y Ida J Tsuji S Nakajima S Kawaguchi M Kobayashi H Doi R Hori T Expression of stromal cell-derived factor 1 and CXCR4 ligand receptor system in pancreatic cancer: a possible role for tumor progression Clin Cancer Res 2000 6 3530 3535 10999740
Bartolome RA Galvez BG Longo N Baleux F Van Muijen GN Sanchez-Mateos P Arroyo AG Teixido J Stromal cell-derived factor-1 promotes melanoma cell invasion across basement membranes involving stimulation of membrane-type 1 matrix metalloproteinase and Rho GTPase activities Cancer Res 2004 64 2534 2543 15059909
Hall JM Korach KS Stromal cell-derived factor 1, a novel target of estrogen receptor action, mediates the mitogenic effects of estradiol in ovarian and breast cancer cells Mol Endocrinol 2003 17 792 803 12586845 10.1210/me.2002-0438
Li YM Pan Y Wei Y Cheng X Zhou BP Tan M Zhou X Xia W Hortobagyi GN Yu D Upregulation of CXCR4 is essential for HER2-mediated tumor metastasis Cancer Cell 2004 6 459 469 15542430 10.1016/j.ccr.2004.09.027
Imai K Kobayashi M Wang J Shinobu N Yoshida H Hamada J Shindo M Higashino F Tanaka J Asaka M Selective secretion of chemoattractants for haemopoietic progenitor cells by bone marrow endothelial cells: a possible role in homing of haemopoietic progenitor cells to bone marrow Br J Haematol 1999 106 905 911 10519991 10.1046/j.1365-2141.1999.01644.x
Arai J Yasukawa M Yakushijin Y Miyazaki T Fujita S Stromal cells in lymph nodes attract B-lymphoma cells via production of stromal cell-derived factor-1 Eur J Haematol 2000 64 323 332 10863978 10.1034/j.1600-0609.2000.90147.x
Fedyk ER Jones D Critchley HO Phipps RP Blieden TM Springer TA Expression of stromal-derived factor-1 is decreased by IL-1 and TNF and in dermal wound healing J Immunol 2001 166 5749 5755 11313418
Hall JM Korach KS Stromal cell-derived factor 1 (SDF-1), a novel target of estrogen receptor action, mediates the mitogenic effects of estradiol in ovarian and breast cancer cells Mol Endocrinol 2003 17 792 803 12586845 10.1210/me.2002-0438
Helbig G Christopherson KW IIBhat-Nakshatri P Kumar S Kishimoto H Miller KD Broxmeyer HE Nakshatri H NF-κB promotes breast cancer cell migration and metastasis by inducing the expression of the chemokine receptor CXCR4 J Biol Chem 2003 278 21631 21638 12690099 10.1074/jbc.M300609200
Bachelder RE Wendt MA Mercurio AM Vascular endothelial growth factor promotes breast carcinoma invasion in an autocrine manner by regulating the chemokine receptor CXCR41 Cancer Res 2002 62 7203 7206 12499259
Kang H Mansel RE Jiang WG The role of stroma-derived factor 1 (SDF-1) in the migration and invasion of breast cancer cells Int J Oncol 2005
Kang W Watkins G Douglas-Jones A Mansel RE Jiang WG The elevated level of CXCR4 expression is correlated with lymph node metastasis in human breast cancer Breast 2005
Kato M Kitayama J Kazama S Nagawa H Expression pattern of CXC chemokine receptor-4 is correlated with lymph node metastasis in human invasive ductal carcinoma Breast Cancer Res 2003 5 R144 R150 12927045 10.1186/bcr627
| 15987445 | PMC1175055 | CC BY | 2021-01-04 16:04:33 | no | Breast Cancer Res. 2005 Apr 4; 7(4):R402-R410 | utf-8 | Breast Cancer Res | 2,005 | 10.1186/bcr1022 | oa_comm |
==== Front
Breast Cancer ResBreast Cancer Research1465-54111465-542XBioMed Central London bcr10251598745110.1186/bcr1025Research ArticleRelationship of patients' age to histopathological features of breast tumours in BRCA1 and BRCA2 and mutation-negative breast cancer families Eerola Hannaleena [email protected]ä Päivi [email protected] Anitta [email protected]äki Kristiina [email protected] Carl [email protected] Heli [email protected] Department of Oncology, Helsinki University Central Hospital, Finland2 Department of Obstetrics and Gynecology, Helsinki University Central Hospital, Finland3 Department of Pathology, Helsinki University Central Hospital, Finland4 Department of Clinical Genetics, Helsinki University Central Hospital, Finland2005 21 4 2005 7 4 R465 R469 21 12 2004 8 2 2005 15 3 2005 17 3 2005 Copyright © 2005 Eerola et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Introduction
Our aim was to evaluate the relationship of patients' age to histopathological features of hereditary breast tumours in a series of breast cancer families not selected for age at diagnosis. In sporadic breast cancer, tumours from premenopausal patients have been shown to differ from those of postmenopausal patients, but this phenomenon has been little studied among familial patients.
Methods
Representative areas of all available breast cancer tissue specimens (n = 262) from 25 BRCA1, 20 BRCA2, and 74 non-BRCA1/2 breast cancer families were punched into a tissue microarray. Immunohistochemical staining of oestrogen receptor, progesterone receptor, ERBB2, and p53 as well as the histology and grade of tumours in these three groups of families were studied in different age groups and compared with each other.
Results
We found that only breast cancers from young (<50 years) BRCA1+ patients represent features documented as being typical of BRCA1-associated cancers, such as high tumour grade, negativity for oestrogen and progesterone receptors, and overexpression of p53. Among the BRCA2 families, the opposite was found, with a significantly higher frequency of tumours negative for oestrogen and progesterone receptors among the older patients than among the other groups, but no distinctive tumour characteristics among the younger BRCA2 patients.
Conclusion
Tumours of BRCA1 and BRCA2 carriers aged 50 years or more differed significantly from those of younger carriers. This difference may reflect different biological behaviour and pathways of tumour development among the older and the younger BRCA1 and BRCA2 patients, with impact also on prognosis and survival.
==== Body
Introduction
Distinct pathological features among BRCA1-associated tumours have been found when such tumours are compared with sporadic cancers; these features include high tumour grade, negativity for oestrogen receptor (ER), overexpression of p53, negativity for progesterone receptor (PR), and a higher proportion of medullary and atypical medullary carcinomas [1-3]. Recently, cDNA expression analyses have suggested a basal epithelial phenotype for BRCA1 tumors [4] and expression of cytokeratins 5/6 have been associated with BRCA1 tumours [5]. Among BRCA2-associated tumours, findings have been inconsistent, and in most cases no significant difference has been found between BRCA2-associated and sporadic cancers [1,2,6,7].
In our previous report [8], we have shown, consistent with earlier studies, that BRCA1-associated cancers were diagnosed younger and were more ER- and progesterone-receptor-negative (PR-), more p53+, and of higher grade than unselected breast tumours or tumours from non-BRCA1/2 breast cancer families. However, in multivariate analysis the independent factors, as compared with non-BRCA1/2 tumours, were age of diagnosis, grade, and PR-negativity. BRCA2 cases did not have such distinctive features compared with non-BRCA1/2 cases.
In large studies on sporadic breast cancer tumours, tumours from premenopausal patients have been shown to differ from those of postmenopausal patients [9-11], but this has been little studied among familial patients. In this study, we had an excellent opportunity to study familial cases without age restriction and to evaluate whether the histology and immunohistochemistry differ in the different age groups (according to whether age of diagnosis is below or over 50 years, with age being used as a surrogate for menopause status) among BRCA1, BRCA2, and non-BRCA1/2 families.
Materials and methods
Family history of cancer was screened for among breast cancer patients in the Department of Oncology, Helsinki University Central Hospital [12]. Families were collected with a simple criterion of at least three first- or second-degree relatives with breast or ovarian cancer, with no restriction regarding age. All the families were tested for BRCA1 and BRCA2 mutations by mutation analysis of the whole coding sequences and exon/intron boundaries of the genes as described elsewhere [13,14], or were tested for all 18 previously reported Finnish BRCA1 and BRCA2 mutations [13-16]. In this study, as previously described [8], we collected all the available paraffin-wax blocks of the primary breast cancers (n = 262) from 119 breast cancer families. Altogether, 51 cancers from the 25 BRCA1 families, 59 cancers from the 20 BRCA2 families, and 152 cancers from the 74 non-BRCA1/2 families were obtained.
The patients' median age at diagnosis of the tumours was 44 years for BRCA1, 47 years for BRCA2, and 55 years for non-BRCA1/2. For comparison of tumours from premenopausal and postmenopausal patients, the age of 50 years was chosen as a surrogate for menopause. Among the BRCA1 patients, 34 (66.7%) were diagnosed when they were below 50 years of age (median age 39) and 17 (33.3%) when they were 50 or more (median age 55); the respective numbers among BRCA2 patients were 35 (59.3%) and 24 (40.7%) (median ages 39 and 56.5, respectively), and among non-BRCA1/2 patients, 58 (38.2%) and 94 (61.8%) (median ages 44 and 65, respectively).
The most representative area of the tumour was punched to produce a hereditary breast cancer tissue microarray including two cores (diameter 0.6 mm) from all the original blocks as described elsewhere [8,17]. The use of microarray tissue blocks makes it possible to stain all the samples at the same time and in the same conditions. Subgroups of the material are therefore very well comparable, and a highly significant correlation between this kind of multicore system and studying the whole sections of the original blocks has also been shown [18,19].
All the tissue microarray slides were stained with routine methods used for pathological diagnostics with ER, PR, ERBB2, and p53 antibodies in the same laboratory [8]. Briefly, five-micrometer sections were cut from paraffin-embedded blocks, dewaxed in xylene, and dehydrated in a series of graded alcohols. The sections were pretreated in a microwave oven and incubated with antibody overnight. ER antibody (1:50) and ERBB2 antibody (NCL-CB11, 1:400) were purchased from Novocastra (Newcastle upon Tyne, UK), and PR (1:250) and p53 antibodies (1:100) were from Dako (Copenhagen, Denmark). The evaluation of the staining results was similar to that used in routine diagnostics, and samples were considered positive when 10%, 10%, and 20% of the cells were stained with ER, PR, and p53, respectively. Samples having a moderate or intense staining of the entire membrane in more than 10% of the tumour cells (immunohistochemical scores of 2+ and 3+) were considered to be ERBB2+. Other staining patterns (0 and 1+) were considered to be negative. We studied the haematoxylin-and-eosin sections of the original blocks to achieve histological diagnosis and grading (all by the same pathologist (PH)). Statistical analysis was done with SPSS version 8.0 for Windows. We tested the differences in dichotomous variables with a χ2 or Fisher's exact test. All P values are two-tailed.
Permissions for this study were obtained from the ethics committees of the Department of Oncology and the Department of Obstetrics and Gynaecology, Helsinki University Central Hospital, and of the Ministry of Social Affairs and Health in Finland. Blood and tumour samples were used in this study with the informed consent of the probands and of the family members.
Results
In BRCA1 families, patients whose cancer was diagnosed when they were under 50 years of age differed significantly from those diagnosed at 50 years or older in the presence of grade 3 tumours (84.4% vs 47.1%, P = 0.009), ER-negativity (83.3% vs 25%, P = 0,001), and p53-positivity (50.0% vs 7.7%, P = 0.014) (Table 1). All of the five cancers with medullary histology were also detected in patients under 50 years old. Patients who were BRCA1+ and were under 50 years old at diagnosis differed significantly in all of these factors from familial non-BRCA1/2 patients (proportion of grade 3 tumours, 84.4% vs 17.3%, respectively, P ≤ 0.0005; of ER-negativity, 83.3% vs 29.3%, P ≤ 0.0005; of PR-negativity, 90.3% vs 31.0%, P ≤ 0.0005; and of p53-positivity, 50.0% vs 25.9%, P = 0.024). However, patients from BRCA1 families diagnosed at age 50 years or older differed significantly only for grade from the non-BRCA1/2 patients in the same age group (47.1% vs 23.3, P = 0.044).
In BRCA2 families, tumours of patients diagnosed at less than 50 years of age differed significantly from those of the older patients for ER-negativity (20.6% vs 52.6%, respectively, P = 0.017) and PR-receptor negativity (35.3% vs 80.0%, P = 0.001) (Table 1). In contrast to BRCA1 tumours, the BRCA2 tumours diagnosed in patients 50 years or older were more often ER- (52.6% vs 25.6%, P = 0.02) and PR- (80.0% vs 54.4%, P = 0.036) than non-BRCA1/2 cancers among the same age group. Tumours of patients diagnosed at less than 50 years of age were very similar to non-BRCA1/2 tumours in the same age group (Table 1).
Pathological features of non-BRCA1/2 tumours did not vary significantly between the two age groups.
Discussion
In this study, we have evaluated whether tumour histology and immunohistochemistry are influenced by age of onset (menopause status) among families with BRCA1, BRCA2, or non-BRCA1/2 tumours. Most of the earlier studies of the characteristics of tumours in BRCA1 and BRCA2 carriers have been based on young patients only. Because there was no age restriction in our selection criterion, we had an excellent opportunity to study patients within the whole age distribution.
In BRCA1 families, tumours from patients diagnosed at over 50 years of age were surprisingly different from those in BRCA1 carriers diagnosed at under 50 years. Only tumours from the younger patients exhibited the distinctive characteristics that have been found to be typical of BRCA1 tumours, with higher grade, negativity for ER and PR, and positivity for p53 distinguishing them from familial non-BRCA1/2 tumours. However, tumours from the older patients in BRCA1 families differed significantly only in grade from tumours in non-BRCA1/2 patients. There were only five cases among this older group of patients, for which the BRCA1 mutation status was unknown. If these patients are excluded from the analysis, the observed frequencies remain ; therefore those do not account for the result.
Previously, Vaziri and colleagues [20] have reported that the tumour immunophenotype of BRCA1-carriers is influenced by the age of diagnosis. As a control group, those authors used age-matched breast cancer patients unselected for family history, whereas in our study we included familial non-BRCA1/2 cancer cases. Vaziri and colleagues observed no differences in ER or PR staining of tumours between BRCA1 carriers diagnosed at 50 years or older and controls with sporadic cancers [20]. Foulkes and colleagues [21] also recently reported that the proportion of ER+ tumours increased with patients' age among the BRCA1 patients included in their study (diagnosed at less than 65 years of age), although they found a strong relationship between BRCA1 carrier status and ER-negativity of tumours in the age group 55 to 65 years. Vaziri and colleagues also studied the expression of markers Ki-67, Cyclin D1, p53, and ERBB2. None of these markers differed significantly in the patients 50 years or older between BRCA1-associated cancers and control cancers, although tumours from the younger BRCA1 age group presented less frequent ER, PR, and cyclin D1 staining and more frequent Ki-67 and B-catenin staining than those from control cancers. p53 expression did not differ in their study in different age groups, nor was p53 more frequently overexpressed among young BRCA1 patients than in controls.
We did not find the BRCA2-associated tumours to differ significantly from familial non-BRCA1/2 tumours among the younger age group. However, tumours of BRCA2 carrier patients diagnosed at 50 years or older had more distinctive features, and were more ER- and PR-, than tumours of younger patients or tumours of the same age group of BRCA1 patients or non-BRCA1/2 patients.
The specific features of BRCA1-associated tumours among the younger age group, and lack of such features among the BRCA2-associated tumours, are consistent with the overall characteristics reported previously among BRCA1 and BRCA2 patients [1,2]. Such features characterise to a large extent the BRCA1 and BRCA2 tumours overall, as a large majority (63% in this study) of all breast tumours in the BRCA1 and BRCA2 families are diagnosed before patients reach 50 years of age.
However, among both BRCA1 and BRCA2 families, tumours from older patients form subgroups that are distinctly different from those of the younger patients. Tumours from the older BRCA1 patients resemble more those among the mutation-negative families, or sporadic tumours. The highest incidence rates and relative breast cancer risk among BRCA1 carriers are seen before age 50 [22], and some tumours from older BRCA1 mutation carriers could also be 'sporadic' cancers. However, the breast tumours from older BRCA1 patients also differed from mutation-negative ones by their higher grade. Furthermore, tumours from the older BRCA2 carriers exhibited distinctly different characteristics from the younger ones or from BRCA1 carrier tumours and mutation-negative ones, suggesting a strong impact of the germline mutation on tumour development among the older patients. It is interesting that the BRCA1 and BRCA2 tumours appear to be opposites with respect to their characteristics in the age groups of younger and older patients.
There are now many models and computer programs to test the probability of BRCA1 or BRCA2 mutations [23-28]. We have also documented previously that efficient predictors for BRCA1 and BRCA2 mutations are early age of breast cancer onset and number of ovarian cancer cases in the family [27]. Simple family history criteria of the strongest predictors (onset of breast cancer under age 40 and presence of ovarian cancer) for a mutation may also provide a rough estimate of a high likelihood of carrying a mutation [27].
However, it would be useful if, besides family history, histopathological markers could also be used to distinguish patients and families likely to carry a BRCA1/2 germline mutation from mutation-negative families and breast cancer patients in general. The use of morphologic and immunohistochemical data has been previously suggested to provide a helpful and cost-effective tool for predicting BRCA1 mutation among young breast cancer patients [2,29,30]. The findings here provide further information specifically with respect to older BRCA1 and BRCA2 patients and warrant further studies for evaluating the probability of mutation by combining information on family history and tumour characteristics in the various age groups.
Conclusion
These findings may reflect different biological behaviour and pathway of tumour development among the older and the younger BRCA1 and BRCA2 patients, with impact also on prognosis and survival. So far, results on survival among BRCA1 and BRCA2 patients have been inconclusive or contradictory, and large meta-analyses specifically according to age groups could shed further light on this. Finally, in the context of genetic counselling, specific tumour characteristics may help evaluate the possibility of a BRCA1 or BRCA2 mutation and the need for mutation testing in a family with a history of breast cancer. It appears crucial, however, to consider such features specifically with respect to the age of the patients.
Abbreviations
ER = oestrogen receptor; PR = progesterone receptor.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
HE drafted the manuscript, participated in the design of the study and data collection, and performed the statistical analysis. PH carried out the immunohistochemistry. AT carried out the molecular genetic studies. KA participated in patient collection and did the genetic counselling of the patients and participated in drafting the manuscript. CB participated in the design of the study and drafting of the manuscript. HN participated in the design of the study, data collection, and drafting of the manuscript. All authors read and approved the final manuscript.
Acknowledgements
We wish to thank the families that participated in this study, Minna Merikivi for her help in sample collection, and the Finnish Cancer Registry for cancer diagnosis and identification numbers for archival material of the pathology laboratories. Grants: Academy of Finland, Clinical Research Fund of Helsinki University Central Hospital, Finnish Breast Cancer Group, Finnish-Norwegian Medical Foundation, The Finnish Cancer Society, and Sigrid Juselius Foundation.
Figures and Tables
Table 1 Features of breast tumours according to breast cancer patient's age (years) at diagnosis
BRCA1 tumours BRCA2 tumours Non-BRCA1/2 tumours
Age <50 Age ≥ 50 P < 0.05a Age <50 Age ≥ 50 P < 0.05a Age <50 Age ≥ 50 P < 0.05a
Histology
Ductal cancer 24 (70.6) 13 (76.5) 19 (54.3) 18 (75.0) 34 (58.6) 68 (72.3)
Lobular cancer 5 (14.7) 3 (17.6) 12 (34.3) 5 (20.8) 15 (25.9) 15 (16.0)
Medullary cancer 5 (14.7) 2 (3.4) 1 (1.1)
Other cancers 1 (5.9) 4 (11.4) 1 (4.2) 7 (12.1) 10 (10.6)
Grade
I 1 (3.1) 2 (11.8) 7 (23.3) 5 (22.7) 13 (25.0) 33 (36.7)
II 4 (12.5) 7 (41.2) 16 (53.3) 10 (45.5) 30 (57.7) 36 (40.0)
III 27 (84.4) 8 (47.1) 0.009 7 (23.3) 7 (31.8) 9 (17.3) 21 (23.3)
I to II 5 (16.6) 9 (53.0) 13 (76.7) 15 (68.2) 43 (82.7) 69 (76.7)
Immunohistochemistry
ER- 25 (83.3) 3 (25.0) 0.001 7 (20.6) 10 (52.6) 0.017 17 (29.3) 23 (25.6)
ER+ 5 (16.7) 9 (75.0) 27 (79.4) 9 (47.4) 41 (70.7) 67 (74.7)
PR- 28 (90.3) 9 (69.2) 12 (35.3) 16 (80.0) 0.001 18 (31.0) 49 (54.4)
PR+ 3 (9.7) 4 (30.8) 22 (64.7) 4 (20.0) 40 (69.0) 41 (45.6)
p53- 15 (50.0) 12 (92.3) 0.014 27 (81.8) 15 (83.3) 43 (74.1) 75 (81.5)
p53+ 15 (50.0) 1 (7.7) 6 (18.2) 3 (16.7) 15 (25.9) 17 (18.5)
ERBB2- 23 (76.7) 12 (92.3) 29 (83.3) 16 (84.2) 45 (81.8) 64 (83.1)
ERBB2+ 7 (23.3) 1 (7.7) 5 (14.7) 3 (15.8) 10 (18.2) 13 (16.9)
Values are no. (frequency %). aComparison by age group. There were no significant differences according to patients' age in the non-BRCA1/2 tumours (rightmost column). ER, oestrogen receptor; ERBB2, ERBB2 oncoprotein; PR, progesterone receptor.
==== Refs
Phillips K Andrulis IL Goodwin PJ Breast carcinomas arising in carriers of mutations in BRCA1 or BRCA2: Are they prognostically different? J Clin Oncol 1999 17 3653 3663 10550164
Lakhani SR van de Vijver MJ Jacquemier J Anderson TJ Osin PP McGuffog L Easton D Consortium BCL The pathology of familial breast cancer: Predictive value of immunohistochemical markers estrogen receptor, progesterone receptor, HER-2, and p53 in patients with mutations in BRCA1 and BRCA2 J Clin Oncol 2002 20 2310 2318 11981002 10.1200/JCO.2002.09.023
Malone KE Daling JR Ostrander EA BRCA1 and medullary breast cancer JAMA 1998 280 1227 9786365 10.1001/jama.280.14.1227
Sorlie T Tibshirani R Parker J Hastie T Marron JS Nobel A Deng S Johnsen H Pesich R Geisler S Repeated observation of breast tumour subtypes in independent gene expression data sets Proc Natl Acad Sci 2003 100 8418 8423 12829800 10.1073/pnas.0932692100
Foulkes W Stefansson IM Chappuis PO Begin LR Goffin JR Wong N Trudel M Akslen LA Germline BRCA1 mutations and basal epithelial phenotype in breast cancer J Natl Cancer Inst 2003 95 1482 1485 14519755
Noguchi S Kasugai T Miki Y Fukutomi T Emi M Nomizu T Clinicopathologic analysis of BRCA1- or BRCA2 associated hereditary breast carcinoma in Japanese Women Cancer 1999 85 2200 22205 10326698 10.1002/(SICI)1097-0142(19990515)85:10<2200::AID-CNCR14>3.0.CO;2-S
Vahteristo P Syrjäkoski K Eerola H Kainu T Holli K Blomqvist C Kallioniemi O Nevanlinna H Response: re: population-based study of BRCA1 and BRCA2 mutations in 1035 unselected Finnish breast cancer patients J Natl Cancer Inst 2001 93 152 154 11208888 10.1093/jnci/93.2.153
Eerola H Heikkilä P Tamminen A Aittomäki K Blomqvist C Nevanlinna H Histopathological features of breast tumours in BRCA1, BRCA2, and in mutation negative breast cancer families Breast Cancer Research 2004
Nixon AJ Neuberg D Hayes DF Gelman R Connolly JL Schnitt S Abner A Recht A Vicini F Harris JR Relationship of patient age to pathologic features of the tumor and prognosis for patients with stage I or II breast cancer J Clin Oncol 1994 12 888 894 8164038
Talley LI Grizzle WE Waterbror JW Brown D Weiss H Frost AR Hormonen receptors and proliferation in breast carcinomas of equivalent histologic grades in pre- and postmenopaussal women Int J Cancer 2002 98 118 127 11857395 10.1002/ijc.10171
Anderson WF Chu KC Chang S Sherman ME Comparison of age-spesific insidence rate patterns for different histopathologic types of breast carcinoma Cancer Epidemiol Biomarkers Prev 2004 13 1128 1135 15247123
Eerola H Blomqvist C Pukkala E Pyrhönen S Nevanlinna H Familial breast cancer in southern Finland: how prevalent are breast cancer families and can we trust the family history reported by patients? Eur J Cancer 2000 36 1143 1148 10854948 10.1016/S0959-8049(00)00093-9
Vehmanen P Friedman LS Eerola H Sarantaus L Pyrhönen S Ponder BAJ Muhonen T Nevanlinna H A low proportion of BRCA2 mutations in Finnish Breast cancer families Am J Hum Genet 1997 60 1050 1058 9150152
Vehmanen P Friedman LS Eerola H McClure M Ward B Sarantaus L Kainu T Syrjäkoski K Pyrhönen S Kallioniemi O Low proportion of BRCA1 and BRCA2 mutations in Finnish breast cancer families: evidence for additional susceptibility genes Hum Mol Genet 1997 6 2309 2315 9361038 10.1093/hmg/6.13.2309
Huusko P Pääkkönen K Launonen V Pöyhönen M Blanco G Kauppila A Puistola U Kiviniemi H Kujala M Leisti J Evidence of founder mutations in Finnish BRCA1 and BRCA2 families Am J Hum Genet 1998 62 1544 1548 9585608 10.1086/301880
Sarantaus L Huusko P Eerola H Launonen V Vehmanen P Rapakko K Gillanders E Syrjäkoski K Kainu T Vahteristo P Multiple founder effects and geographical clustering of BRCA1 and BRCA2 families in Finland Eur J Hum Genet 2000 8 757 763 11039575 10.1038/sj.ejhg.5200529
Vahteristo P Bartkowa J Eerola H Syrjäkoski K Ojala S Kilpivaara O Tamminen A Kononen J Aittomäki K Heikkilä P A CHK2 genetic variant contributing to a substantial fraction of familial breast cancer Am J Hum Genet 2002 71 432 438 12094328 10.1086/341943
Vaziri AJ Krumroy LM Elson P Budd GT Darlington G Myles J Tubbs RR Casey G Breast tumour immunophenotype of BRCA1-mutation carriers is influenced by age at diagnosis Clin Cancer Res 2001 7 1937 1945 11448907
Antoniou AC Pharoah P Narod S Risch H Eyfjord JE Hopper JL Loman N Olsson H Johansson OT Borg A Average risks of breast and ovarian cancer associated with mutations in BRCA1 or BRCA2 detected in case series unselected for family history: a combined analysis of 22 studies Am J Hum Genet 2003 72 1117 1130 12677558 10.1086/375033
Foulkes W Metcalfe K Sun P Hanna WM Lynch H Ghadirian P Tung N Olopade O Weber BL McLennan J Estrogen reseptor status in BRCA1- and BRCA2-related breast cancer: The influence of age, grade, and histological type Clin Cancer Res 2004 10 2029 2034 15041722
Kononen J Bubendorf L Kallioniemi A Bärlund M Schraml P Leighton S Torhorst J Mihatch MJ Sauter G Kallioniemi O Tissue microarrays for high-throughput molecular profiling of tumor specimens Nature Med 1998 4 844 847 9662379 10.1038/nm0798-844
Gillet CE Springall RJ Barnes DM Hanby AM Multiple tissue core arrays in histopathology research: a validation study J Pathol 2000 192 549 553 11113875 10.1002/1096-9896(2000)9999:9999<::AID-PATH721>3.0.CO;2-0
Shattuck-Eidens D Oliphant A McClure M McBride C Gupte J Rubano T Pruss D Tavtigian SV Teng DH Adey N BRCA1 sequence analysis in women at high risk for susceptibility mutations. Risk factor analysis and implications for genetic testing JAMA 1997 278 1242 1250 9333265 10.1001/jama.278.15.1242
Couch FJ DeShano ML Blackwood MA Calzone K Stopfer J Campeau L Ganguly A Rebbeck T Weber BL BRCA1 mutations in women attending clinics that evaluate the risk of breast cancer N Engl J Med 1997 336 1409 1415 9145677 10.1056/NEJM199705153362002
Chan-Claude J Dong J Schmidt S Shayeghi M Komitowski D Becher H Stratton MR Royer-Pokora B Using gene carrier probability to select high risk families for identifying germline mutations in breast cancer susceptibility genes J Med Genet 1998 35 116 121 9507390
Parmigiani G Berry D Aguilar O Determining carrier probabilities for breast cancer-susceptibility genes BRCA1 and BRCA2 Am J Hum Genet 1998 62 145 158 9443863 10.1086/301670
Vahteristo P Eerola H Tamminen A Blomqvist C Nevanlinna H A probability model for predicting BRCA1 and BRCA2 mutations in breast and breast-ovarian cancer families Brit J Cancer 2001 84 704 708 11237395 10.1054/bjoc.2000.1626
Gilpin CA Carson N Hunter AGW A preliminary validation of a family history assessment form to select women at risk for breast or ovarian cancer for referral to a genetic center Clin Genet 2000 58 299 308 11076055 10.1034/j.1399-0004.2000.580408.x
Lidereau R Eisinger F Champème M-H Noguès C Bièche I Birnbaum D Pallud C Jacquemier J Sobol H Major Improvement in the efficacy of BRCA1 mutation screening using morphoclinical features of breast cancer Cancer Res 2000 60 1206 1210 10728676
Eisinger F Noguès C Guinebretiere J Peyrat J Bardou V Noguchi T Vennin P Sauvan R Lidereau R Birnbaum D Novel indications for BRCA1 screening using individual clinical and morphological features Int J Cancer 1999 84 263 267 10371344 10.1002/(SICI)1097-0215(19990621)84:3<263::AID-IJC11>3.0.CO;2-G
| 15987451 | PMC1175056 | CC BY | 2021-01-04 16:54:50 | no | Breast Cancer Res. 2005 Apr 21; 7(4):R465-R469 | utf-8 | Breast Cancer Res | 2,005 | 10.1186/bcr1025 | oa_comm |
==== Front
Breast Cancer ResBreast Cancer Research1465-54111465-542XBioMed Central London bcr10261598744910.1186/bcr1026Research ArticleBioluminescent human breast cancer cell lines that permit rapid and sensitive in vivo detection of mammary tumors and multiple metastases in immune deficient mice Jenkins Darlene E [email protected] Yvette S [email protected] Yoko [email protected] Joan [email protected] Tony [email protected] Xenogen Corporation, Alameda, California, USA2 Currently employed by Chiron Corporation, Emeryville, California, USA2005 8 4 2005 7 4 R444 R454 23 11 2004 11 2 2005 9 3 2005 17 3 2005 Copyright © 2005 Jenkins et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Introduction
Our goal was to generate xenograft mouse models of human breast cancer based on luciferase-expressing MDA-MB-231 tumor cells that would provide rapid mammary tumor growth; produce metastasis to clinically relevant tissues such as lymph nodes, lung, and bone; and permit sensitive in vivo detection of both primary and secondary tumor sites by bioluminescent imaging.
Method
Two clonal cell sublines of human MDA-MB-231 cells that stably expressed firefly luciferase were isolated following transfection of the parental cells with luciferase cDNA. Each subline was passaged once or twice in vivo to enhance primary tumor growth and to increase metastasis. The resulting luciferase-expressing D3H1 and D3H2LN cells were analyzed for long-term bioluminescent stability, primary tumor growth, and distal metastasis to lymph nodes, lungs, bone and soft tissues by bioluminescent imaging. Cells were injected into the mammary fat pad of nude and nude-beige mice or were delivered systemically via intracardiac injection. Metastasis was also evaluated by ex vivo imaging and histologic analysis postmortem.
Results
The D3H1 and D3H2LN cell lines exhibited long-term stable luciferase expression for up to 4–6 months of accumulative tumor growth time in vivo. Bioluminescent imaging quantified primary mammary fat pad tumor development and detected early spontaneous lymph node metastasis in vivo. Increased frequency of spontaneous lymph node metastasis was observed with D3H2LN tumors as compared with D3H1 tumors. With postmortem ex vivo imaging, we detected additional lung micrometastasis in mice with D3H2LN mammary tumors. Subsequent histologic evaluation of tissue sections from lymph nodes and lung lobes confirmed spontaneous tumor metastasis at these sites. Following intracardiac injection of the MDA-MB-231-luc tumor cells, early metastasis to skeletal tissues, lymph nodes, brain and various visceral organs was detected. Weekly in vivo imaging data permitted longitudinal analysis of metastasis at multiple sites simultaneously. Ex vivo imaging data from sampled tissues verified both skeletal and multiple soft tissue tumor metastasis.
Conclusion
This study characterized two new bioluminescent MDA-MB-231-luc human breast carcinoma cell lines with enhanced tumor growth and widespread metastasis in mice. Their application to current xenograft models of breast cancer offers rapid and highly sensitive detection options for preclinical assessment of anticancer therapies in vivo.
==== Body
Introduction
Development of breast cancer mouse models that provide consistent primary mammary tumors and metastasis to clinically relevant tissues such as lymph nodes, lungs, and bone remain a challenge in the preclinical evaluation of therapies for human breast cancer. Current xenograft models of breast carcinoma involve murine or human breast cancer cell lines implanted into the mammary fat pad of mice or injected systemically by intravenous or intracardiac routes. Tumor cells injected into the mammary tissue yield reproducible tumors, but can require weeks to several months for primary tumor development and produce varied spontaneous metastasis depending on the cell line and mouse strain used in the study [1].
One common human breast cancer cell line used in xenograft animals models is MDA-MB-231. These cells originated from a human metastatic ductal breast carcinoma sample [2], are estrogen independent, and exhibit preferential growth in the mammary fat pad of immune compromised mice [3]. MDA-MB-231 cells develop primary tumors that produce spontaneous metastasis to lymph nodes and micrometastases to the lungs [4]. Detection of metastasis has relied primarily upon histological or PCR analysis of selected tissues at experimental end-points. Spontaneous metastasis to bone or soft organs from primary mammary tumors has not been reported.
Reproducible bone metastasis in breast cancer xenograft models has been achieved with intracardiac injection of MDA-MB-231 cells [5,6]. Passaging tumor cells harvested from the bone lesions several times in vivo has created MDA-MB-231 sublines with exclusive propensity for bone metastasis [7-10]. The bone metastases are typically identified in animals by radiographic or histological procedures. Recently, researchers have begun to apply luciferase-based imaging methods to detect widespread metastasis in mouse breast cancer models [10-13]. In studies using luciferase-expressing MDA-MB-231 tumor cell sublines specifically selected for skeletal metastasis, in vivo imaging was able to monitor experimental bone metastasis in mice to a level comparable to that of X-ray analysis [10,13]. Our goal was to develop a bioluminescent human breast cancer cell line that would offer a similar level of detection for both primary and metastatic tumors and would more fully mimic clinical breast cancer by metastasizing to multiple tissues, including lymph nodes, lungs, bone, and visceral organs.
This report describes bioluminescent xenograft mouse models based on more widely metastatic derivatives of MDA-MB-231 cells. These two luciferase-expressing cell lines, D3H1 and D3H2LN, were isolated for stable firefly luciferase expression in vitro and were passaged in mice to enhance their tumorigenic and metastatic properties. We evaluated the effect of long-term in vivo growth on the stability of cellular bioluminescence. In vivo and ex vivo imaging was used to monitor and compare the primary tumor growth rates and metastatic potential from mammary tumors and after intracardiac injection of the two MDA-MB-231 sublines over time.
Materials and methods
Tumor cell line
Human breast cancer cell line MDA-MB-231 was obtained from the American Type Culture Collection (Rockville, MA, USA). Cells were cultured in Minimum Essential Medium with Earl's Balanced Salts Solution MEM/EBSS medium supplemented with 10% fetal bovine serum, 1% nonessential amino acids, 1% L-glutamine, and 1% sodium pyruvate (all from Hyclone, Logan, UT, USA) at 37°C in a humidified atmosphere containing 5% carbon dioxide.
Co-transfection and selection of cell line
MDA-MB-231 cells were co-transfected with plasmids expressing the firefly luciferase gene (luc; pGL3-Red, Chris Contag, Stanford University, Stanford, CA, USA) and zeocin resistance gene (pSV40/Zeo; Invitrogen, Carlsbad, CA, USA) using lipofectamine (Invitrogen). Transfected cells were then selected for antibiotic resistance (zeocin; Invitrogen) and surviving colonies were screened for bioluminescence in complete media supplemented with 150 μg/ml D-luciferin (Biosynth International, Inc, Naperville, IL, USA) by in vitro imaging using the IVIS™ camera system (Xenogen, Alameda, CA, USA; see below). Bioluminescent, antibiotic resistant, single cell clones were amplified in culture and characterized for stable luminescence in vitro and tumorigenic potential in vivo. One clonal cell line, MDA-MB-231-luc-D3, was initially selected. This D3 clone was passaged through mice as a primary tumor in the mammary fat pad to produce the D3H1 cell line. A spontaneous lymph node metastasis from a D3H1 mammary fat pad tumor was also harvested and designated D3H2LN. These three bioluminescent derivatives of MDA-MB-231 cells were used for further in vitro or in vivo studies.
Bioluminescent imaging
Bioluminescent imaging was performed with a highly sensitive, cooled CCD camera mounted in a light-tight specimen box (IVIS™; Xenogen), using protocols similar to those described previously [14-17]. Imaging and quantification of signals were controlled by the acquisition and analysis software Living Image® (Xenogen). For in vitro imaging, bioluminescent cells were serially diluted from 4000 to 8 cells in complete media into black, clear bottomed, 96-well plates (Costar, Acton, MA, USA). D-luciferin (Biosynth International, Inc.) at 150 μg/ml in media was added to each well 5–10 min before imaging. Imaging time was 1 min/plate. For in vivo imaging, animals were given the substrate D-luciferin by intraperitoneal injection at 150 mg/kg in DPBS Dulbecco's Phosphate Buffered Saline (Invitrogen, Carlsbad, CA, USA), and anesthetized (1–3% isoflurane). Mice were then placed onto the warmed stage inside the light-tight camera box with continuous exposure to 1–2% isoflurane. Imaging times ranged from 1 s to 3 min, depending on the tumor model and time point. Generally, two to three mice were imaged at a time. The low levels of light emitted from the bioluminescent tumors or cells were detected by the IVIS™ camera system, integrated, digitized, and displayed. Regions of interest from displayed images were identified around the tumor sites and were quantified as total photon counts or photons/s using Living Image® software (Xenogen). Background bioluminescence in vivo was in the region of 1 × 104 photon counts or 1–2 × 105 photons/s. For ex vivo imaging, 150 mg/kg D-luciferin was injected into the mice just before necropsy. Tissues of interest were excised, placed into 24-well tissue culture plates with 300 μg/ml D-luciferin in DPBS, and imaged for 1–2 min. Tissues were subsequently fixed in 10% formalin (Sigma, St. Louis, MO, USA) and prepared for standard histopathology evaluation.
Mouse strain and animal care
Strict animal care procedures set forth by the Institutional Animal Care and Use Committee based on guidelines from the US National Institutes of Health Guide for the Care and Use of Laboratory Animals were followed for all experiments [18]. Mice used in these studies were either female athymic nude-nu mice aged 8–10 weeks (Harlan, Indianapolis, IN, USA) or female nude-beige (NIH-bg-nu-xidBR) mice aged 8–10 weeks (Charles River Laboratories, Wilmington, MA, USA).
Mammary fat pad spontaneous metastasis model
Female nude mice or female nude-beige mice aged 8–10 weeks were anesthetized by exposure to 1–3% isoflurane and injected with 50 μl of 2 × 106 MDA-MB-231 cells suspended in 50% DPBS/50% matrigel into the abdominal mammary fat pad. At 10–15 min after luciferin injection, mice were placed in the IVIS™ Imaging System and imaged from the ventral view. Tumor growth was monitored weekly by bioluminescent imaging and external caliper measurements (tumor size = [length × width × height] × 0.52) for 5–9 weeks. In some experiments, the lower portion of each animal was shielded before reimaging in order to minimize the bioluminescence from the primary tumor so that the signals from the metastatic regions could be observed in vivo. The front limbs were secured with tape to better expose the axillary/brachial lymph node areas. The imaging time ranged from 1 s to 1 min, depending on the size of the primary tumors, but was consistently 3 min for detection of metastases.
Intracardiac experimental metastasis model
Female nude mice (age 8–10 weeks) were anesthetized by intramuscular injection of 120 mg/kg ketamine hydrochloride with 6 mg/kg xylazine on the day of injections, and by exposure to 1–3% isoflurane on subsequent imaging days. On day 0, anesthetized animals were injected with 1 × 105 MDA-MB-231 cells suspended in 100 μl sterile DPBS into the left ventricle of the heart by nonsurgical means [19]. Anesthetized mice were placed in the IVIS™ Imaging System and imaged from both dorsal and ventral views approximately 10–15 min after intraperitoneal injection of D-luciferin. A successful intracardiac injection was indicated on day 0 by images showing systemic bioluminescence distributed throughout the animal. Only mice with evidence of a satisfactory injection continued in the experiment. Assessment of subsequent metastasis was monitored in vivo once a week by imaging for up to 5 weeks.
Histopathology
To confirm the presence of neoplastic cells, selected tissues were excised from the mice at necropsy and were preserved in 10% formalin solution (Sigma) immediately after ex vivo imaging. Tissues were prepared for histopathology (paraffin preparation, sectioning, and hematoxylin and eosin staining) and analyzed by subsequent microscopic evaluation by IDEXX Veterinary Services (West Sacramento, CA, USA).
Statistical analyses
The mean bioluminescence (photons/s), tumor volume, and corresponding standard errors of the mean were determined for each experiment. Regression plots were used to describe the relationship between bioluminescence and cell number and tumor volume; R2 values are reported to assess the quality of the regression model.
Results
Stable luciferase expression of MDA-MB-231-luc derivative cell lines
Parental MDA-MB-231 cells were transfected with a plasmid encoding firefly luciferase, and several stable clones were selected in vitro. A resulting bioluminescent cell line with reasonable bioluminescence in vitro (D3) was isolated and injected into the mammary fat pad of nude mice. Following 12 weeks of growth in vivo, tumor cells were harvested from a primary tumor and re-propagated in vitro to produce the MDA-MB-231 subclone line D3H1. These cells were injected once more into the mammary fat pad of mice to yield a second cell line, D3H2LN, harvested from a lymph node metastasis of a D3H1 mammary tumor.
To determine whether prolonged growth in vivo affected bioluminescence of the D3H1 or D3H2LN cells over time relative to the original D3 luciferase-expressing line, serial dilutions from cell cultures of each were imaged and compared in vitro (Fig. 1). The mean photon emissions ranged from approximately 100 to 208 photons/s per cell, and we found no decrease in bioluminescence as the MDA-MB-231 luciferase-expressing subclones were passaged through animals. The D3H2LN cells were in fact slightly brighter in vitro than the D3H1 or original D3 cells. The data indicate that the level of luciferase expression remained relatively stable for up to 12 weeks of tumor growth for the D3H1 cells, and after an additional 12 weeks of in vivo growth for the D3H2LN cells.
Enhanced tumorigenicity of D3H1 and D3H2LN cell lines in the mammary fat pad
The mammary tumor growth rates of the two luciferase-expressing derivatives were compared with the nonglowing parental MDA-MB-231 cells (Fig. 2a). Mean external caliper measurements of mammary tumors indicated that the D3H1 cells grew at a slightly faster rate than did the original parental cells. The D3H2LN cells exhibited a dramatic increase in tumor growth relative to both the D3H1 and the parental cells in vivo. By 5 weeks of growth in nude mice, the D3H2LN tumors had reached a volume more than two to seven times larger than the tumors of the other cell lines at 8 weeks after injection (Fig. 2a). Similar results were found with nude-beige mice (Fig. 2b), although tumor growth in general was slower for both D3H1 and D3H2LN cells in these experiments as compared with the rates in nude mice.
The growth of D3H1 and D3H2LN mammary tumors in nude or nude-beige mice was also monitored in vivo by bioluminescent imaging. The mean photons emitted from the tumors over time in each cohort were compared with the corresponding tumor volumes of the same animals. Bioluminescent data correlated with tumor volume, with overall R2 values of 0.95–0.98 (Fig. 2a, b). The only exception occurred at the final time points for the D3H2LN cells in nude mice, when the extremely fast growing tumors at weeks 4 and 5 continued to increase in volume but the bioluminescence leveled off or decreased. The developing necrotic cores in these tumors, confirmed by postmortem histological analysis (histology data not shown), presumably contributed to the diminished photon emission at the later time points.
Spontaneous metastasis from primary mammary fat pad tumors
Mice with D3H1 and D3H2LN mammary tumors were evaluated in vivo for spontaneous metastasis to the lungs or thoracic lymph nodes by shielding the bright primary tumors growing in the abdominal fat pad and reimaging the thorax of each animal. Representative images over time from a nude-beige mouse with D3H2LN tumors are presented in Fig. 3. A summary of the data from nude and nude-beige mice with D3H1 or D3H2LN tumors is presented in Table 1.
Similar to the D3H2LN tumor volume data shown in Fig. 2b, D3H2LN mammary tumor bioluminescence started to increase about 4 weeks after cell injection in nude-beige mice (Fig. 3, upper panels). Metastatic signals also appeared at that time in the right axillary region in 33% (2/6) of the mice, indicating spontaneous thoracic metastasis on the same side of the animal as the developing distal primary tumor. Subsequent thoracic images demonstrated an incremental increase in metastatic growth over time, and by week 6 all six mice had developed right axillary signals. Also at this time point, half of the mice exhibited additional metastatic signals in the left axilla, and by the end of the experiment all animals had developed bilateral thoracic metastases (Fig. 3).
The location and size of the in vivo bioluminescent signals suggested lymph node metastasis, and so the brachial and axillary lymph nodes of all mice were excised at necropsy, imaged ex vivo, and preserved for histological evaluation. Lung lobes were also included in the analysis as an alternative site for metastasis. Figure 3 (middle panel) shows the ex vivo images of the lymph nodes and lung lobes from a representative mouse. Brachial lymph nodes in all mice and lung lobes in 67% (4/6) of the cohort were bioluminescent ex vivo. The brighter ex vivo signal of the right brachial lymph nodes compared with the left lymph nodes correlated with the relatively higher intensity observed in vivo in the right thoracic area at the final time points. In the lung lobes, the ex vivo signal was at least 10-fold lower than the brachial lymph nodes, suggesting a much lower tumor burden in lungs than in lymph nodes. Subsequent histological evaluation confirmed extensive and pervasive tumor growth in the brachial lymph nodes and limited, sporadic micrometastasis in the lung lobes (Fig. 3, lower panels). Selected axillary lymph nodes, which were negative for bioluminescence ex vivo, were also found to be negative by histological examination (data not shown).
Tumor growth and metastasis of D3H2 tumors versus D3H2LN tumors
The mammary tumor take rate was lower and metastatic tumor growth was slower in studies with D3H1 cells than in those with D3H2LN cells. Table 1 shows the results for both nude and nude-beige mice. Metastatic spread to lymph nodes was also more limited in mice with D3H1 primary tumors, and spontaneous metastasis to both right and left brachial lymph nodes was rare or nonexistent. In contrast, all mice with D3H2LN mammary tumors developed right lymph node metastasis, and detection of bilateral spread was found in 100% of the nude beige mice. The detection of micrometastasis to lung lobes was observed in 38–67% of the mice with D3H2LN mammary tumors, whereas lung micrometastasis was nearly absent in mice that had D3H1 mammary tumors. Overall, nude-beige mice had higher frequencies of lung or lymph node metastasis than did nude mice, regardless of the tumor cell type used in the study.
Detection of metastases following intracardiac injection of MDA-MB-231-luc cells
We used a nonsurgical intracardiac injection of D3H1 or D3H2LN cells to compare experimental metastasis after systemic circulation of cells. Mice were imaged immediately after injection of cells and once a week thereafter. Ventral and dorsal images of a representative nude mouse injected with D3H2LN cells are shown in Fig. 4. Animals with successful intracardiac injection (8/12) on day zero showed low level, widespread bioluminescence minutes after the injection. Within 2 weeks, localized bioluminescent signals indicating metastasis began to appear at a few sites in each animal. By 3–4 weeks after injection, all eight mice exhibited clear indications of tumor growth at multiple sites in the head, thorax, abdomen, legs, or spine. None of these tumors was superficially obvious or palpable.
Ex vivo imaging was conducted with 17–20 different tissues excised from each mouse after the final in vivo imaging session at week 5. These tissues included untrimmed specimens of bones (jaw, skull, ribs, scapulae, clavicles, femur/tibia, vertebrae) along with isolated lungs, lymph nodes, salivary glands, and soft tissues such as brain, liver, pancreas, uterus, kidney and adrenals. In general, ex vivo images corroborated widespread metastasis in all mice injected with D3H2LN cells (Table 2). Representative ex vivo images of skeletal and brain metastasis are shown in Fig. 5, along with views of their corresponding histology sections. In all mice (8/8) metastasis to some type of bone-associated tissue was detected by ex vivo imaging, with frequencies greater than 50% for spine, skull, scapula, leg, or clavicle. A majority of animals injected with D3H2LN cells exhibited bioluminescent evidence of metastasis to the rib cage (8/8), brain (7/8), lungs (6/8), and thoracic lymph nodes (5/8). Less frequent metastasis (2–4/8) was observed ex vivo in the liver, pancreas, kidney/adrenals, uterus, heart, or salivary glands.
The panel of tissues examined by ex vivo imaging was also evaluated by histological analysis in two of the eight mice that presented with widespread metastasis from D3H2LN cells. Tumor lesions were sporadic and small in many of the tissues and some had to be re-sectioned and evaluated two or three times in order to locate what often was described as a few micrometastases per section. Nevertheless, metastasis was identified by microscopic examination in the majority of the tissues from at least one of the two mice sampled (Table 2, Fig. 5). In nearly all the lungs and rib sections, micrometastases were detected on the pleural surface, and not within the lung parenchyma or bone proper. It was generally not possible to detect micrometastasis to the larger soft organs by histology in the tissues from the two mice examined, although ex vivo imaging detected low levels of bioluminescence (data not shown).
Mice injected with D3H1 cells exhibited widespread metastasis in vivo after intracardiac injection but with lower frequency at each site compared with animals injected with the D3H2LN cells (Table 2). The first in vivo indication of metastasis in any of the D3H1 cohort occurred at week 5, which was 3 weeks later than in animals injected with the same number of D3H2LN cells (Fig. 5), and it took 10 weeks for metastasis to develop in a majority of animals injected with D3H1 cells versus 5 weeks for mice injected with D3H2LN cells. Although metastasis was confirmed by histology in most of the tissue samples from animals with D3H1 cells, isolated micrometastasis was frequently found associated with attendant muscle or fat rather than within the organs themselves (data not shown).
Discussion
Bioluminescent imaging has been used to detect primary cancer growth and metastasis in a growing number of animal models [14-20]. In xenograft tumor models, human cell lines expressing luciferase have permitted studies that yield real-time, noninvasive monitoring of tumor sites in the same cohort of animals over time. This study is the first application of in vivo bioluminescent imaging to monitor breast cancer tumor growth in animals and to detect spontaneous metastasis of tumor cells from the mammary fat pad to lymph nodes and lungs. The luciferase-expressing subclones of MDA-MB-231 characterized in this report also produced multiple metastases at high frequencies to clinically relevant tissues such as bone and brain following intracardiac injection of cells. With the ability to produce lymph node and lung metastasis from primary tumors and widespread metastasis after intracardiac injection, the D3H1 and D3H2LN subclones of MDA-MB-231 cells fully mimic the range of breast cancer development in humans.
Previous studies have noted lymph node and lung metastasis from parental MDA-MB-231 mammary tumors in mice but have largely relied on postmortem histology or PCR to document the presence of limited metastasis, especially to the lungs, from MDA-MB-231 primary tumors [4]. Our imaging experiments yielded similar findings but they did not require time consuming tissue processing and analysis. The low bioluminescence of the micrometastasis to the lung from mammary tumors was detected immediately in excised tissues by ex vivo analysis. The more pervasive lymph node metastasis was detected in both live animals over time and by ex vivo examination. Subsequent microscopic evaluation of the metastatic tissues initially identified as positive or negative by bioluminescent imaging was used to confirm tumor spread and allowed precise localization and sizing of lesions at the cellular level.
MDA-MB-231 luciferase-expressing cells were injected in the mammary fat pad of mice and passaged one or two times in vivo to generate a more robust xenograft model of human breast cancer with enhanced primary tumor growth and subsequent metastasis. In evaluating the photon emission of these cell lines, we found no decrease in bioluminescence between the D3H1 cells, D3H2LN cells, and the original unpassaged D3 subclone from the parental MDA-MB-231 cells, despite 12 weeks of tumor growth for D3H1 cells and 24 weeks of accumulated in vivo growth time for the D3H2LN line. Additionally, the D3H1 and D3H2LN cell lines maintained consistent levels of bioluminescence in vitro for up to 2 months or approximately eight to 10 passages in culture (data not shown). These data demonstrate remarkable long-term stability of luciferase expression after multiple rounds of in vivo growth and continuous periods of in vitro culturing.
The intracardiac injection model is used to generate widespread arterial dissemination of tumor cells by bypassing the lungs in order to seed cells to various organs. This method has been employed with luciferase-positive MDA-MB-231 derivative cell lines that were selected for enhanced in vivo metastasis to bone [10,13]. The metastatic spread of these cells after intracardiac injection was primarily to bone sites, with limited visceral metastasis to the adrenals or pancreas. Our D3H1 and D3H2LN sublines expand the utility of such models by producing consistent metastatic lesions to brain and various visceral organs as well as skeletal sites following intracardiac injection. In previous reports researchers compared bioluminescent imaging with radiometric measurements of bone metastasis, and found in vivo imaging to be the faster and more sensitive detection method [10]. Similarly, our intracardiac experiment compared imaging with histological examination and showed that, in many cases, tissues that had been found to be positive by in vivo or ex vivo imaging required multiple rounds of sectioning and histology to identify the micrometastasis.
In both nude and nude-beige mice, the D3H2LN tumors produced a faster rate of tumor growth in the mammary fat pad as well as enhanced spontaneous metastasis to the lymph nodes and lungs compared with the D3H1 cells (Fig. 2 and Table 1). The D3H2LN cells also exhibited higher frequencies of metastasis to bone, brain, lungs, and soft tissues after intracardiac injection (Table 2). Taken together, these findings suggest a more invasive phenotype in the D3H2LN cells and demonstrate an enhanced utility in animal metastasis studies. We did find that the rapid growth of D3H2LN cells in some mammary fat pad experiments created necrotic tumors that showed a decrease in bioluminescent signal in vivo at later time points, despite an incremental increase in tumor volume (Fig. 2a). A similar disparity between volume measurements and bioluminescence with larger tumors was reported in a study using subcutaneous tumors of human prostate cells expressing luciferase [11]. Here, the authors noted that tumor volume became static after treatment of animals, while a drop in bioluminescence of the same tumors indicated a tumoricidal effect on cells that was subsequently confirmed by histology.
Conclusion
The bioluminescent stability of the D3H1 and D3H2LN cell lines, coupled with their enhanced tumorigenicity and widespread metastatic potential, provides a sensitive in vivo model system for preclinical assessment of breast cancer growth, dissemination, and response to anticancer therapies.
Abbreviations
DPBS = Dulbecco's Phosphate Buffered Saline; PCR = polymerase chain reaction.
Competing interests
All authors are current or former employees of Xenogen Corporation. Xenogen manufactures the IVIS Imaging System™ and provides commercially available bioluminescent tumor cell lines as part of their Bioware™.
Authors' contributions
DEJ conceived the study, directed the experiments, and wrote the manuscript. YSH and YO carried out the in vivo experiments, organized the data, and contributed to drafts of the manuscript. JD carried out in vitro experiments and participated in drafting the final manuscript. TP supervised the project and participated in drafting the manuscript.
Acknowledgements
The authors wish to thank Shang-Fan Yu and Anne Pletcher (In Vivo Technologies, Inc) for intracardiac injections and training; Joycelyn Bishop for assistance with editing and formatting; Carrie Scatena and Anne Clermont for data analysis and figure formatting; Jay Petersen and John Hunter for comments and editing; and Bonnie Lemos and Shari Starr for expert animal care.
Figures and Tables
Figure 1 In vitro bioluminescence of MDA-MB-231 sublines D3, D3H1 and D3H2LN. Cells from each subline were serially diluted in duplicate wells from 4000 to 8 cells/well. Luciferin substrate was added to each well and the plate was imaged to obtain photons/s per cell. Wells with media (no cells) or cells alone (no luciferin) were included as controls. The range of bioluminescence for each subline represents data collected from two to three experiments.
Figure 2 Mammary tumor growth is enhanced in D3H1 and D3H2LN bioluminescent sublines of MDA-MB-231 cells. Cells were injected into the mammary fat pad of (a) nude or (b) nude-beige mice, and mean tumor volumes were determined by caliper measurements over time. The tumors from the bioluminescent sublines D3H1 and D3H2LN were also monitored using in vivo imaging, and the correlation between mean photons/s and mean tumor volume are indicated as R2 values.
Figure 3 D3H2LN mammary tumors grow rapidly and produce bilateral lymph node metastasis. Bioluminescent D3H2LN cells were injected into nude-beige mice, and tumor growth and metastasis were monitored over time in vivo. Data for a representative mouse are shown. Tumor take rate was high (6/8 mice) and lymph node metastasis to both the right and left thoracic areas was detected by week 9 in all mice with mammary tumors (upper panels). Ex vivo images of excised brachial lymph nodes and lung lobes are shown (middle panels). Histological analysis of the same tissues confirmed perfuse metastasis to the lymph nodes and isolated micrometastasis to the lung lobes (lower panels).
Figure 4 Multiple metastases detected in vivo following intracardiac injection of D3H2LN cells. Ventral (upper panels) and dorsal (lower panels) images taken over time from a representative nude mouse injected with D3H2LN cells are shown. Pseudocolor scale bars were consistent for all images of ventral or dorsal views in order to show relative changes at metastatic sites over time.
Figure 5 Ex vivo and histological data confirm metastases from intracardiac injection of D3H2LN cells. Representative data from selected tissues from two mice that were evaluated by both ex vivo imaging and subsequent histological analysis are shown.
Table 1 Mammary fat pad model tumor take and metastasis
Cell line Experimental parameters Detected metastases
Mouse strain n Take rate End-point (weeks) LN in vivo Lung ex vivo
D3H1 Nude 9, 10 40–78% 8–13 75–86% R, 0% L 0–16%
Nude-beige 6 66% 13 67% R, 16% L 0%
D3H2LN Nude 9, 8 89–100% 5 100% R, 0% L 38%
Nude-beige 8 75% 9 100% R, 100% L 67%
L, left; LN, lymph node; R, right.
Table 2 Ex vivo detection of metastasis after intracardiac injection (nude mice)
H2LN (n = 8) D3H1 (n = 5)
Experimental end-point Week 5 Week 10
Skeletal metastasis 8/8 4/5
Spine 88%a 60%a
Skull 75%a 40%a
Scapula 75% 40%a
Femur/tibia 63%a 0%a, b
Clavicle 50% 0%
Upper/lower jaw 13–23% ND 0–20%
Ribs and associated tissues 8/8a 4/5a
Brain 7/8a 2/5a
Lungs 6/8a 2/5a
Lymph nodes (axillary/brachial) 5/8a 2/5
Other soft tissues (heart, liver, pancreas, kidney, adrenals, uterus, salivary gland) 4/8 2/5
aMetastasis confirmed by histology in at least one animal sample. bThree out of five mice exhibited in vivo leg signal, and metastasis was confirmed by histology. ND, tissue not analyzed by histology.
==== Refs
Price JE Metastasis from human breast cancer cell lines Breast Cancer Res Treat 1996 39 93 102 8738609
Cailleau R Young R Olive M Reeves WJ Jr Breast tumor cell lines from pleural effusions J Natl Cancer Inst 1974 53 661 674 4412247
Price JE Polyzos A Zhang RD Daniels LM Tumorigenicity and metastasis of human breast carcinoma cell lines in nude mice Cancer Res 1990 50 717 721 2297709
Rose DP Connolly JM Liu XH Effects of linoleic acid on the growth and metastasis of two human breast cancer cell lines in nude mice and the invasive capacity of these cell lines in vitro Cancer Res 1994 54 6557 6562 7987856
Yoneda T Sasaki A Mundy GR Osteolytic bone metastasis in breast cancer Breast Cancer Res Treat 1994 32 73 84 7819589 10.1007/BF00666208
Yoneda T Arterial microvascularization and breast cancer colonization in bone Histol Histopathol 1997 12 1145 1149 9302573
Yoneda T Williams PJ Hiraga T Niewolna M Nishimura R A bone-seeking clone exhibits different biological properties from the MDA-MB-231 parental human breast cancer cells and a brain-seeking clone in vivo and in vitro J Bone Miner Res 2001 16 1486 1495 11499871
Michigami T Hiraga T Williams PJ Niewolna M Nishimura R Mundy GR Yoneda T The effect of the bisphosphonate ibandronate on breast cancer metastasis to visceral organs Breast Cancer Res Treat 2002 75 249 258 12353814 10.1023/A:1019905111666
Kang Y Siegel PM Shu W Drobnjak M Kakonen SM Cordon-Cardo C Guise TA Massague J A multigenic program mediating breast cancer metastasis to bone Cancer Cell 2003 3 537 549 12842083 10.1016/S1535-6108(03)00132-6
Wetterwald A van der Pluijm G Que I Sijmons B Buijs J Karperien M Lowik CW Gautschi E Thalmann GN Cecchini MG Optical imaging of cancer metastasis to bone marrow: a mouse model of minimal residual disease Am J Pathol 2002 160 1143 1153 11891210
Mendel DB Laird AD Xin X Louie SG Christensen JG Li G Schreck RE Abrams TJ Ngai TJ Lee LB In vivo antitumor activity of SU11248, a novel tyrosine kinase inhibitor targeting vascular endothelial growth factor and platelet-derived growth factor receptors: determination of a pharmacokinetic/pharmacodynamic relationship Clin Cancer Res 2003 9 327 337 12538485
Murray LJ Abrams TJ Long KR Ngai TJ Olson LM Hong W Keast PK Brassard JA O'Farrell AM Cherrington JM SU11248 inhibits tumor growth and CSF-1R-dependent osteolysis in an experimental breast cancer bone metastasis model Clin Exp Metastasis 2003 20 757 766 14713109 10.1023/B:CLIN.0000006873.65590.68
Minn AJ Kang Y Serganova I Gupta GP Giri DD Doubrovin M Ponomarev V Gerald WL Blasberg R Massague J Distinct organ-specific metastatic potential of individual breast cancer cells and primary tumors J Clin Invest 2005 115 44 55 15630443 10.1172/JCI200522320
Jenkins DE Oei Y Hornig YS Yu SF Dusich J Purchio T Contag PR Bioluminescent imaging (BLI) to improve and refine traditional murine models of tumor growth and metastasis Clin Exp Metastasis 2003 20 733 744 14713107 10.1023/B:CLIN.0000006815.49932.98
Jenkins DE Yu SF Hornig YS Purchio T Contag PR In vivo monitoring of tumor relapse and metastasis using bioluminescent PC-3M-luc-C6 cells in murine models of human prostate cancer Clin Exp Metastasis 2003 20 745 756 14713108 10.1023/B:CLIN.0000006817.25962.87
Scatena CD Hepner MA Oei YA Dusich JM Yu SF Purchio T Contag PR Jenkins DE Imaging of bioluminescent LNCaP-luc-M6 tumors: a new animal model for the study of metastatic human prostate cancer Prostate 2004 59 292 303 15042605 10.1002/pros.20003
Rehemtulla A Stegman LD Cardozo SJ Gupta S Hall DE Contag CH Ross BD Rapid and quantitative assessment of cancer treatment response using in vivo bioluminescence imaging Neoplasia 2000 2 491 495 11228541 10.1038/sj.neo.7900121
Arguello F Baggs RB Frantz CN A murine model of experimental metastasis to bone and bone marrow Cancer Res 1988 48 6876 6881 3180096
Contag CH Jenkins D Contag PR Negrin RS Use of reporter genes for optical measurements of neoplastic disease in vivo Neoplasia 2000 2 41 52 10933067 10.1038/sj.neo.7900079
Institute of Laboratory Animal Resources (ILAR) Guide for the Care and Use of Laboratory Animals: 1996 National Academy Press, Washington DC
| 15987449 | PMC1175057 | CC BY | 2021-01-04 16:04:33 | no | Breast Cancer Res. 2005 Apr 8; 7(4):R444-R454 | utf-8 | Breast Cancer Res | 2,005 | 10.1186/bcr1026 | oa_comm |
==== Front
Breast Cancer ResBreast Cancer Research1465-54111465-542XBioMed Central London bcr10271598745010.1186/bcr1027Research ArticleCYP17 5'-UTR MspA1 polymorphism and the risk of premenopausal breast cancer in a German population-based case–control study Verla-Tebit Emaculate [email protected] Shan [email protected] Jenny [email protected] Division of Clinical Epidemiology, German Cancer Research Center (Deutsches Krebsforschungszentrum), Heidelberg, Germany2 Molecular Biology Laboratory, Department of Obstetrics and Gynecology, University of Ulm, Ulm, Germany2005 12 4 2005 7 4 R455 R464 13 12 2004 31 1 2005 14 2 2005 18 3 2005 Copyright © 2005 Verla-Tebit et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Introduction
Studies on the association between the cytochrome P450c17α gene (CYP17) 5'-untranslated region MspA1 genetic polymorphism and breast cancer risk have yielded inconsistent results. Higher levels of estrogen have been reported among young nulliparous women with the A2 allele. Therefore we assessed the impact of CYP17 genotypes on the risk of premenopausal breast cancer, with emphasis on parity.
Methods
We used data from a population-based case–control study of women aged below 51 years conducted from 1992 to 1995 in Germany. Analyses were restricted to clearly premenopausal women with complete information on CYP17 and encompassed 527 case subjects and 904 controls, 99.5% of whom were of European descent. The MspA1 polymorphism was analyzed using PCR-RFLP (PCR–restriction fragment length polymorphism) assay.
Results
The frequencies of the variant allele among the cases and controls were 43% and 41%, respectively. Overall, CYP17 A1/A2 and A2/A2 genotypes compared with the A1/A1 genotype were not associated with breast cancer, with adjusted odds ratios (ORs) of 1.04 and 1.23, respectively. Among nulliparous women, however, breast cancer risk was elevated for the A1/A2 (OR = 1.31; 95% confidence interval (CI) 0.74 to 2.32) and the A2/A2 genotype (OR = 2.12; 95% CI 1.04 to 4.32) compared with the A1/A1 genotype, with a trend towards increasing risk associated with number of A2 alleles (P = 0.04). Otherwise, the CYP17 polymorphism was found neither to be an effect modifier of breast cancer risks nor to be associated with stage of disease.
Conclusion
Our results do not indicate a major influence of CYP17 MspA1 polymorphism on the risk of premenopausal breast cancer, but suggest that it may have an impact on breast cancer risk among nulliparous women. The finding, however, needs to be confirmed in further studies.
==== Body
Introduction
The risk of breast cancer is related to genetic, environmental, and lifestyle factors that influence the level of exposure to estrogens and other sex hormones [1]. Regarding genetic factors, high-penetrance cancer-susceptibility genes such as BRCA1 and BRCA2 are associated with some cases of familial breast cancer, though this association accounts for only about 5% of all breast cancer cases [2], while low-penetrance genes together with endogenous and lifestyle factors are likely to account for a higher proportion of breast cancer cases [3]. These low-penetrance genes include genes involved in the metabolism of sex hormones. One such gene is CYP17, which codes for the enzyme cytochrome P450c17α, responsible for catalyzing steroid 17α-hydroxylase and 17,20-lyase activities at key branch points in the estrogen biosynthesis pathway [4]. An increase or decrease in activity of these enzymes may alter the level of endogenous estrogen (estradiol), thereby influencing susceptibility to breast cancer [5,6]. One of the polymorphisms of the CYP17 gene is a thymidine substitution for cytosine (T to C) giving rise to an MspA1 restriction site at nucleotide 27 in the 5'-untranslated region (5'-UTR) promoter [7]. The MspA1 polymorphism has three genotypes: a homozygous wild type (A1/A1), a heterozygous variant (A1/A2), and the homozygous variant (A2/A2) [6]. The T-to-C substitution was initially hypothesized to create an Sp-1 binding site, which could lead to up-regulation of transcriptional activation of the variant allele and higher activity of the enzyme [8,9], but this was not observed in experimental studies [10-12].
The A2 allele has been associated with higher levels of estrogen than the wild-type allele [13,14]. In premenopausal women, the A2 allele is also associated with higher levels of dehydroepiandrosterone sulfate, and in postmenopausal women, with higher levels of total estradiol [15]. Estrogen is a known risk factor for breast cancer and many reproductive factors that are associated with risk, such as nulliparity, late age at first pregnancy, early menarche, and late menopause, are considered markers of lifetime exposure to estrogens [1]. It has been hypothesized that the presence of the variant A2 allele could contribute to an increase in breast cancer risk by virtue of higher estrogen levels. Several epidemiological studies have investigated the association between the CYP17 MspA1 polymorphism and breast cancer risk, with inconsistent findings [16-19]. A systematic review of studies on CYP17 and breast cancer risk by Dunning and colleagues concluded that risk was not significantly altered by CYP17 genetic polymorphisms [5]. In addition, a meta-analysis of 15 case–control studies published between 1994 and 2001 showed that the CYP17 MspA1 polymorphism may be a weak modifier of breast cancer risk but is not a significant independent risk factor [6]. However, the meta-analysis failed to include one study which, if included, could suggest a possible small increase in risk of breast cancer associated with the A2 allele [20]. Ambrosone and colleagues [21] also found that CYP17 acts as an effect modifier of breast cancer risk, especially with factors that influence endogenous estrogen levels. Similar to findings of other studies, they reported a protective effect of late age at menarche [8,13,22], and an increased risk with late age at first full-term pregnancy and use of oral contraceptives, among premenopausal women with the A1/A1 genotype.
It has been shown that premenopausal nulliparous women with the A2/A2 genotype have higher mean levels of serum estradiol than those with the A1/A1 genotype [14,23], implying that nulliparous women with the A2/A2 genotype may have a higher risk of breast cancer than nulliparous women with the A1/A1 genotype. Very few studies have looked at the risk of breast cancer associated with CYP17 in relation to parity [22,24]. This study therefore aimed to assess the risk of breast cancer and CYP17 genotype according to parity (nulliparous versus parous) and other risk factors for breast cancer among premenopausal women.
Materials and methods
Study design and study population
We used data from a population-based case–control study conducted from 1992 to 1995 in two regions of southern Germany (Freiburg and the Rhein–Neckar–Odenwald regions) [25]. The ethics committee of the University of Heidelberg reviewed the study protocol, and subjects who participated gave their informed consent. Subjects eligible for participation were German speaking, were under 51 years of age, resided within the study region, and had no previous history of breast cancer. Cases newly diagnosed with either in-situ or invasive breast cancer within the study period were identified through frequent monitoring of hospital admissions, surgery schedules, and pathology records from about 40 hospitals serving the study regions. There were also periodic checks of pathology institutions serving these hospitals, in order to identify any overlooked cases. There were 1,020 eligible case subjects, of whom 1,005 (98.5%) were alive when identified. Of these living subjects, 706 (70.2%) completed the study questionnaire and constituted the original population of case subjects, 152 (15.1%) refused to participate, 85 (8.5%) failed to respond, 51 (5.1%) were not contacted because of the physician's refusal to allow contact, and 11 (1.1%) were unable to participate because of health problems.
Controls were randomly selected from population registers of the study regions. An attempt was made to recruit two population controls per case, matched by age and study region. Subjects were not eligible as controls if they could not speak German, had moved out of the study region, had a previous history of primary breast cancer, were mentally handicapped, or had died. Of 2,257 eligible population controls, 1,381 (61.2%) participated (these were considered the original control population), 658 (29.1%) refused to participate, and 218 (9.7%) did not respond.
All subjects were asked to complete a self-administered questionnaire on demographic factors, anthropometric measures, menstrual, reproductive, and breast feeding histories, use of contraceptives and exogenous hormones, medical and screening histories, first- and second-degree family history of breast cancer, occupational exposures, smoking history, and alcohol consumption. Information on exposure for cases and controls was truncated at a reference date, which was the date of diagnosis for cases and the date of completion of the questionnaire for controls. All subjects were asked to provide a blood sample, which was used for genotyping. The median time between diagnosis and interview for cases was 2 months.
The study population was homogeneous, with 95.1% being of German origin and 88.6% of the non-Germans being of European descent. A total of 99.5% of the sudy subjects (Germans and non-Germans) were of European descent. The mean age was 41.6 years (± 5.8 standard deviations) for the cases and 41.7 years (± 5.7 standard deviations) for the controls.
Genotyping
Genomic DNA was extracted from the blood samples drawn into ethylenediaminetetraacetic acid tubes using Blood and Cell Culture DNA kits as described by the manufacturer (Qiagen GmbH, Hilden, Germany). The CYP17 5'-UTR MspAI polymorphism was analyzed using the previously described PCR-RFLP assay [7]. Briefly, a PCR fragment containing the base-pair change was amplified from genomic DNA by using primers (sense, 5'-CATTCGCACCTCTGGAGTC-3' ; antisense, 5'-GGCTCTTGGGGTACTTG-3'). After amplification, the PCR products were digested with the restriction endonuclease MspAI, subjected to electrophoresis through a 3% agarose gel, and visualized by staining the gel with ethidium bromide. Different genotypes could then be distinguished based on the size of the digested fragments.
Statistical analysis
We restricted our analysis to women clearly defined as premenopausal, since risk factors for breast cancer vary depending on menopausal status. Only women who still had menstrual cycles or reported natural amenorrhea for less than 6 months or more before the reference date (date of diagnosis for case subjects and date of completion of questionnaire for controls) were considered premenopausal. Women who reported natural amenorrhea 6 months before the reference date or bilateral oophorectomy were considered postmenopausal and hence not included in the analysis. Menopausal status for those who reported hysterectomy alone was classified as unknown and these subjects were also excluded, leaving 558 (79.0%) cases and 1,116 (80.8%) controls. Of these premenopausal subjects, 527 case subjects (94.4%) and 904 controls (81.0%) had complete information on CYP17 genotype, and therefore these 527 cases and 904 controls were included in the analysis.
The distribution of demographic characteristics and potential risk factors of breast cancer in this study population was compared with that of the original study population. Allele and genotype frequencies among cases and controls were calculated and deviation from Hardy–Weinberg equilibrium was examined using the χ2 test. The distributions of potential risk factors for breast cancer by CYP17 genotype in cases and controls were compared. Odds ratios (ORs) and 95% confidence intervals (95% CIs) were computed using multivariate conditional logistic regression analysis. Maximum-likelihood estimates were produced using the PHREG procedure in the SAS statistical software package (release 8.2; SAS Institute Inc, Cary, NC, USA). One-year age stratification was used to optimize age adjustment. Assessment of the association between CYP17 genotype and breast cancer was adjusted for potential confounders, including age at menarche, having ever used an oral contraceptive, total months of breastfeeding, family history of breast cancer in first-degree relatives, parity (defined as the number of full-term pregnancies resulting in either a live or a stillbirth), age at first full-term pregnancy (for parous women only), current body mass index, alcohol consumption, and level of education. Other variables such as study region, marital status, and smoking did not materially affect the risk estimates and were therefore not included in the model.
The effect of CYP17 genotype by parity (nulliparous and parous) and by other risk factors was examined to identify differential effects. We tested for trends in the logistic analyses by categorizing the exposure variables and treating the scored variables as continuous. All P values computed in the analyses were two-tailed. We tested for multiplicative interaction by computing the cross product of the variables and including it in the model alongside its separate components. We also assessed the distribution of the CYP17 genotype with respect to stage of the disease (local: stages 1 and 2; advanced: stages 3 and 4) and investigated trends with the Cochran–Armitage test.
Results
There were no major differences in the distribution of demographic characteristics and potential risk factors of breast cancer between the original study population and this study population (Table 1). Subjects in this study were about a year younger than those in the originally selected population, because we included only premenopausal women, who are generally younger than the excluded postmenopausal women. The frequency of the variant allele (A2) in the study population was similar for the cases and the controls: 43% and 41%, respectively. The genotype distribution was in agreement with that predicted under Hardy–Weinberg equilibrium, for both cases (P = 0.217) and controls (P = 0.380).
Table 2 shows the distribution of some risk factors for breast cancer among case subjects and controls by CYP17 genotype. χ2 tests for distribution revealed no significant differences among cases and controls in any of the genotype groups (A1/A1, A1/A2, A2/A2) with respect to age at menarche, age at first full-term pregnancy, parity, total months of breastfeeding, current body mass index, and educational level. Case subjects with the A1/A2 genotype were more likely than controls with this genotype to have used oral contraceptives. Compared with controls, case subjects with the A1/A1 genotype had significantly more family history of breast cancer and consumed more alcohol.
Overall, there was no significant association between the CYP17 genetic polymorphism and breast cancer risk (Table 3). The odds ratios for A1/A2 and A2/A2 genotypes were 1.04 and 1.23, respectively, in comparison with the A1/A1 genotype. Stratification by parity revealed a significantly increased risk in carriers of the A2/A2 genotype when compared with the A1/A1 genotype among nulliparous women (OR = 2.12). The risk associated with the A1/A2 genotype was elevated but did not reach statistical significance (OR = 1.31). There was a trend towards increasing risk with the number of variant alleles carried, which was statistically significant among nulliparous women (P = 0.04) but nonsignificant among parous women (P for interaction = 0.87) (Table 3). Table 4 shows the joint effects of CYP17 genotypes and parity. In comparison with parous women with the A1/A1 genotype, the greatest risk increase was seen for nulliparous women with the A2/A2 genotype (OR = 1.48).
In Table 5, we show results regarding some potential risk factors of breast cancer by CYP17 genotype, both overall and with further stratification by parity. Because of limited power for subgroup analyses, we combined A1/A2 and A2/A2 genotypes, as both groups are considered putative high-risk groups. Late age at menarche was not associated with breast cancer risk, irrespective of genotype and parity. The odds ratio for nulliparous women with the A1/A1 genotype who had ever used oral contraceptives was elevated compared with those who had never used oral contraceptives (OR = 2.60). The effect of breastfeeding among parous women (those who had ever breastfed versus those who had never breastfed) did not differ by genotype. However, among the parous women who had ever breastfed, breastfeeding for more than 12 months was associated with a risk reduction (OR = 0.56) when compared with 1 to 12 months of breastfeeding (Table 5). This effect did not differ according to CYP17 genotype (P for interaction = 0.48). Age at first full-term pregnancy was not associated with breast cancer risk in this study and the result was not altered when genotype status was taken into consideration.
We did not find any evidence of an association between CYP17 genotype and stage of breast cancer, with 65.5% of those with local disease and 68.1% of those with the advanced disease, respectively, being carriers of the A2 allele.
Discussion
The impact of CYP17 genetic polymorphism on the risk of breast cancer gained a lot of interest after Feigelson and colleagues first reported an increase in risk of advanced breast cancer for carriers of the A2 allele [8]. With a few exceptions [19,26], most subsequent studies did not find an overall increase in risk with the A2/A2 genotype [17,21,22,24,27-31]. The A2 allele has been shown to be associated with higher levels of estrogen in two studies [13,14], although a recent study did not observe higher levels of estrogen with the A2 allele [32]. Hong and colleagues recently reported that in premenopausal women, the A2 allele is associated with higher levels of dehydroepiandrosterone sulfate, which is a precursor to estrogens and androgens [15]. Despite the higher levels of this precursor in these subjects, there was no corresponding elevation of estradiol levels, which could be due partly to the difficulty of assessing representative estrogen levels based on a single measure [15,33].
We did not observe an overall increase in risk associated with the A2/A2 genotype compared with the A1/A1 genotype. However, we found an increased risk associated with the A2/A2 compared with the A1/A1 genotype among nulliparous women. This observation supports findings from a previous study indicating that nulliparous women with the A2/A2 genotype have higher mean levels of serum estradiol than those with the A1/A1 genotype [14]. Jernström and colleagues [34] also found out that the urinary ratio of the less potent 2-hydroxyestrone to the more potent 16α-hydroxyestrone is lower among nulliparous women with the A2/A2 genotype compared with the A1/A1 genotype. A low urinary ratio of 2-hydroxyestrone to 16α-hydroxyestrone has been reported to be associated with increased risk of breast cancer in premenopausal women [35]. This could explain our finding of increased risk of nulliparous premenopausal women with the A2/A2 genotype. Despite the biologically plausible mechanisms, this result should be interpreted with caution, because the numbers of subjects were small in this group and we did not observe any significant gene–parity effect modification. An increased risk has also been reported for nulliparous women and women who had had their first full-term pregnancy after the age of 30 years for carriers of at least one A2 allele among Chinese women in Singapore, though this was more pronounced in the postmenopausal group with that allele [24]. In the same line, a lower risk was observed for parous women with the A1/A1 genotype compared with nulliparous women with the same genotype [22]. Altogether, these findings suggest that the increased risk associated with A2 alleles may be more easily observable in nulliparous women because the greater lifetime exposure to circulating steroid hormones associated with this genotype is not altered by reproductive events.
We found no evidence of the previously reported effect modification for later age at menarche by the A1/A1 genotype [8,13,21,22]. We also did not find any association with respect to age at first full-term pregnancy. These associations were still absent after stratification by parity. A number of studies have also not been able to detect an association [17,29,30], including one of the largest case–control studies on this topic [31]. We observed an increased risk for oral contraceptive use compared with those who had never used oral contraceptives among nulliparous women with the A1/A1 genotype, although this might be a chance finding, especially as the number of subjects in this subgroup was small. Selection bias could also explain the findings if nulliparous case subjects with the A1/A1 genotype who participated in the study are more likely to use oral contraceptives. This is unlikely, however, as subjects are aware neither of their own genetype nor of the risk associated with it. Ambrosone and colleagues reported similar findings, though this was for all premenopausal women and not only in the nulliparous group [21]. They argued that oral contraceptive use might affect risk only in an environment of lower estrogens, seen in carriers of the A1/A1 genotype. Among parous women who had ever breastfed, we found a risk reduction for greater than 12 months of breastfeeding compared with 1 to 12 months of breastfeeding, but this effect did not differ with CYP17 genotype.
We also assessed the impact of CYP17 on stage of breast cancer, and, as in other studies [22,30,31], we were not able to confirm the increased risk for advanced breast cancer reported previously [8,26]. The association with stage was found in studies that included subjects having different racial and ethnic backgrounds. However, a possible bias due to population stratification can be excluded, as this effect was observed across all the ethnic groups [26].
We found fewer A2/A2 carriers in cases with a positive family history of breast cancer in first-degree relatives than in all controls, irrespective of family history of breast cancer (12.3% versus 17.4%; P = 0.21). This contradicts the findings of Spurdle and colleagues [27], who observed more of the A2/A2 genotype among cases with a positive family history of breast cancer in first- or second-degree relatives than in all controls. Their study subjects were below the age of 40, whereas most of our study subjects (70%) were aged 40 or above. In addition, they reported a deviation from Hardy–Weinberg equilibrium among cases with a positive family history of breast cancer, whereas we did not observe a deviation in this group (χ2 = 0.25; P = 0.62). Jernström and colleagues in their study on nulliparous young women also reported that carriers of the A2/A2 genotype had less family history of breast cancer in first- or second-degree relatives than carriers of the A1/A2 and A1/A1 genotypes combined [34].
Conclusion
Our results do not suggest a major influence of CYP17 genetic polymorphism on the risk of premenopausal breast cancer generally, but they do suggest an increase in risk for nulliparous women with the A2/A2 genotype. The resolution of the question regarding the role of the CYP17 genotype in breast cancer risk may require a better understanding of the functional variants discussed. A more comprehensive haplotype analysis would help to clarify whether the CYP17 variant allele itself is causal or is in linkage disequilibrium with some other variant that has a causal relation with breast cancer.
Abbreviations
CI = confidence interval; OR = odds ratio; PCR = polymerase chain reaction; RFLP = restriction fragment length polymorphism; UTR = untranslated region.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
VTE performed the statistical analysis and drafted the manuscript. WGS carried out the molecular genetic studies and participated in the preparation of the manuscript. CCJ conceived the study, participated in its design and coordination, and contributed to the statistical analysis and the preparation of the manuscript. All authors read and approved the final manuscript.
Acknowledgements
We are grateful to the women who participated in this research project and the oncologists and gynecologists in the 40 clinics of the study regions for allowing us to contact their patients; and to Tanja Koehler for intensive technical assistance on genotyping and Ursula Eilber for competent data management. This work was supported by the Medical Faculty of the University of Ulm (P.589 and P.685), the Deutsche Krebshilfe e.V., and the Graduiertenkolleg 793 of the University of Heidelberg.
Figures and Tables
Table 1 Comparison of demographic characteristics and potential risk factors for breast cancer in two study populations
Population statistics and characteristics Present study population Original study populationa
Cases (n = 527) Controls (n = 904) Cases (n = 706) Controls (n = 1,381)
Mean age (years) at diagnosis or recruitment 41.6 41.7 42.5 42.6
Mean BMI (kg/m2) 24.0 23.9 24.1 24.2
Mean age (years) at menarche 13.1 13.1 13.1 13.1
Mean age (years) at first full-term pregnancyb 24.4 24.9 24.2 24.3
Population characteristics (%)
Study region
Rhein–Neckar–Odenwald 71.7 70.7 70.0 69.3
Freiburg 28.3 29.3 30.0 30.7
Marital statusc
Single 9.3 9.8 7.8 8.4
Married/with partner 81.0 80.0 81.2 78.3
Widowed, divorced or separated 9.7 10.2 10.9 13.2
Nationality
German 91.8 97.0 91.2 95.8
Non-German 8.2 3.0 8.8 4.2
Educational level
Low 13.5 11.7 14.7 14.2
Intermediate 63.7 60.7 63.3 60.3
High 22.8 27.6 22.0 25.5
Parity
0 23.5 22.0 21.7 20.8
1 27.7 24.1 29.0 24.5
2 39.5 37.5 38.5 38.0
3+ 9.3 16.4 10.8 16.7
Oral contraceptive used
No 16.1 19.7 18.1 19.8
Yes 83.9 80.1 81.9 80.0
Breastfeedingb
Never 28.8 26.5 31.3 29.2
Ever 71.2 73.5 68.7 70.8
First-degree family history of breast cancer
No 87.7 94.7 87.7 94.9
Yes 12.3 5.3 12.3 5.1
Daily average alcohol intake
0 g 20.7 16.4 21.7 17.3
1–18 g 65.6 74.5 63.8 74.1
>18 g 13.7 9.1 14.5 8.7
aAll who completed questionnaires. bLimited to parous women. cTwo controls from the original study population had unknown marital status. dThree controls from the original study and two controls from the present study have missing data. BMI, body mass index.
Table 2 Distribution of breast cancer risk factors according to CYP17 genotype for premenopausal women in Germany
Patient characteristics CYP17 genotype
A1/A1 (n = 503) A1/A2 (n = 668) A2/A2 (n = 260)
Cases, no. (%) Controls no. (%) Cases, no. (%) Controls no. (%) Cases, no. (%) Controls no. (%)
Age at menarche (years)a
<13 68 (37.8) 123 (38.2) 84 (34.4) 149 (35.1) 34 (33.0) 51 (32.5)
≥13 112 (62.2) 199 (61.6) 158 (64.8) 275 (64.9) 68 (66.0) 106 (67.5)
Age at first full-term pregnancy (years)b
<25 69 (49.6) 125 (49.4) 106 (56.1) 172 (52.1) 40 (53.3) 55 (45.1)
≥25 70 (50.4)) 128(50.6) 83 (43.9) 158 (47.9) 35 (46.7) 67 (54.9)
Parity
0 birth 41 (22.8) 70 (21.7) 55 (22.5) 94 (22.2) 28 (27.2) 35 (22.3)
1–2 births 123 (68.3) 201 (62.2) 168 (68.9) 260 (61.3) 63 (61.2) 96 (61.1)
>2 births 16 (8.9) 52 (16.1) 21 (8.6) 70 (16.5) 12 (11.6) 26 (16.6)
Total months of breastfeedingb
0 43 (30.9) 65 (25.7) 53 (28.0) 94 (28.5) 20 (26.7) 28 (22.9)
1–12 86 (61.9) 156 (61.7) 117 (61.9) 186 (56.4) 50 (66.8) 74 (60.7)
>12 10 (7.2) 32 (12.6) 19 (10.1) 50 (15.2) 5 (6.7) 20 (16.4)
Ever use oral contraceptivec
No 30 (17.1) 62 (19.4) 31 (13.0) 84 (20.0)* 24 (23.3) 32 (20.5)
Yes 146 (82.9) 258 (80.6) 207 (87.0) 336 (80.0) 79 (76.7) 124 (79.5)
Family history of breast cancer in first-degree relative
No 150 (83.3) 306 (94.7)* 217 (88.9) 407 (96.0) 95 (92.2) 143 (91.1)
Yes 30 (16.7) 17 (5.3) 27 (11.1) 17 (4.0) 8 (7.8) 14 (8.9)
Current BMI (k/m2)d
<20.0 26 (14.4) 35 (10.8) 39 (16.0) 47 (11.1) 10 (9.7) 25 (15.9)
20.0–25.0 95 (52.8) 191 (59.1) 130 (53.3) 245 (57.8) 64 (62.1)) 87 (55.4)
25.1–30.0 38 (21.1) 71 (22.0) 51 (20.9) 95 (22.4) 20 (19.4) 30 (19.1)
>30.0 20 (11.1) 25 (7.7) 24 (9.8) 37 (8.7) 9 (8.7) 15 (9.6)
Educational level
Low 24 (13.4) 39 (12.1) 31 (12.7) 50 (11.8) 16 (15.5) 17 (10.8)
Middle 114 (63.3) 200 (61.9) 159 (65.2) 262 (61.8) 63 (61.2) 87 (55.4)
High 42 (23.3) 84 (26.0) 54 (22.1) 112 (26.4) 24 (23.3) 53 (33.8)
Alcohol consumption (g/day)
0 39 (21.7) 52 (16.1)* 52 (21.3) 72 (17.0) 18 (17.5) 24 (15.3)
1–18 118 (65.5) 246 (76.2) 158 (64.8) 309 (72.9) 70 (68.0) 119 (75.8)
>18 23 (12.8) 25 (7.7) 34 (13.9) 43 (10.1) 15 (14.5) 14 (8.8)
aData were missing for one control with A1/A1 genotype, two cases with A1/A2 genotype, and one case with A2/A2 genotype. bLimited to parous women. cData were missing for four cases and three controls with A1/A1 genotype; six cases and four controls with A1/A2 genotype, and one control with A2/A2 genotype. dData were missing for one case and one control with A1/A1 genotype. * χ2 test, P < 0.05. BMI, body mass index.
Table 3 Association between CYP17 genetic polymorphisms and premenopausal breast cancer risk by parity in Germany
CYP17 genotype Cases, no. (%) (n = 527) Controls, no. (%) (n = 904) Age-adjusted OR (95% CI)a Multivariate-adjused OR (95% CI)b
All subjectsc
A1/A1 180 (34.2) 323 (35.7) 1.00 1.00
A1/A2 244 (46.2) 424 (46.9) 1.02 (0.80–1.31) 1.04 (0.81–1.34)
A2/A2 103 (19.6) 157 (17.4) 1.18 (0.87–1.61) 1.23 (0.89–1.69)
Trend test P 0.34 0.24
Parous womend
A1/A1 139 (34.5) 253 (35.9) 1.00 1.00
A1/A2 189 (46.9) 330 (46.8) 1.02 (0.78–1.35) 1.04 (0.78–1.37)
A2/A2 75 (18.6) 122 (17.3) 1.13 (0.79–1.61) 1.14 (0.79–1.65)
Trend test P 0.54 0.50
Nulliparous women
A1/A1 41 (33.0) 70 (35.2) 1.00 1.00
A1/A2 55 (44.4) 94 (47.2) 1.23 (0.72–2.07) 1.31 (0.74–2.32)
A2/A2 28 (22.6) 35 (17.6) 1.74 (0.90–3.38) 2.12 (1.04–4.32)
Trend test P 0.11 0.04
aAge-adjusted odds ratio (OR) and 95% confidence interval (CI). bMultivariate-adjusted OR and 95% CI. Adjusted for age at menarche in years (≤ 12, 13 to 14, 15+); total months of breastfeeding (0, 1 to 12, >12); alcohol consumption (0, 1–18, >18 g/day); current body mass index as continuous variable; educational level (low, middle, high); and family history of breast cancer in first-degree relative (no, yes); oral contraceptive use (no, yes). cAlso adjusted for parity (0, 1 to 2 births, 3+ births). dAlso adjusted for age at first full-term pregnancy (<25, ≥ 25 years). P for interaction = 0.87.
Table 4 Joint effects of CYP17 polymorphisms and parity on risk of premenopausal breast cancer in Germany
CYP17 genotype Cases, no. (%) Controls, no. (%) OR (95% CI)a OR (95% CI)b
Parous women
A1/A1 139 (26.4) 253 (28.0) 1.00 1.00
A1/A2 189 (35.9) 330 (36.5) 1.03 (0.78–1.36) 1.05 (0.79–1.39)
A2/A2 75 (14.2) 122 (13.5) 1.12 (0.79–1.60) 1.16 (0.80–1.66)
Nulliparous women
A1/A1 41 (7.8) 70 (7.7) 1.04 (0.67–1.61) 1.01 (0.62–1.67)
A1/A2 55 (10.4) 94 (10.4) 1.04 (0.70–1.56) 1.05 (0.66–1.67)
A2/A2 28 (5.3) 35 (3.9) 1.43 (0.83–2.46) 1.48 (0.82–2.70)
aAge-adjusted odds ratio (OR) and 95% confidence interval (CI). bMultivariate adjusted OR and 95% CI. Adjusted for age at menarche in years (≤ 12, 13 to 14, 15+); oral contraceptive use (no, yes); total months of breastfeeding (0, 1 to 12, >12); alcohol consumption (0, 1 to 18, >18 g/day); current body mass index as continuous variable; educational level (low, middle, high).
Table 5 Association between some potential risk factors of breast cancer by CYP17 genotype among premenopausal women in Germany
Factor All data A1/A1 A1/A2 and A2/A2
Cases (n = 527) Controls (n = 904) ORa (95% CI) Cases (n = 180) Controls (n = 323) ORa (95% CI) Cases (n = 347) Controls (n = 581) ORa (95% CI)
Age at menarcheb
<13 years 186 323 1.00 68 123 1.00 118 200 1.00
≥13 years 338 580 1.00 (0.80–1.27) 112 199 1.01 (0.67–1.51) 226 381 1.00 (0.75–1.34)
Age at menarche, nulliparous women
<13 years 43 64 1.00 17 24 1.00 26 40 1.00
≥13 years 79 135 0.79 (0.48–1.31) 24 46 0.98 (0.38–2.55) 55 89 0.91 (0.47–1.78)
Age at menarche, parous women
<13 years 143 259 1.00 51 99 1.00 92 160 1.00
≥13 years 259 445 1.11 (0.85–1.45) 88 153 1.08 (0.68–1.72) 171 292 1.09 (0.78–1.53)
Pill usec
Never 85 178 1.00 30 62 1.00 55 116 1.00
Ever 442 724 1.23 (0.91–1.66) 150 260 1.22 (0.72–2.09) 292 464 1.24 (0.86–1.79)
Pill use, nulliparous women
Never 25 48 1.00 10 23 1.00 15 25 1.00
Ever 99 150 1.64 (0.87–3.08) 31 46 2.60 (0.81–8.43) 68 104 1.35 (0.59–3.10)
Pill use, parous women
Never 60 130 1.00 20 39 1.00 40 91 1.00
Ever 343 574 1.26 (0.89–1.80) 119 214 1.10 (0.58–2.09) 224 360 1.35 (0.88–2.05)
Age at first full-term preganancyd
<25 years 215 352 1.00 69 125 1.00 146 227 1.00
≥25 years 188 353 0.86 (0.66–1.14) 70 128 0.98 (0.60–1.59) 118 225 0.79 (0.56–1.11)
Breastfeeding statusd
Never 116 187 1.00 43 65 1.00 73 122 1.00
Ever 287 518 0.90 (0.68–1.20) 96 188 0.70 (0.43–1.16) 191 330 1.02 (0.71–1.46)
Total months of breastfeeding among parous ever breastfed women
1–12 253 416 1.00 86 156 1.00 167 260 1.00
>12 34 102 0.56 (0.36–0.87) 10 32 0.69 (0.30–1.60) 24 70 0.53 (0.31–0.90)
aMultivariate adjusted odds ratio (OR) and 95% confidence interval (CI). Adjusted for alcohol consumption (0, 1 to 18, >18 g/day); current body mass index as continuous variable; educational level (low, middle, high); family history of breast cancer in first-degree relative (no, yes); parity (0, 1 to 2 births, 3+ births), and other variables in the table where appropriate. bThree cases and one control had missing data. cTwo controls had missing data. dLimited to parous women.
==== Refs
Feigelson HS Henderson BE Estrogens and breast cancer Carcinogenesis 1996 17 2279 2284 8968038
Bennett IC Gattas M Teh BT The genetic basis of breast cancer and its clinical implications Aust N Z J Surg 1999 69 95 105 10030809 10.1046/j.1440-1622.1999.01515.x
Johnson-Thompson MC Guthrie J Ongoing research to identify environmental risk factors in breast carcinoma Cancer 2000 88 1224 1229 10705359 10.1002/(SICI)1097-0142(20000301)88:5+<1224::AID-CNCR8>3.0.CO;2-H
Picado-Leonard J Miller WL Cloning and sequence of the human gene for P450c17 (steroid 17α-hydroxylase/17,20 lyase): similarity with the gene for P450c21 DNA 1987 6 439 448 3500022
Dunning AM Healey CS Pharoah PDP Teare MD Ponder BAJ Easton DF A systematic review of genetic polymorphisms and breast cancer risk Cancer Epidemiol Biomarkers Prev 1999 8 843 854 10548311
Ye Z Parry JM The CYP17 MspA1 polymorphism and breast cancer risk: a meta-analysis Mutagenesis 2002 17 119 126 11880540 10.1093/mutage/17.2.119
Carey AH Waterworth D Patel K White D Little J Novelli P Franks S Williamson R Polycystic ovaries and premature male pattern baldness are associated with one allele of the steroid metabolism gene CYP17 Hum Mol Genet 1994 3 1873 1876 7849715
Feigelson HS Coetzee GA Kolonel LN Ross RK Henderson BE A polymorphism in the CYP17 gene increases the risk of breast cancer Cancer Res 1997 57 1063 1065 9067272
Kristensen VN Haraldsen EL Anderson KB Lonning PE Erikstein B Karesen R Gabrielsen OS Borresen-Dale A-L CYP17 and breast cancer risk: The polymorphism in the 5' flanking area of the gene does not influence binding to Sp-1 Cancer Res 1999 59 2825 2828 10383140
Kadonaga JT Carner KR Masiarz FR Tjian R Isolation of cDNA encoding transcription factor Sp 1 and functional analysis of the DNA binding domain Cell 1987 51 1079 1090 3319186 10.1016/0092-8674(87)90594-0
Lin CJ Martens JW Miller WL NF-1C, Sp1, and Sp3 are essential for transcription of the human gene for P450c17 (steroid 17 alpha-hydroxylase/17,20 lyase) in human adrenal NCIH295A cells Mol Endocrinol 2001 15 1277 1293 11463853 10.1210/me.15.8.1277
Miyoshi Y Ando A Ooka M Shiba E Taguchi T Tamaki Y Noguchi S Association of CYP17 genetic polymorphism with intra-tumoral estradiol concentrations but not with CYP17 messenger RNA levels in breast cancer tissue Cancer Lett 2003 195 81 86 12767515 10.1016/S0304-3835(02)00211-2
Haiman CA Hankinson SE Spiegelman D Colditz GA Willett WC Speizer FE Kelsey KT Hunter DJ The relationship between a polymorphism in CYP17 with plasma hormone levels and breast cancer Cancer Res 1999 59 1015 1020 10070957
Feigelson HS Shames LS Pike MC Coetzee GA Stanczyk FZ Henderson BE Cytochrome P450c17α gene (CYP17) polymorphism is associated with serum estrogen and progesterone concentrations Cancer Res 1998 58 585 587 9485002
Hong C-C Thompson HJ Jiang C Hammond GL Tritchler D Yaffe M Boyd NF Association between the T27C polymorphism in the cytochrome P450 c17α (CYP17) gene and risk factors for breast cancer Breast Cancer Res Treat 2004 88 217 230 15609124 10.1007/s10549-004-0780-7
Miyoshi Y Iwao K Ikeda N Egawa C Noguchi S Genetic polymorphism in CYP17 and breast cancer risk in Japanese women Eur J Cancer 2000 36 2375 2379 11094312 10.1016/S0959-8049(00)00334-8
Hamajima N Iwata H Obata Y Matsuo K Mizutani M Iwase T Miura S Okuma K Ohashi K Tajima K No association of the 5' promoter region polymorphism of CYP17 with breast cancer risk in Japan Jpn J Cancer Res 2000 91 880 885 11011114
Huang C-S Chern H-D Chang K-J Cheng C-W Hsu S-M Shen C-Y Breast cancer risk associated with genotype polymorhism of the estrogen-metabolizing genes CYP17, CYP1A1 and COMT: A multigenic study on cancer susceptibility Cancer Res 1999 59 4870 4875 10519398
Bergman-Jungestrom M Gentile M Lundin A-C Wingren S Association between CYP17 gege polymorphism and risk of breast cancer in young women Int J Cancer 1999 84 350 353 10404084 10.1002/(SICI)1097-0215(19990820)84:4<350::AID-IJC3>3.0.CO;2-L
Feigelson HS Mckean-Cowdin R Henderson BE Concerning the CYP17 MspA1 polymorphism and breast cancer risk: a meta-analysis [letter to the editor] Mutagenesis 2002 17 445 446 12202634 10.1093/mutage/17.5.445
Ambrosone CB Moysich KB Furberg H Freudenheim JL Bowman ED Ahmed S Graham S Vena JE Shields PG CYP17 genetic polymorphism, breast cancer and breast cancer risk factors Breast Cancer Res 2003 5 R45 R51 12631398 10.1186/bcr570
Mitrunen K Jourenkova N Kataja V Eskelinen M Kosma VM Benhamou S Vainio H Uusitupa M Hirvonen A Steroid metabolism gene CYP17 polymorphism and the development of breast cancer Cancer Epidemiol Biomarkers Prev 2000 9 1343 1348 11142420
Thompson PA Ambrosone C Molecular epidemiology of genetic polymorhisms in estrogen metabolizing enzymes in human breast cancer J Natl Cancer Inst Monogr 2000 27 125 134 10963624
Wu AH Seow A Arakawa K van den Berg D Lee H-P Yu MC HSD17B1 and CYP17 polymorphism and breast cancer risk among Chinese women in Singapore Int J Cancer 2003 104 450 457 12584742 10.1002/ijc.10957
Chang-Claude J Eby N Kiechle M Bastert G Becher H Breastfeeding and breast cancer risk by age 50 among women in Germany Cancer Causes Control 2000 11 687 695 11065005 10.1023/A:1008907901087
Feigelson HS McKean-Cowdin R Coetzee GA Stam DO Kolonel LN Henderson BE Building a multigenic model of breast cancer susceptibility: CYP17 and HSD17B1 are two important candidates Cancer Res 2001 61 785 789 11212283
Spurdle AB Hopper JL Dite GS Chen X Cui J McCredie MRE Giles GG Southey MC Venter DJ Easton DF CYP17 promoter polymorphism and breast cancer in Australian women under age forty years J Natl Cancer Inst 2000 92 1674 1681 11036113 10.1093/jnci/92.20.1674
Gudmundsdottir K Thorlacius S Jonasson JG Sigfusson BF Tryggvadottir L Eyfjord JE CYP17 promoter polymorphism and breast cancer risk in males and females in relation to BRCA2 status Br J Cancer 2003 88 933 936 12644832 10.1038/sj.bjc.6600839
Helzlsouer KJ Huang H-Y Strickland PT Hoffman S Alberg AJ Comstock GW Bell DA Association between CYP17 polymorphisms and the development of breast cancer Cancer Epidemiol Biomarkers Prev 1998 7 945 949 9796641
Weston A Pan C-f Bleiweiss IJ Ksieski HB Roy N Maloney N Wolff MS CYP17 Genotype and breast cancer risk Cancer Epidemiol Biomarkers Prev 1998 7 941 944 9796640
Dunning AM Healey CS Pharoah PD Foster Na Lipscombe JM Easton DF Day NE Ponder BA No association between a polymorphism in the steroid metabolism gene CYP17 and the risk of breast cancer Br J Cancer 1998 77 2045 2047 9667690
García-Closas M Herbstman J Schiffman M Glass A Dorgan JF Relationship between serum hormone concentrations, reproductive history, alcohol consumption and genetic polymorphisms in pre-menopausal women Int J Cancer 2002 102 172 178 12385014 10.1002/ijc.10651
Michaud DS Mansen JE Spiegelman D Barbieri RL Sepkovic D Bradlow HL Hankinson SE Reproducibility of plasma and urinary sex hormone levels in premenopausal women over a one-year period Cancer Epidemiol Biomarkers Prev 1999 8 1059 1064 10613337
Jernström H Vesprini D Bradlow HL Narod SA Correspondence- Re: CYP17 promoter polymorphism and breast cancer in Ausralian women under age forty years J Natl Cancer Inst 2001 93 554 11287451 10.1093/jnci/93.7.554
Muti P Bradlow HL Micheli A Krogh V Freudenheim JL Schunemann HJ Stanulla M Yang J Sepkovic DW Trevisan M Estrogen metabolism and risk of breast cancer: a prospective study of the 2:16alpha-hydroxyestrone ratio in pre-menopausal and postmenopausal women Epidemiology 2000 11 635 640 11055622 10.1097/00001648-200011000-00004
| 15987450 | PMC1175058 | CC BY | 2021-01-04 16:54:50 | no | Breast Cancer Res. 2005 Apr 12; 7(4):R455-R464 | utf-8 | Breast Cancer Res | 2,005 | 10.1186/bcr1027 | oa_comm |
==== Front
Breast Cancer ResBreast Cancer Research1465-54111465-542XBioMed Central London bcr10281598746410.1186/bcr1028Research ArticleInhibition of insulin-like growth factor-1 receptor signaling enhances growth-inhibitory and proapoptotic effects of gefitinib (Iressa) in human breast cancer cells Camirand Anne [email protected] Mahvash [email protected] Fiona [email protected] Michael [email protected] Lady Davis Institute for Medical Research and Department of Oncology, McGill University, Montréal, QC, Canada2 Biology Department, Faculty of Science, Az-Zahra University, Vanak, Tehran, Iran2005 12 4 2005 7 4 R570 R579 10 11 2004 20 1 2005 23 2 2005 18 3 2005 Copyright © 2005 Camirand et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Introduction
Gefitinib (Iressa, ZD 1839, AstraZeneca) blocks the tyrosine kinase activity of the epidermal growth factor receptor (EGFR) and inhibits proliferation of several human cancer cell types including breast cancer. Phase II clinical trials with gefitinib monotherapy showed an objective response of 9 to 19% in non-small-cell lung cancer patients and less than 10% for breast cancer, and phase III results have indicated no benefit of gefitinib in combination with chemotherapy over chemotherapy alone. In order to improve the antineoplastic activity of gefitinib, we investigated the effects of blocking the signalling of the insulin-like growth factor 1 receptor (IGF-1R), a tyrosine kinase with a crucial role in malignancy that is coexpressed with EGFR in most human primary breast carcinomas.
Methods
AG1024 (an inhibitor of IGF-1R) was used with gefitinib for treatment of MDA468, MDA231, SK-BR-3, and MCF-7 breast cancer lines, which express similar levels of IGF-1R but varying levels of EGFR. Proliferation assays, apoptosis induction studies, and Western blot analyses were conducted with cells treated with AG1024 and gefitinib as single agents and in combination.
Results
Gefitinib and AG1024 reduced proliferation in all lines when used as single agents, and when used in combination revealed an additive-to-synergistic effect on cell growth inhibition. Flow cytometry measurements of cells stained with annexin V-propidium iodide and cells stained for caspase-3 activation indicated that adding an IGF-1R-targeting strategy to gefitinib results in higher levels of apoptosis than are achieved with gefitinib alone. Gefitinib either reduced or completely inhibited p42/p44 Erk kinase phosphorylation, depending on the cell line, while Akt phosphorylation was reduced by a combination of the two agents. Overexpression of IGF-1R in SK-BR-3 cells was sufficient to cause a marked enhancement in gefitinib resistance.
Conclusion
These results indicate that IGF-1R signaling reduces the antiproliferative effects of gefitinib in several breast cancer cell lines, and that the addition of an anti-IGF-1R strategy to gefitinib treatment may be more effective than a single-agent approach.
==== Body
Introduction
The signaling activity of receptor protein tyrosine kinases (PTKs) is crucial to the control of apoptosis, differentiation, and proliferation processes; consequently, dysfunction or deregulation of these molecules can lead to uncontrolled growth and neoplastic progression. The abnormal activation of PTKs in the pathology of many cancers has called attention to these receptors as potential targets for therapeutic intervention [1-4]. Some neoplastic conditions arise from excessive activity of a single PTK, for example Bcr-Abl in chronic myeloid leukaemia [5], or c-kit or platelet-derived growth factor receptor-α in gastrointestinal stromal cell tumours [6], and these conditions are effectively treated using the PTK inhibitor Gleevec (Imatinib mesylate) [7]. However, most cancers have complex biochemical causes and may involve dysfunction of several PTKs as well as crosstalk between downstream signaling pathways. One approach to address the multiplicity problem involves cotargeting different PTKs [8-17], but for maximal efficacy, the choice of PTKs to be simultaneously blocked in any specific cancer type is crucial.
The epidermal growth factor receptor (EGFR, erbB1, or HER1) is a 170-kDa member of the erbB family of PTKs, which are transmembrane receptors with important roles in development, differentiation, proliferation, and migration [18]. The activation of EGFR by ligand binding causes dimerization and autophosphorylation of the receptor and subsequent recruitment of downstream molecules, leading to mitogenic signaling [19]. EGFR is overexpressed in a large subset of primary breast carcinomas, and EGFR ligands such as TGF-α and amphiregulin are found in 50 to 90% of primary breast carcinomas [20]. The co-occurrence of these sets of factors is associated with poor prognosis and resistance to hormonal therapy [21].
Several anti-EGFR molecules have been shown to cause neoplastic growth inhibition [22]. Among these, gefitinib (Iressa; AstraZeneca) is an orally active synthetic anilinoquinazoline (4-(3-chloro-4-fluroanilino)-7-methoxy-6-(3-morpholinopropoxy) quinazoline) that inhibits EGFR but also has activity against erbB2 and vascular endothelial growth factor receptor 2 (VEGFR-2) at 100-fold higher than those needed for EGFR inhibition [23]. It has proved an effective inhibitor of proliferation in experimental human breast cancer cell systems, either alone or in combination with other antineoplastic agents [10,11,14,24-32]. Gefitinib as second- or third-line monotherapy in phase II trials of non-small-cell lung cancer patients provided objective tumour response rates of 9 to 19% [22,33,34]. A response rate of 10.8% was also seen in head and neck cancer patients [35], but phase II trials in advanced breast cancer patients showed partial response in fewer than 10% of patients [36-38]. Non-small-cell lung cancer phase III trials where gefitinib was used in combination with traditional chemotherapy (paclitaxel, gemcitabine, or cisplatin) showed no added benefit of gefitinib to patients over chemotherapy alone [39,40]. The acceptable safety profile of gefitinib was, however, confirmed by these studies, and the results motivate studies to determine if PTK cotargeting might improve the efficacy of the drug.
A potential cotarget receptor in breast cancer is the insulin-like growth factor 1 receptor (IGF-1R). In its mature form, IGF-1R is a heterotetrameric receptor (two extracellular 125-kDa α chains and two transmembrane 95-kDa β chains) that autophosphorylates after ligand binding and activates several downstream signaling routes, including the phosphatidylinositol 3-kinase (PI3K) and mitogen-activated protein kinase (MAPK) pathways. Signaling through IGF-1R stimulates proliferation, promotes angiogenesis and metastasis, and inhibits apoptosis [41-45]. There is now abundant evidence indicating that signaling through the IGF-1R pathway is important in many cancers, including breast cancer [4,42,46-49], and recent preclinical work has shown that IGF-1R could be used as a successful cotarget with EGFR in primary human glioblastoma cells [13,50], with c-kit in small-cell lung cancer [15,17], and with HER2/erbB2 in breast cancer cells [12,16]. AG1024 (3-bromo-5-t-butyl-4-hydroxy-benzylidenemalonitrile) is a synthetic tyrphostin that inhibits ligand-stimulated IGF-1R autophosphorylation in intact cells, with an inhibitory concentration 50% (IC50) of 7 μM and can affect the insulin receptor at 9- to 10-fold higher concentrations (IC50 57 μM) [51]. Tyrphostins bind to the active site of receptors and modify its conformation to prevent the substrate and ATP from binding [51]. Through its anti-IGF-1R activity, AG1024 inhibits cell proliferation and induces apoptosis in several cell systems, including non-small-cell lung cancer [52], small-cell lung cancer [17], melanoma [53], and breast cancer [54].
In this study, AG1024 and gefitinib were used to cotarget IGF-1R and EGFR activity in several human breast cancer cell lines that express IGF-1R similarly but present different levels of EGFR. We show that combination treatment causes additivity or synergy in growth inhibition and apoptosis induction, and we speculate that adding an anti-IGF-1R strategy to gefitinib treatment may be more effective than single-agent gefitinib therapy.
Materials and methods
Chemicals and drugs
Gefitinib (ZD 1839, Iressa) was a gift from AstraZeneca (Macclesfield, UK). AG1024 was purchased from Calbiochem-EMD Biosciences (La Jolla, CA, USA).
Cell lines and proliferation assays
Breast cancer cell lines MCF-7, MDA468, MDA231, and SK-BR-3 were obtained from ATCC (Manassas, VA, USA). Cells were cultured at 37°C with 5% CO2 in RPMI 1640 (MCF-7) or McCoy medium (all other cell lines) with 10% fetal bovine serum (FBS) (InVitrogen, Gaithersburg, MD, USA), except in growth inhibition assays, where the FBS supplement was reduced to 1%. Cell proliferation was measured with the Alamar Blue dye reduction method (Biosource International, Camarillo, CA, USA). Growth tests were conducted with 104 cells/well in 200 μl media in 96-well plates, and three replicates per dose combination were used for each experiment. Experiments shown here are representative of three repeats. Stock solutions of tyrphostin AG1024 and gefitinib were made in dimethyl sulfoxide to 10 mM, stored at -20°C, and diluted in medium containing 1% FBS just before use. The concentration of dimethyl sulfoxide in the final culture was kept below 0.2% (v/v). All procedures involving tyrphostins were conducted in low light intensity.
Flow cytometry for receptors
Medium was removed from breast cancer cells growing in monolayers, and cells were collected by scraping in 1 ml 4°C FACS (fluorescence-activated cell sorter) buffer (3% fetal bovine serum, 0.02% NaN3 in PBS). Cells were centrifuged and washed in FACS buffer; approximately 106 cells were stained with phycoerythrin-conjugated anti-IGF-1R (BD Pharmingen, San Diego, CA, USA), or with fluorescein-isothiocyanate-conjugated (FITC-conjugated) anti-EGFR antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) for 30 min at 4°C in the dark, washed twice in FACS, and resuspended in the same buffer. Analysis was conducted for 20,000 cells using a FACSCalibur flow cytometer (BD Biosciences, Burlington, MA, USA) with CellQuest software (BD Biosciences Immunocytometry Systems, Franklin Lakes, NJ, USA). Normal mouse IgG1 (Santa Cruz Biotechnology) was used for isotype determination. All tests were conducted in duplicate and the experiments shown here are representative of two repeats.
Flow cytometry for apoptosis induction
Growth medium was removed from breast cancer cells growing in monolayers; adherent cells were briefly trypsinized, detached, combined with floating cells from the original growth medium, centrifuged, and washed twice with PBS. Approximately 106 cells were stained for 30 min with annexinV–FITC and propidium iodide using the ApoTarget kit (Biosource International). Analysis was conducted on a FACSCalibur flow cytometer using CellQuest software (see receptor section, above). For quantification of caspase-3 activation, cells (approximately 0.5 × 106) were obtained as for testing with annexinV and propidium iodide, but were washed in media, resuspended in 150 μl media containing 10% FBS and 0.5 μl Red-DEVD-FMK (Caspase-3 detection kit, Calbiochem-EMD Biosciences), and incubated for 30 min at 37°C in a cell-culture incubator with 5% CO2. The stained cells were centrifuged, washed twice with the wash buffer provided in the kit, resuspended in 500 μl of the same buffer, and analyzed for fluorescence on a FACSCalibur flow cytometer using CellQuest software. All apoptosis tests were conducted in duplicate and results shown are representative of three experiments.
Western blotting and immunoprecipitation
Cells growing in monolayers in 10-cm culture plates were treated with various doses of AG1024, gefitinib, or vehicle for 24 or 72 hours, then lysed in nondenaturing buffer (1% Nonidet NP-40, 20 mM TrisCl pH 8.0; 0.5 mM sodium orthovanadate, pH 9.0; and proteinase inhibitors (Roche, Mannheim, Germany)), and particulate material was removed by centrifugation at 4°C. Samples (50 μg) of the supernatant were separated on 10% or 15% polyacrylamide gels. After transfer to TransBlot nitrocellulose membranes (BioRad, Hercules, CA, USA), the proteins were reacted overnight with the following primary antibodies at 1:1,000 dilution: anti-Akt, anti-phospho-Akt (Ser473), anti-Erk1/Erk2 (p44/42) anti-phospho-Erk1/Erk2 (Thr202/Tyr204), and anti-EGFR (Cell Signalling Technologies, Beverly, MA, USA). Anti-phospho-EGFR (Tyr1173) was from Upstate (Charlottesville, VA, USA). Blots were then reacted for 1 hour with 1:2,000 horseradish-peroxidase-conjugated antirabbit immunoglobulin G (Pharmacia-Amersham, Piscataway, NJ, USA). Tubulin 1:200 (Santa Cruz Biotechnology) and antimouse immunoglobulin G (Pharmacia-Amersham) were used to check evenness of loading. Membranes were reacted with enhanced chemiluminescence (ECL) reagents (Pharmacia-Amersham) and exposed to X-OMAT LS film (Kodak, Rochester, NJ, USA). For immunoprecipitation, 500-μg samples of soluble protein in a final volume of 500 μl were incubated with 10 μl antiphosphotyrosine monoclonal antibody (BD Pharmingen, Mississauga, ON, Canada) with rotation at 4°C overnight. A mixture (20 μl) of protein A and G+ Sepharose beads (Santa Cruz Biotechnology) was then added, and the samples were rotated at 4°C for 1 hour. Beads were collected by centrifugation, washed once with lysis buffer, heated for 5 min at 95°C in SDS–PAGE loading buffer, and separated by electrophoresis. Membranes after transfer were reacted with an anti-IGF-1R β-subunit antibody (Santa Cruz Biotechnology) and processed as above for enhanced chemiluminescence detection. Western blot analyses were repeated twice.
Statistical analysis
Statistical validity was evaluated using Student's t-test or the Student Newmany–Keuls test for multiple pairwise comparisons of means with Statistical Analysis System software, version 8 (SAS Institute, Cary, NC, USA), with P values ≤ 0.05 considered significant.
Results
Surface expression of IGF-1R and EGFR in breast cancer cell lines
The breast cancer cell lines tested exhibit similar surface expression of the IGF-1 receptor, but the number of EGF receptors varied considerably, with MDA468 cells showing very high expression, MDA231 intermediate levels, SK-BR-3 low expression, and MCF-7 no significant presence of EGFR (Fig. 1).
Inhibition of IGF-1R signaling enhances the effect of gefitinib on the proliferation of breast cancer cell lines
In the culture conditions used here, proliferation IC50 values (means ± standard deviations) for AG1024 were 3.5 μM ± 0.4 for MDA468; 3.5 μM ± 0.5 for MCF-7; 4.5 μM ± 0.4 for MDA231; and 2.5 ± 0.4 for SK-BR-3 cells. The respective IC50 values for gefitinib were 8.0 μM ± 1.0; 9.2 μM ± 2.3; 11.5 μM ± 3.0; and 6.5 μM ± 1.5.
The use of treatments combining AG1024 and gefitinib revealed that the cotargeting approach achieved a greater growth inhibition (Fig. 2a). Combination index (CI) values calculated according to the classic isobologram equation [55] evaluate the interactions between agents as additive (CI approximately 1), antagonistic (CI >1), or synergistic (CI <1). The results (Fig. 2b) indicate synergy (for MDA468) or additivity (other cell lines) of interaction between AG1024 and gefitinib.
Adding an anti-IGF-1R strategy to gefitinib treatment increases levels of apoptosis
Flow cytometric analyses of breast cancer cells treated with AG1024, gefitinib, or both, and stained with annexinV and propidium iodide (cells treated for 3 days) or with red–DEVD–FMK for caspase-3 activation (cells treated for 1 day) are shown in Fig. 3a,b. In all cell lines, and for both methods of detecting apoptosis, conditions were found where addition of AG1024 significantly increased apoptosis levels over those seen with gefitinib alone.
Effect of treatment with AG1024 or gefitinib on protein and phosphorylation levels of Akt and p44/p42 Erk kinases
After 24 hours of treatment, gefitinib decreased the levels of Erk phosphorylation in most cell lines, and completely eliminated Erk phosphorylation in MDA468 (Fig. 4). In contrast, the phosphorylation levels of Akt were reduced by the combination of the two agents. Erk and Akt protein levels were not affected by the 24-hour treatments. Tubulin levels confirmed equal loading (not shown).
Overexpression of IGF-1R greatly reduces sensitivity to gefitinib
SK-BR-3 cells transfected to overexpress the IGF-1 receptor (SK-BR-3-IR) [12] were tested for sensitivity to gefitinib. Figure 5a illustrates the high IGF-1R expression levels observed by flow cytometry in SK-BR-3-IR cells compared with the levels in the SK-BR-3 parental line shown in Fig. 1. Increased expression of IGF-1R caused a very marked enhancement in resistance to the growth-inhibitory effects of gefitinib (Fig. 5b).
Effect of treatment with AG1024 or gefitinib on tyrosine phosphorylation of IGF-1R and EGFR
An example of the effect of AG1024, gefitinib, or both (24-hour treatment) on the phosphorylation levels of IGF-1 and EGF receptors in 1% serum conditions is illustrated in Fig. 6. In MCF-7 cells (left), AG1024 at 2.5 μM eliminated phosphorylation of IGF-1R, while gefitinib did not affect the phosphorylation state of IGF-1R. EGFR phosphorylation levels were decreased by gefitinib, but only slightly affected by AG1024 treatment (MDA468 cells, Fig. 6, right). Protein levels for both receptors were unaffected by treatment in the conditions used here.
Discussion
Several reports have suggested that cotargeting protein tyrosine kinases results in substantial enhancement of growth inhibition [8-17]. In the present study, the choice of the IGF-1 receptor as cotarget is based on the knowledge that this receptor drives important cell survival pathways [41-45] and that reduction of its antiapoptotic effects increases the efficacy of treatments targeting several other neoplasia-related PTKs [12,13,15-17]. The results presented here show the effects of adding an anti-IGF-1R strategy to gefitinib treatment in human breast cancer cell lines chosen for their similar expression of IGF-1R but their different EGFR levels (Fig. 1).
Gefitinib and AG1024 used as single agents show antiproliferative activity on all cell lines tested, and their combination produces an additive-to-synergistic enhancement of growth inhibition (Fig. 2a,b). The mechanism of action on cancer cells of EGFR blockers such as AG 1478, mAb225, and gefitinib is generally cytostatic and proceeds via a G0/G1 arrest [56]. Most breast cancer cells are growth-arrested by gefitinib, but only a subset shows induction of apoptosis (cytotoxic effect) [31], and high doses of the drug are needed to induce apoptosis in normal mammary epithelial cells and primary cultures of mammary carcinoma cells [24]. Blocking the antiapoptotic IGF-1R pathway with AG1024 improves apoptosis induction over the level due to treatment with gefitinib alone (Fig. 3). All the cell lines tested exhibited this effect, regardless of the levels of expression of EGFR. In fact, the growth-inhibitory effect of gefitinib has been reported to be independent of the levels of expression of EGFR in human breast cancer cells [10,24-26,31] and other cancer cell lines [57]. As the EGFR expression level is not a good predictor of gefitinib sensitivity [58], EGFR expression status in tumours cannot be used to exclude patients from gefitinib trials [59]. It has been shown that the presence of somatic mutations in the EGFR gene in lung cancer samples correlates with sensitivity to gefitinib [60,61]. However, even in the absence of detectable EGFR (as in MCF-7 cells: our results, Fig. 1, and [10]), gefitinib and AG1024 still have additive capability, raising the possiblity of a non-EGFR-specific gefitinib effect that can be enhanced by the anti-IGF-1R agent.
Western blot analysis (Fig. 4) showed that after a 24-hour treatment, gefitinib affects phosphorylation levels of p44/p42 Erk and Akt kinases, but that combination treatment with the anti-IGF-1R agent causes a further reduction in levels of Akt phosphorylation. The effect is particularly visible for MDA468 cells, which probably reflects the fact that these cells show a synergistic rather than additive growth reduction pattern. Interestingly, MDA468 cells (PTEN-null) (phosphatase-and-tensin-homolog-null) have been reported to show a relative resistance to gefitinib that can be reversed through the use of the PI3K inhibitor LY294002 [62] or PTEN reconstitution [30], pointing to a crucial role for receptors that signal through the PI3K cascade, such as IGF-1R. MDA468 cells are also the most sensitive to gefitinib inhibition of Erk phosphorylation. In longer treatments (not shown), the levels of protein expression for Akt and Erk are decreased by AG1024 or by the combination of agents. AG1024 treatment has been reported to decrease the expression of several proteins known as regulators of apoptosis and the cell cycle [53,54], and the inhibitor may therefore also provide a longer-term inhibitory effect by mechanisms involving protein degradation.
An important point, illustrated in Fig. 5, is that overexpression of the IGF-1 receptor results in increased resistance to gefitinib. This observation implies that one way in which breast cancer cells resist gefitinib is through the signaling activity of IGF-1R. Since gefitinib does not affect phosphorylation of the IGF-1 receptor (Fig. 6 and [63]), our results suggest that the antiapoptotic pathways driven by IGF-1 signalling should be targeted in order to optimize the antineoplastic effects of gefitinib. While our model system involves increased IGF-1R activity due to receptor overexpression, it must be noted that increased IGF-1R signaling in clinical breast cancer might also arise from mechanisms involving abnormally high IGF-2 expression or from derangements in IGF-binding protein physiology [42].
The findings described here suggest that the antineoplastic effects of gefitinib may be significantly underestimated if examined only under conditions in which IGF-IR is fully functional. Several anti-IGF-1R compounds are now being developed for clinical evaluation [64-67], and it should soon be feasible to conduct trials to test the hypothesis that the efficacy of gefitinib treatments is enhanced by IGF-1R targeting. The data presented here support further research into breast cancer therapeutic strategies combining gefitinib with anti-IGF-1R agents.
Conclusion
In several human breast cancer cell lines, addition of the IGF-1R inhibitor AG1024 to gefitinib reduced cell proliferation in an additive or synergistic fashion and enhanced the induction of apoptosis over levels achieved by gefitinib alone. This effect was independent of levels of expression of the EGF receptor. Overexpression of IGF-1R in SK-BR-3 cells was sufficient to cause a marked enhancement in gefitinib resistance. IGF-1R signaling can therefore limit the antiproliferative effects of gefitinib in vitro, and we speculate that for a subset of human breast cancers, adding an anti-IGF-1R strategy to gefitinib treatment may be more effective than a single-agent approach.
Abbreviations
CI = combination index; EGFR = epidermal growth factor receptor; FACS = fluorescence-activated cell sorter; FBS = fetal bovine serum; FITC = fluorescein isothiocyanate; IC50 = inhibitory concentration 50%; IGF-1R = insulin like growth factor 1 receptor; PI3K = phosphatidylinositol 3-kinase; PTEN = phosphatase and tensin homolog; PTK = protein tyrosine kinase; VEGFR = vascular endothelial growth factor receptor.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
Anne Camirand: study design, data collection, statistical analysis, data interpretation, manuscript preparation, literature search and funds collection. Mahvash Zakikhani: data collection, statistical analysis, data interpretation, manuscript preparation, literature search. Fiona Young: data collection. Michael Pollak: study design, funds collection.
All authors read and approved the final manuscript.
Acknowledgements
This work was supported by a grant to MP and AC from the Susan G Komen Foundation for Breast Cancer Research. Gefitinib was a gift from AstraZeneca (Macclesfield, UK).
Figures and Tables
Figure 1 Surface expression of IGF-1R and EGFR in human breast cancer cell lines. Untreated cells were stained with phycoerythrin-conjugated anti-IGF-1R (insulin like growth factor-1 receptor) or with fluorescein-isothiocyanate-conjugated anti-EGFR (anti-epidermal growth factor receptor) antibody. Shaded peaks show flow cytometry analysis of the number of insulin-like growth factor 1 (top row) and epidermal growth factor (bottom row) receptors on the surface of MDA468, MCF-7, MDA231, and SK-BR-3 human breast cancer cells. Outlined peak represents isotype control (normal mouse IgG1). Counts indicate number of events.
Figure 2 Inhibition of breast cancer cell growth by AG1024 and gefitinib singly and in combination. (a) Cells in exponential stages of growth were exposed to increasing concentrations of inhibitors for 72 hours in media containing 1% fetal bovine serum. Triplicates were used for each dose combination for each experiment. (b) Proliferation combination index (CI) values were calculated using the classic isobologram equation [55] and indicate synergy (CI < 1) or additivity (CI approximately 1).
Figure 3 Treatment with AG1024 enhances apoptotic effects of gefitinib. (a) Flow cytometric analysis of apoptosis in cells stained with annexin V and propidium iodide after 72-hour treatment of breast cancer cells with AG1024 (AG), gefitinib (GE), or both. (b) Flow cytometric analysis of caspase-3 induction by Red-DEVD-FMK fluorescence after 24-hour treatment of cells with AG1024, gefitinib, or both. Addition of AG1024 to gefitinib treatment significantly enhanced apoptotic induction over levels achieved by gefitinib alone. Values on horizontal axes are concentrations (μM). *P < 0.05.
Figure 4 Effect of treatment on Erk and Akt kinases phosphorylation and protein levels. Western blot analysis showing phosphorylation (P) (top) and protein (bottom) levels of Akt and p44/p42 Erk kinases in cells treated for 24 hours with AG1024 (AG), gefitinib (GE), or both.
Figure 5 Overexpression of IGF-1R results in enhanced resistance to gefitinib antiproliferation activity. (a) Flow cytometry estimation of surface expression levels of epidermal growth factor 1 receptor (IGF-1R) (shaded peak) in the SK-BR-3-IR line, which differs from its parental line, SK-BR-3, only by the transfected IGF-1R gene [12]. Outlined peak is normal mouse IgG1 isotype. Counts indicate number of events. (b) Effect of a 72-hour treatment with gefitinib alone on the proliferation of SK-BR-3 and SK-BR-3-IR cells. Triplicates were used for each dose.
Figure 6 Single agent treatment effect on tyrosine phosphorylation (P; top) and protein expression of receptors. Western blot showing phosphorylation (P, top) and receptor protein levels (bottom). MCF-7 cells were used for detection of phosphorylation of the insulin like growth factor 1 receptor (IGF-1R) and MDA468 cells, for phosphorylation analysis of the epidermal growth factor receptor (EGFR). AG, AG1024; GE, gefitinib.
==== Refs
Arteaga CL Moulder SL Yakes FM HER (erbB) tyrosine kinase inhibitors in the treatment of breast cancer Semin Oncol 2002 29 4 10 12138392 10.1053/sonc.2002.34047
Shawver LK Slamon D Ullrich A Smart drugs: Tyrosine kinase inhibitors in cancer therapy Cancer Cell 2002 1 117 123 12086869 10.1016/S1535-6108(02)00039-9
Dancey J Sausville EA Issues and progress with protein kinase inhibitors for cancer treatment Nat Rev Drug Discov 2003 2 296 313 12669029 10.1038/nrd1066
Surmacz E Growth factor receptors as therapeutic targets: strategies to inhibit the insulin-like growth factor-I receptor Oncogene 2003 22 6589 6597 14528284 10.1038/sj.onc.1206772
Melo JV The diversity of BCR-ABL fusion proteins and their relationship to leukemia phenotype Blood 1996 88 2375 2384 8839828
Heinrich MC Corless CL Duensing A McGreevey L Chen C-J Joseph N Singer S Griffith DJ Haley A Town A PDGFRA activating mutations in gastrointestinal stromal tumors Science 2003 299 708 710 12522257 10.1126/science.1079666
Druker BL Talpaz M Resta DJ Peng B Buchdunger E Ford JM Lydon NB Kantardjian H Capdeville R Ohno-Jones S Efficacy and safety of a specific inhibitor of the Bcr-Abl tyrosine kinase in chronic myeloid leukemia New Engl J Med 2001 344 1031 1037 11287972 10.1056/NEJM200104053441401
Wu X Fan Z Masui H Rosen N Mendelsohn J Apoptosis induced by an anti-EGFR monoclonal antibody in human colorectal carcinoma cell line and its delay by insulin J Clin Invest 1995 95 1897 1905 7706497
Ye D Mendelsohn J Fan Z Augmentation of a humanized anti-HER2 mAb 4D5 induced growth inhibition by a human-mouse chimeric anti-EGF receptor mAb C225 Oncogene 1999 18 731 738 9989823 10.1038/sj.onc.1202319
Moasser MM Basso A Averbuch SD Rosen N The tyrosine kinase inhibitor ZD1839 ("Iressa") inhibits HER2-driven signaling and suppresses the growth of HER2-overexpressing tumor cells Cancer Res 2001 61 7184 7188 11585753
Moulder SL Yakes FM Muthuswamy SK Bianco R Simpson JF Arteaga CL Epidermal growth factor receptor (HER1) tyrosine kinase inhibitor ZD1839 (Iressa) inhibits HER2/neu (erbB2)-overexpressing breast cancer cells in vitro and in vivo Cancer Res 2001 61 8887 8895 11751413
Lu Y Zi X Zhao Y Mascarenhas D Pollak M Insulin-like growth factor-I receptor signaling and resistance to trastuzumab (Herceptin) J Natl Cancer Inst 2001 93 1852 1857 11752009 10.1093/jnci/93.24.1852
Chakravarti A Loeffler JS Dyson NJ Insulin-like growth factor receptor I mediates resistance to anti-epidermal growth factor receptor therapy in primary human glioblastoma cells through continued activation of phosphoinositide 3-kinase signaling Cancer Res 2002 62 200 207 11782378
Normanno N Campiglio M De LA Somenzi G Maiello M Ciardiello F Gianni L Salomon DS Menard S Cooperative inhibitory effect of ZD1839 (Iressa) in combination with trastuzumab (Herceptin) on human breast cancer cell growth Ann Oncol 2002 13 65 72 11863114 10.1093/annonc/mdf020
Warshamana-Grene GS Litz J Buchdunger E Hofmann F Garcia-Echeverria C Krystal GW The insulin-like growth factor-I (IGF-I) receptor kinase inhibitor NVP-ADW742, in combination with STI571, delineates a spectrum of dependence of small cell lung cancer on IGF-I and stem cell factor signalling Mol Cancer Ther 2004 3 527 535 15141010
Camirand A Lu Y Pollak M Co-targeting HER2/ErbB2 and insulin-like growth factor-1 receptors causes synergistic inhibition of growth in HER2-overexpressing breast cancer cells Med Sci Monit 2002 8 BR521 BR526 12503030
Camirand A Pollak M Co-targeting IGF-1R and c-kit: synergistic inhibition of proliferation and induction of apoptosis in H 209 small cell lung cancer cells Br J Cancer 2004 90 1825 1829 15150607
Rusch V Mendelsohn J Dmitrovsky E The epidermal growth factor receptor and its ligands as therapeutic targets in human tumors Cytokine Growth Factor Rev 1996 7 133 141 8899291 10.1016/1359-6101(96)00016-0
Yarden Y The EGFR family and its ligands in human cancer: signalling mechanisms and therapeutics opportunities Eur J Cancer 2001 37 S3 S8 11597398 10.1016/S0959-8049(01)00230-1
Salomon DS Brandt R Ciardiello F Normanno N Epidermal growth factor-related peptides and their receptors in human malignancies Crit Rev Oncol Hematol 1995 19 183 232 7612182
Woodburn JR The epidermal growth factor receptor and its inhibition in cancer therapy Pharmacol Ther 1999 82 241 250 10454201 10.1016/S0163-7258(98)00045-X
Blackledge G Averbuch S Gefitinib ('Iressa', ZD 1839) and new epidermal growth factor receptor inhibitors Br J Cancer 2004 90 566 572 14760365 10.1038/sj.bjc.6601550
Ranson M ZD1839 (Iressa): for more than just non-small cell lung cancer Oncologist 2002 7 16 24 12202784
Ciardiello F Caputo R Bianco R Damiano V Pomatico G De Placido S Bianco AR Tortora G Antitumor effect and potentiation of cytotoxic drugs activity in human cancer cells by ZD-1839 (Iressa), an epidermal growth factor receptor-selective tyrosine kinase inhibitor Clin Cancer Res 2000 6 2053 2063 10815932
Ciardiello F Caputo R Bianco R Damiano V Fontanini G Cuccato S De Placido S Raffaele Bianco A Tortora G Inhibition of growth factor production and angiogenesis in human cancer cells by ZD1839 (Iressa)1, a selective epidermal growth factor receptor tyrosine kinase inhibitor Clinical Cancer Research 2001 7 1459 1465 11350918
Anderson NG Ahmad T Chan K Dobson R Bundred NJ ZD1839 (Iressa), a novel epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor, potently inhibits the growth of EGFR-positive cancer cell lines with or without erbB2 overexpression Int J Cancer 2001 94 774 782 11745477 10.1002/ijc.1557
Nicholson RI Hutcheson IR Harper ME Knowlden JM Barrow D McClelland RA Jones HE Wakeling AE Gee JMW Modulation of epidermal growth factor receptor in endocrine-resistant, oestrogen receptor-positive breast cancer Endocr Relat Cancer 2001 8 175 182 11566608 10.1677/erc.0.0080175
Morris C The role of EGFR-directed therapy in the treatment of breast cancer Breast Cancer Res Treat 2002 75 S51 S55 12353824 10.1023/A:1020370018668
Arteaga C Targeting HER1/EGFR: a molecular approach to cancer therapy Semin Oncol 2003 30 (3 Suppl 7) 3 14 10.1016/S0093-7754(03)00185-4
Bianco R Shin I Ritter CA Yakes FM Basso A Rosen N Tsurutani J Dennis PA Mills GB Arteaga CL Loss of PTEN/MMAC1/TEP in EGF receptor-expressing tumor cells counteracts the antitumor action of EGFR tyrosine kinase inhibitors Oncogene 2003 22 2812 2822 12743604 10.1038/sj.onc.1206388
Campiglio M Locatelli A Olgiati C Normanno N Somenzi G Vigano L Fumagalli M Menard S Gianni L Inhibition of proliferation and induction of apoptosis in breast cancer cells by the epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor ZD1839 ('Iressa') is independent of EGFR expression level J Cell Physiol 2004 198 259 268 14603528 10.1002/jcp.10411
Gee JMW Harper ME Hutcheson IR Madden TA Barrow D Knowlden JM McClelland RA Jordan N Wakeling AE Nichelson RI The antiepidermal growth factor receptor agent gefitinib (ZD1839/Iressa) inproves antihormone response and prevents development of resistance in breast cancer in vitro Endocrinology 2003 144 5105 5117 12960029 10.1210/en.2003-0705
Kris MG Natale RB Herbst RS Lynch TJ Prager D Belani CP Schiller JH Kelly K Spiridonidis H Sandler A Efficacy of gefitinib, an inhibitor of the epidermal growth factor receptor tyrosine kinase, in symptomatic patients with non-small cell lung cancer, a randomized trial JAMA 2003 290 2149 2158 14570950 10.1001/jama.290.16.2149
Fukuoka M Yano S Giaccone G Tamura T Nakagawa K Douillard J-Y Nishiwaki Y Vansteenkiste J Kudoh S Rischin D Multi-institutional randomized phase II trial of gefitinib for previously treated patients with adanced non-small-cell lung cancer J Clin Oncol 2003 21 2237 2246 12748244 10.1200/JCO.2003.10.038
Cohen EEW Rosen F Stadler WM Recant W Stenson K Huo D Vokes EE Phase II trial of ZD1839 in recurrent or metastatic squamous cell carcinoma of the head and neck J Clin Oncol 2003 21 1980 1987 12743152 10.1200/JCO.2003.10.051
Baselga J Albanell J Ruiz A Lluch A Gascon S Guillen V Sauleda S Averbuch S Rojo F Phase II and tumor pharmacodynamic study of gefinitib (ZD 1839) in patients with advanced breast cancer [abstract] Proc Am Soc Clin Oncol 2003 22 S24
Robertson JFR Gutteridge E Cheung KL Owers R Koehler M Hamilton L Gee J Nicholson RI Gefitinib (ZD 1839) is active in acquired tamoxifen (TAM)-resistant oestrogen receptor (ER)-positive and ER-negative breast cancer: Results from a phase II study [abstract] Proc Am Soc Clin Oncol 2003 22 S23
von Minckwitz G Jonat W Beckmann M du Bois A Kleeberg U Kühnle H Kettner E Hilfrich J Torode J Schneider A A multicenter phase II trial to evaluate gefitinib ('Iressa', ZD1839) (500 mg/day) in patients with metastatic breast cancer after previous chemotherapy treatment [abstract] Proceedings of the 12th European Conference on Clinical Oncology, Copenhagen, Denmark 21–25 sept. 2003, abstract 437
Giaccone G Herbst RS Manegold C Scagliotti G Rosell R Miller V Natale RB Schiller JH Von Pawel J Pluzanska A Gefitinib in combination with gemcitabine and cisplatin in advanced non-small-cell lung cancer: a phase III trial – INTACT 1 J Clin Oncol 2004 22 777 784 14990632 10.1200/JCO.2004.08.001
Herbst RS Giaccone G Schiller JH Natale RB Miller V Manegold C Scagliotti G Rosell R Oliff I Reeves JA Gefitinib in combination with paclitaxel and carboplatin in advanced non-small-cell lung cancer: a phase III trial – INTACT2 J Clin Oncol 2004 22 785 794 14990633 10.1200/JCO.2004.07.215
Pollak M Insulin-like growth factor physiology and cancer risk Eur J Cancer 2000 36 1224 1228 10882860 10.1016/S0959-8049(00)00102-7
Pollak M Schernhammer ES Hankinson SE Insulin-like growth factors and neoplasia Nat Rev Cancer 2004 4 505 518 15229476 10.1038/nrc1387
Khandwala HM McCutcheon IE Flyvbjerg A Friend KE The effects of insulin-like growth factors on tumorigenesis and neoplastic growth Endocr Rev 2000 21 215 244 10857553 10.1210/er.21.3.215
Yu H Rohan T Role of the insulin-like growth factor family in cancer development and progression J Natl Cancer Inst 2000 92 1472 1489 10995803 10.1093/jnci/92.18.1472
Wang Y Sun Y Insulin-like growth factor receptor-1 as an anti-cancer target: blocking transformation and inducing apoptosis Curr Cancer Drug Targets 2002 2 191 207 12188907 10.2174/1568009023333863
Werner H LeRoith D The role of the insulin-like growth factor system in human cancer Adv Cancer Res 1996 68 183 223 8712068
Dunn SE Ehrlich M Sharp NJ Reiss K Solomon G Hawkins R Baserga R Barrett JC A dominant negative mutant of the insulin-like growth factor-I receptor inhibits the adhesion, invasion, and metastasis of breast cancer Cancer Res 1998 58 3353 3361 9699666
Wood TL Yee D Introduction: IGFs and IGFBPs in the normal mammary gland and in breast cancer J Mammary Gland Biol Neoplasia 2000 5 1 5 10791763 10.1023/A:1009580913795
Yee D The insulin-like growth factor system as a target in breast cancer Semin Oncol 2002 29 86 95 12138402 10.1053/sonc.2002.34060
Steinbach JP Eisenmann C Klumpp A Weller M Co-inhibition of epidermal growth factor receptor and type 1 insulin-like growth factor receptor synergistically sensitizes human malignant glioma cells to CD95L-induced apoptosis Biochem Biophys Res Commun 2004 321 524 530 15358139 10.1016/j.bbrc.2004.06.175
Parrizas M Gazit A Levitzki A Wertheimer E and LeRoith D Specific inhibition of insulin-like growth factor-I and insulin receptor tyrosine kinase activity and biological function of tyrphostins Endocrinology 1997 138 1427 1433 9075698 10.1210/en.138.4.1427
Hurbin A Dubrez L Coll JL Favrot MC Inhibition of apoptosis by amphiregulin via an insulin-like growth factor-1 receptor-dependent pathway in non-small cell lung cancer cell lines J Biol Chem 2002 277 49127 49133 12356750 10.1074/jbc.M207584200
von Willebrand M Zacksenhaus E Cheng E Glazer P Halaban R The tyrphostin AG1024 accelerates the degradation of phosphorylated forms of retinoblastoma protein (pRb) and restores pRb tumor suppressive function in melanoma cells Cancer Res 2003 63 1420 1429 12649208
Wen B Deutsch E Marangoni E Frascona V Maggiorella L Abdulkarim B Chavaudra N Bourhis J Tyrphostin AG 1024 modulates radiosensitivity in human breast cancer cells Br J Cancer 2001 85 2017 2021 11747348 10.1054/bjoc.2001.2171
Berenbaum MC Criteria for analyzing inetactions between biologically active agents Adv Cancer Res 1981 35 269 335 7041539
Busse D Doughty RS Ramsey TT Russel WE Price JO Flanagan WM Shawver LK Arteaga CL Reversible G(1) arrest induced by inhibition of the epidermal growth factor receptor tyrosine kinase requires up-regulation of p27(KIP1) independent of MAPK activity J Biol Chem 2000 275 6987 6995 10702262 10.1074/jbc.275.10.6987
Sirotnak FM Zakowski MF Miller VA Scher HI Kris MG Efficacy of cytotoxic agents against human tumor xenografts is markedly enhanced by coadministration of ZD1839 (Iressa), an inhibitor of EGFR tyrosine kinase Clin Cancer Res 2000 6 4885 4892 11156248
Bailey LR Janas M Schmidt K Bindslev N Wolf M Grous J Askaa J Herbst R Johnson DH Giacone G Evaluation of epidermal growth factor receptor (EGFR) as a predictive marker in patients with non-small-cell lung cancer (NSCLC) receiving first-line gefitinib combined with platinum-based chemotherapy [abstract] J Clin Oncol 2004 22 S7013
Arteaga C Epidermal growth factor receptor dependence in human tumors: more than just expression? Oncologist 2002 7 31 39 12202786
Paez JG Janne PA Lee JC Tracy S Greulich H Gabriel S Herman P Kaye FJ Lindeman N Boggon TJ EGFR mutations in lung cancer: correlation with clinical response to gefitinib therapy Science 2004 304 1497 1500 15118125 10.1126/science.1099314
Lynch TH Bell DW Sordella R Gurubhagavatula S Okimoto RA Brannigan BW Harris PL Haserlat SM Supko JG Haluska FG Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to Gefitinib N Engl J Med 2004 350 2129 2139 15118073 10.1056/NEJMoa040938
She QB Solit D Basso A Moasser MM Resistance to Gefitinib in PTEN-null HER-overexpressing tumor cells can be overcome through restoration of PTEN function or pharmacologic modulation of constitutive phosphatidylinositol 3'-kinase/Akt pathway signaling Clinical Cancer Research 2003 9 4340 4346 14555504
Gilmore AP Valentijn AJ Wang P Ranger AM Bundred N O'Hare MJ Wakeling A Korsmeyer SJ Streuli CH Activation of BAD by therapeutic inhibition of epidermal growth factor receptor and transactivation by isulin-like growth factor receptor J Biol Chem 2002 277 27643 27650 12011069 10.1074/jbc.M108863200
Maloney EK McLaughlin JL Dagdigian NE Garrett LM Connors KM Zhou X-M Blattler WA Chittenden T Singh R An anti-insulin-like growth factor-I receptor antibody that is a potent inhibitor of cancer cell proliferation Cancer Res 2003 63 5073 5083 12941837
Garcia-Echeverria C Pearson MA Marti A Meyer T Mestan J Zimmermann J Gao J Brueggen J Capraro H-G Cozens R Evans DB Fabbro D Furet P Graus-Porta D Liebetanz J Martiny-Baron G Ruetz S Hofmann F In vivo anti-tumor activity of NVP-AEW541 – A novel, potent and selective inhibitor of the IGF-IR kinase Cancer Cell 2004 5 231 239 15050915 10.1016/S1535-6108(04)00051-0
Mitsiades CS Mitsiades NS McMullan CJ Poulaki V Shringarpure R Akiyama M Hideshima T Chauhan D Joseph M Libermann TA Inhibition of the insulin-like growth factor receptor-1 tyrosine kinase activity as a therapeutic strategy for multiple myeloma, other hematologic malignancies, and solid tumors Cancer Cell 2004 5 221 230 15050914 10.1016/S1535-6108(04)00050-9
Goetsch L Gonzalez A Leger O Beck A Pauwels PJ Haeuw JF Corvaia NA A ecombinant humanized anti-insulin-like growth factor receptor type I antibody (h7C10) enhances the antitumor activity of vinorelbine and anti-epidermal growth factor receptor therapy against human cancer xenographs Int J Cancer 2005 113 316 328 15386423 10.1002/ijc.20543
| 15987464 | PMC1175059 | CC BY | 2021-01-04 16:04:33 | no | Breast Cancer Res. 2005 Apr 12; 7(4):R570-R579 | utf-8 | Breast Cancer Res | 2,005 | 10.1186/bcr1028 | oa_comm |
==== Front
Breast Cancer ResBreast Cancer Research1465-54111465-542XBioMed Central London bcr10311598745210.1186/bcr1031Research ArticlePreventive and curative effect of melatonin on mammary carcinogenesis induced by dimethylbenz[a]anthracene in the female Sprague–Dawley rat Lenoir Véronique [email protected] de Jonage-Canonico Marianne Beau 1Perrin Marie-Hélène 1Martin Antoine 2Scholler Robert [email protected]é Bernard [email protected] Laboratoire de neuroendocrinologie, CNRS-FRE2718, UFR Biomédicale des Saints-Pères, Université René Descartes, Paris2 Service Central d'Anatomie et de Cytologie Pathologique, Hôpital Avicenne, Bobigny, France2005 29 4 2005 7 4 R470 R476 8 12 2004 4 2 2005 16 3 2005 31 3 2005 Copyright © 2005 Lenoir et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Introduction
It has been well documented that the pineal hormone, melatonin, which plays a major role in the control of reproduction in mammals, also plays a role in the incidence and growth of breast and mammary cancer. The curative effect of melatonin on the growth of dimethylbenz [a]anthracene-induced (DMBA-induced) mammary adenocarcinoma (ADK) has been previously well documented in the female Sprague–Dawley rat. However, the preventive effect of melatonin in limiting the frequency of cancer initiation has not been well documented.
Methods
The aim of this study was to compare the potency of melatonin to limit the frequency of mammary cancer initiation with its potency to inhibit tumor progression once initiation, at 55 days of age, was achieved. The present study compared the effect of preventive treatment with melatonin (10 mg/kg daily) administered for only 15 days before the administration of DMBA with the effect of long-term (6-month) curative treatment with the same dose of melatonin starting the day after DMBA administration. The rats were followed up for a year after the administration of the DMBA.
Results
The results clearly showed almost identical preventive and curative effects of melatonin on the growth of DMBA-induced mammary ADK. Many hypotheses have been proposed to explain the inhibitory effects of melatonin. However, the mechanisms responsible for its strong preventive effect are still a matter of debate. At least, it can be envisaged that the artificial amplification of the intensity of the circadian rhythm of melatonin could markedly reduce the DNA damage provoked by DMBA and therefore the frequency of cancer initiation.
Conclusion
In view of the present results, obtained in the female Sprague–Dawley rat, it can be envisaged that the long-term inhibition of mammary ADK promotion by a brief, preventive treatment with melatonin could also reduce the risk of breast cancer induced in women by unidentified environmental factors.
==== Body
Introduction
It has been well documented that the pineal hormone, melatonin, besides its well established circadian rhythm, plays a major role in the control of reproduction in mammals [1,2]. The role of the pineal gland, the major source of melatonin, in the development of breast [3] and mammary [4,5] cancer has been clearly documented. Also, it has been shown that melatonin may exert in vivo [6,7] and in vitro [8-10] an oncostatic activity by at least a direct action on the mammary tissue [11] and on the activation of estrogen receptor (ER) for DNA binding [12], or by modulation of the expression of ER mRNA [13], or by an increase of the ER binding activity [14]. Recently, it was shown that melatonin acts as a calmodulin antagonist, inducing conformational changes in the ERα–calmodulin complex and thus impairing the binding of 17β-estradiol (E2) and the ERα–calmodulin complex to DNA and therefore preventing ERα-dependent transcription [15]. Also, 17β-estradiol treatment of pineal glands in an in vitro perifusion system leads to complete blunting of the isoproterenol-induced stimulation of melatonin secretion in 7,12-dimethylbenz [a]anthracene-treated (DMBA-treated) female Sprague–Dawley rats [16].
On the other hand, a single intragastric admininistration of DMBA has been shown to induce mammary tumors in young, cycling female Sprague–Dawley rats [17]. This carcinogen interacts with rapidly proliferating cells in the terminal end buds, forming DNA adducts, which in turn participate in transforming the normal terminal end bud cells to malignant pathways [18-20]. The susceptibility of Sprague–Dawley rats to DMBA is maximal at 55 to 60 days of age and is abolished by ovariectomy, suggesting the inducible action of the carcinogen depends on ovarian secretions [21]. Furthermore, during the latency period, estrous cycles are associated with blunted preovulatory surges of luteinizing hormone and follicle-stimulating hormone [22], an increased surge of 17β-estradiol [23], and disruption of the expression patterns of the genes for hypothalamic gonadotropin-releasing hormone and its pituitary receptor [24]. Moreover, ovariectomized rats pretreated with DMBA exhibit blunted release of luteinizing hormone in response to in vivo estradiol replacement, and reduced release of gonadotropin-releasing hormone as measured in vitro using synaptosomes from the mediobasal hypothalamus [25]
Considering the estrogenic properties of the DMBA molecule [26], it is possible that the carcinogen exerts its long-lasting effects at least on the plasma membrane of estrogen-sensitive neurons [27]. In addition, DMBA can interact with the ER and partially mimic both the positive and negative feedback actions of estradiol in ovariectomized rats [28].
The aim of this study was to compare the potency of melatonin to limit the frequency of cancer initiation with its potency to inhibit tumor progression once initiation by DMBA was achieved.
Our research team has previously shown that the administration of melatonin (10 mg/kg daily) for 6 months markedly reduces the percentage of tumor-bearing animals in DMBA-treated female rats, by 65% [16].
This study compared, over 1 year, the effect of preventive treatment with melatonin (10 mg/kg daily), administered for only 15 days before the administration of DMBA, to the effect of a curative treatment with the same dose of melatonin administered for 6 months after the administration of DMBA.
Materials and methods
Animals
Sixty 40-day-old female Sprague–Dawley rats (Charles River, L'Arbresle, France), 40 days of age, were used. They were housed (five animals per cage) under conditions of controlled temperature (20 to 22°C) and light (12 hours light /12 hours darkness; lights on from 0700 to 1900 hours). The animals had free access to commercial pelleted rat food (UAR, Villemoisson, Epinay-sur-Orge, France) and water. After 5 days of acclimatization, they were randomly assigned to three experimental groups of 20 animals each.
The first group of animals received a daily intragastric administration (in 1 ml of hydroxyethylcellulose at 1%) of melatonin (10 mg/kg daily, given no more than 3 hours before lights were turned off) for 15 days (from 45 to 59 days of age). At age 60 days, at noon on the day after cessation of the treatment, they were given a single intragastric administration of DMBA (Sigma, Saint Quentin Fallavier, France) (75 mg/kg) diluted in 1 ml of sesame oil, as described elsewhere [17].
The second group of animals received a single intragastric administration of DMBA (75 mg/kg) at noon when they were 55 days old. Then, beginning the next day, they received a daily intragastric administration of melatonin (10 mg/kg daily, given no more than 3 hours before the lights were turned off) for 6 months.
The third group of animals was given a single intragastric administration of DMBA (75 mg/kg, given at noon) at 55 days of age.
All the animals were followed up for 12 months after the administration of DMBA.
The occurrence of palpable mammary tumors was recorded every 2 weeks after the intragastric administration of DMBA.
Tumor histology
The number and the size of mammary tumors were recorded and a histological analysis was performed. Tumors were removed when the largest diameter was at least 2 cm. They were dissected, trimmed free of surrounding connective tissue, and placed in 37% formaldehyde. Tumors were fixed in paraffin wax, sectioned, and stained with hematoxylin and eosin. Sections were examined by one of us (AM).
The lesions observed from the removed samples ranged widely from benign to malignant, with pathology closely resembling that seen in human breast tumors. The observed tumors were generally classified as adenocarcinomas (ADKs), with marked nuclear irregularities and numerous mitoses. Only clearly characterized ADKs were taken into account in the results presented in this study.
Statistical analysis
The normality of data was analyzed with the Kolmogorov–Smirnov and Shapiro–Wilk tests.
As a preliminary step in the statistical process, a two-factor analysis of variance was performed: it was based on the comparison of means weighted by the time factor, because the classical underlying assumptions did not hold.
Because the samples were not drawn from gaussian populations, Student's t and the Welsh tests could not be used; therefore, the Mann–Whitney test was applied.
The percentages of animals diagnosed with mammary ADK in each data set were compared using the Fisher exact test (one-sided P-values).
Results
ANOVA two-way analysis
There was strong evidence (P < 0.01) that the ADK number increases within months after DMBA and that there was a very strong decrease of the number of ADK in treated rat groups (P = 0.01). In order to analyse these differences more closely, we looked to see whether the various time groups differed, whether any such difference was due to a decrease in the number of ADK-bearing animals, and whether the ADK number per ADK-bearing animals was reduced.
Comparison of the mean numbers of mammary ADKs between the groups
In control animals, there was an almost linear increase in the mean number of ADKs between the second and ninth months after DMBA administration (Fig. 1). The highest number (3.05 ± 0.59) was found 9 months after the administration of DMBA. In the melatonin-treated rats, there was a significant reduction of the number of ADKs in both the group given preventive treatment and that given curative treatment (Mann–Whitney test). Although the test used is nonparametric, the sample statistics mean and standard error of the mean are shown in Fig. 1 for a better description of the groups.
Twelve months after DMBA administration, values were 1.30 ± 0.30 in the the group given preventive treatment and 0.95 ± 0.23 in the group given curative treatment, vs 3.05 ± 0.59 in the control group (Fig. 1). Also, there was no statistical difference between the preventively and curatively treated groups (1.30 ± 0.30 vs 0.95 ± 0.23) (Mann–Whitney test). Nevertheless, the curative group differed more significantly from the control group than the preventive group (0.002 <P < 0.01), from 6 to 10 months.
Percentage of rats with at least one palpable mammary ADK
In control animals, the first palpable mammary ADK (latency period) appeared 2 months after DMBA administration (Fig. 2). By 6 months after the administration of DMBA, 75% of control rats exhibited at least one palpable mammary ADK.
In animals that were pretreated daily with melatonin for either 15 days before the administration of DMBA or for 6 months after the administration of DMBA, there was a marked reduction in the percentage of animals with mammary ADKs as compared with the control group. Six months after DMBA administration, the ADK rate was 43% in the group given preventive treatment (P = 0.04 vs control group) and 35% in the group given curative treatment (P = 0.01 vs control group), vs 75% in the control group. The difference between the rates in the two experimental groups was not statistically significant. However, the inhibitory effect last significantly longer in the group given preventive treatment than in that given curative treatment (10 months vs 7 months, respectively).
Comparison of the mean number of mammary ADK number between groups of ADK-bearing animals (Fig. 3)
In control animals, between the fifth and the ninth month after DMBA administration there was an almost linear increase in the mean number of ADKs per ADK-bearing animal. At 9 months after administration of DMBA, the mean number (± standard error of the mean) was 4 ± 0.6 ADKs per ADK-bearing animal.
In the melatonin-treated female rats given preventive treatment, there was a reduction in the mean number of ADK per ADK-bearing animal but it did not reach a high degree of statistical significance.
The number of ADKs in the melatonin-treated female rats given curative treatment (1.38 ± 0.18 at 9 months after DMBA administration and 1.58 ± 0.26 at 12 months after DMBA administration) was significantly lower than in the control group (4 ± 0.59) (0.02 <P < 0.05 for month 7 to month 11 and 0.01 <P < 0.02 for month 12).
The differences between the preventive and the curative group (P > 0.1) and between the control and the preventive group (P ≥ 0.05) were not statistically significant. However, the difference between the control and the preventive group was close or equal to 0.05 from month 7 to month 12.
Discussion
The aim of this study was to compare the effect of daily intragastric administration of melatonin for 15 days before the administration of DMBA in the female rat with the effect of such treatment for 6 months after the administration of DMBA. Our results clearly show the preventive and curative effects of melatonin on the growth of DMBA-induced mammary ADK.
The curative effect of melatonin on the growth of DMBA-induced mammary ADK in the female rat has been previously well documented [4,5,16]. In the present study, the maximal inhibitory effect of melatonin on the percentage of rats with mammary ADK was by 62% at 4 months and 68% after 6 months of treatment (at the end of treatment). After the end of treatment, the intensity of the inhibitory effect started to decrease, but it was still significant up to 4 months after the end of treatment (10 months after DMBA administration).
The preventive (protective) effect of a brief period of melatonin treatment before DMBA administration has not been documented under the same experimental conditions. The maximal inhibitory effect of melatonin on the percentage of female rats with mammary ADK was a reduction to 62% relative to controls, at 4 months after DMBA administration. The Inhibition was significant up to 7 months after DMBA administration, but not later. Very similar results were obtained when the mean number of ADKs per rat or the mean number of ADKs per ADK-bearing rat were considered. Regarding the mean number of ADKs per ADK-bearing rat, there was a loss of significance, but there was no statistical difference between the effect of the preventive treatment and the curative treatment.
Interestingly enough, there was no statistical difference between the oncostatic activity afforded by melatonin according to whether it was administered by the preventive or the curative protocol: the incidence and number of ADKs were very similar, up to 7 months after DMBA administration. A major finding of this study is that 7 months after the induction of the carcinogenic process, both the percentage of animals with ADK and the mean number of ADKs per animal were still reduced in animals treated with melatonin for 15 days before DMBA administration.
In this study, the preventive treatment consisted of only 15 days of treatment, for technical reasons: the treatment could not be started before the rats were 40 days old, to avoid interference with the onset of estral cycling. A longer period of pretreatment with melatonin before the administration of DMBA might have led to a more pronounced inhibitory effect on the growth of DMBA-induced mammary ADK.
Many hypotheses have been proposed to explain the inhibitory effects of melatonin. They include modulation of the reproductive neuroendocrine axis (down-regulation of the gonadotropic axis and decrease of estrogen and prolactin levels); immunoenhancing activity; antioxidative properties; and direct antitumorigenic activity [8,9,29-32]. The direct oncostatic effect of metatonin has been mostly studied on in vitro models of estrogen-responsive breast cancer. In MCF-7 cells, melatonin interferes with the estrogen response pathway through the suppression of ERα expression and transactivation to affect the level of growth regulatory genes mediating the mitogenic action of estradiol [33,34]. Also, melatonin inhibits EGF and MAPK-induced cell proliferation via the suppression of linoleic acid uptake and metabolism [35] and it promotes cell differentiation (differentiation of terminal end buds) by reducing the invasiveness of MCF-7 cells and increasing gap functional contacts [35]
It is thought that DMBA acts, at the mammary gland level, throughout the formation of a cascade of metabolites (generated at least in the liver and the mammary gland) which ends with its ultimate carcinogenic form, the 3,4-dihydro-diol-1,2-epoxide, before adduct formation with DNA occurs [36,37]. Melatonin may inhibit adduct formation with DNA because of its free radical scavenging/antioxidant action [38]. In that view, it may be envisaged that the strong artificial amplification of the intensity of the circadian rhythm of melatonin, provoked by an exogenous supply, before the administration of DMBA could play a major preventive role against the induction of the carcinogenic process [39].
In the group of rats given long-term melatonin treatment beginning a day after DMBA administration, DMBA metabolism and adduct formation certainly still occur for at least 5 days (unpublished results). Therefore, the results obtained in the curative group represent the sum of an inhibition of both the initiation and the promotion process. Obviously, if the melatonin treatment had started a week or two after DMBA administration, we might have seen different results. However, in another carcinogen model of mammary tumorigenesis – the N-nitrosomethylurea model of hormone-responsive rat mammary carcinogenesis – melatonin was without effect on carcinogenesis when its administration was restricted to the initiation phase but was quite effective during the promotion phase [40].
Also, changes in circadian rhythms have already been documented to be associated with carcinogenesis [41,42], and it has been shown that the administration of melatonin could restore the deficiency of the circadian clock provoked by the low secretion of melatonin induced by DMBA [16]. Therefore, it can be envisaged that amplifying the intensity of the circadian rhythm of melatonin might help to prevent the induction of the carcinogenic process by DMBA.
Conclusion
In view of the present results obtained in the female Sprague–Dawley rat and of the well-documented oncostatic properties of melatonin, it can be envisaged that the long-term inhibition of mammary ADK promotion by a brief preventive treatment with melatonin could also reduce the risk of breast cancer induced in women by unidentified environmental factors.
Abbreviations
ADK = adenocarcinoma; DMBA = dimethylbenz [a]anthracene; ER = estrogen receptor.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
VL carried out the experimental studies and performed a detailed analysis of the data. MBYDJC equally carried out the experimental studies. MHP made a contribution to the acquisition of data. AM performed the histological analysis. RS participated in the design of the study and performed the statistical analysis. BK conceived the study and participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.
Acknowledgements
The authors gratefully acknowledge the editorial assistance and manuscript preparation of Ms Isabelle Michel.
Figures and Tables
Figure 1 Average number (mean ± standard error of the mean) of mammary adenocarcinomas (ADKs) per rat after daily administration of melatonin (10 mg/kg) for either 15 days before (preventive treatment; open diamonds) or 6 months after (curative treatment; filled diamonds) the administration of dimethylbenz [a]anthracene (DMBA). In both cases, the rats were followed up for 12 months after the administration of DMBA. Hatched bar, duration of preventive treatment; filled bar, duration of curative treatment. *P < 0.05 vs controls; **P < 0.01 vs controls.
Figure 2 Incidence of mammary adenocarcinomas (ADKs) after daily administration of melatonin (10 mg/kg) for either 15 days before (preventive treatment; open diamonds) or 6 months after (curative treatment; filled diamonds) the administration of dimethylbenz [a]anthracene (DMBA). In both cases, the rats were followed up for 12 months after the administration of DMBA. Hatched bar, duration of preventive treatment; filled bar, duration of curative treatment. *P < 0.05 vs controls; **P < 0.01 vs controls.
Figure 3 Average number (mean ± standard error of the mean) of mammary adenocarcinomas (ADKs) per ADK-bearing rat after daily administration of melatonin (10 mg/kg) for either 15 days before (preventive treatment; open diamonds) or 6 months after (curative treatment; filled diamonds) the administration of dimethylbenz [a]anthracene (DMBA). In both cases, the rats were followed up for 12 months after the administration of DMBA. Hatched bar, duration of preventive treatment; filled bar, duration of curative treatment. *P < 0.05 vs controls; **P < 0.01 vs controls.
==== Refs
Reiter RJ The pineal gland and its hormones in the control of reproduction in mammals Endocr Rev 1980 1 109 131 6263600
Vanececk J Inhibitory effect of melatonin on GnRH-induced LH release Rev Reprod 1999 4 67 72 10357093 10.1530/ror.0.0040067
Anisimov VN The role of the pineal gland in breast cancer development Crit Rev Oncol Hematol 2003 46 221 234 12791421
Aubert CH Janiaud P Leclavez J Effect of pinealectomy and melatonin on mammary tumor growth in Sprague-Dawley rats under different conditions of lighting J Neural Transm 1980 47 121 130 6768843
Tamarkin L Cohen M Roselle D Reicher C Lippman M Chabner B Melatonin inhibition and pinealectomy enhancement of 7,12-dimethyl-benz(a)anthracene-induced mammary tumors in the rat Cancer Res 1981 41 4432 4436 6796259
Blask DE Hill SM Orstead KM Massa JS Inhibition effects of the pineal hormone melatonin and underfeeding during the promotional phase of 7,12 DMBA induced mammary tumorigenesis J Neural Transm 1986 67 125 138 3097252 10.1007/BF01243365
Blask DE Sauer L Dauchy R Holowachuk EW Ruhoff MS Kopff HS Melatonin inhibition of cancer growth in vivo involves suppression of tumor fatty acid metabolism via melatonin receptor-mediated signal transduction events Cancer Res 1999 59 4693 4701 10493527
Hill SM Blask DE Effects of the pineal hormone melatonin on the proliferation and morphological characteristics of human breast cancer cells (MCF-7) in culture Cancer Res 1988 48 6121 6126 3167858
Cos S Sanchez-Barcelo EJ Webb SM, Puig-Domingo M, Moller M, Pevet P Melatonin reduces the invasion capacity of MCF-7 human breast cancer cells. In vitro studies Pineal Update: From Molecular Mechanisms to Clinical Implications 1997 New York: PJD 377 383
Cos S Fernandez R Güezmes A Sanchez-Barcelo EJ Influence of melatonin on invasive and metastatic properties of MCF-7 human breast cancer cells Cancer Res 1998 58 4383 4390 9766668
Recio J Cardinali DP Sanchez-Barcelo EJ 2-[125I]iodomelatonin binding sites in murine mammary tissue Biol Signals 1994 3 85 90 7951651
Rato AG Pedrero JG Martinez MA del Rio B Lazo PS Ramos S Melatonin blocks the activation of estrogen receptor for DNA binding FASEB J 1999 13 857 868 10224229
Molis TM Spriggs LL Hill SM Modulation of estrogen receptor mRNA expression by melatonin in MCF-7 human breast cancer cells Mol Endocrinol 1994 8 1681 1690 7708056 10.1210/me.8.12.1681
Danforth DN Tamarkin L Lippman ME Melatonin increases oestrogen receptor binding activity of human breast cancer cells Nature 1983 305 323 325 6621687 10.1038/305323a0
Del Rio B Garcia Pedrero JM Martinez-Campa C Zuazua P Lazo PS Ramos S Melatonin, an endogenous-specific inhibitor of estrogen receptor α via calmodulin J Biol Chem 2004 279 38294 38302 15229223 10.1074/jbc.M403140200
Beau Yon De Jonage-Canonico M Lenoir V Martin A Scholler R Kerdelhué B Long term inhibition by estradiol of progesterone of melatonin secretion after administration of a mammary carcinogen, the dimethyl benz(a)anthracene, in Sprague-Dawley female rat; inhibitory effect of melatonin on mammary carcinogenesis Breast Cancer Res Treat 2003 79 365 377 12846421 10.1023/A:1024059824430
Huggins C Grand LC Brillantes FP Mammary cancer induced by a single feeding of polynuclear hydrocarbons and its suppression Nature 1961 189 204 207 13716610
Russo J Tav IK Russo IH Differentiation of the mammary gland and susceptibility to carcinogenesis Breast Cancer Res Treat 1982 2 5 73 6216933
Russo J Tait I Russo IH Susceptibility of the mammary gland to carcinogenesis. III. The cell of origin of rat mammary carcinogens Am J Pathol 1983 113 50 66 6312803
Daniel FB Jovco NJ DNA adduct formation by 7,12-dimethylbenz[a]anthracene and its noncarcinogenic 2-fluoro analog in the female Sprague-Dawley rats J Natl Cancer Inst 1983 70 111 118 6401826
Dao TL The role of ovarian hormones in initiating the induction of mammary cancer in rats by polynuclear hydrocarbons Cancer Res 1962 22 973 981 14025066
Kerdelhué B El Abed A Inhibition of preovulatory gonadotropin secretion and stimulation of prolactin secretion by 7,12-dimethylbenz[a]anthracene in Sprague-Dawley rats Cancer Res 1979 39 4700 4705 115578
El Abed A Kerdelhué B Castanier M Scholler R Stimulation of estradiol-17β secretion by 7,12-dimethylbenz[a]anthracene during mammary tumor induction in Sprague-Dawley rats J Steroid Biochem 1987 26 733 738 3112469 10.1016/0022-4731(87)91047-8
Jakubowski M Lenoir V Jimenez-Linan M Duval P Israel L Roberts JL Kerdelhué B Long-term effects of the mammary carcinogen 7,12-dimethylbenz[a]anthracene on hypothalamic gonadotropin-releasing hormone and Its pituitary receptor gene expression, during the promotion stage, in female Sprague-Dawley rats Breast Cancer Res Treat 2002 73 23 29 12083628 10.1023/A:1015282229388
Kerdelhué B Peck EJ In vitro LHRH release: correlation with the LH surge and alteration by a mammary carcinogen Peptides 1981 2 219 222 6794013 10.1016/0167-0115(81)90026-4
Chang C Topographic recognition of cyclic hydrocarbons and related compounds by receptors for androgens, estrogens and glucocorticoids J Steroid Biochem 1987 27 123 131 3695475 10.1016/0022-4731(87)90303-7
Garcia-Segura LM Bojda-Diolez F Lenoir V Naftolin F Kerdelhué B Estrogen-like effects of the mammary carcinogen 7,12-dimethylbenz(a)anthracene upon hypothalamic neuronal membranes Brain Res Bull 1992 28 625 628 1617445 10.1016/0361-9230(92)90113-C
Pasqualini C Sarrieau A Dussalliant M Corbani M Diolez-Bojda F Rostene W Kerdelhué B Estrogen-like effects of 7,12-dimethylbenz(a)anthracene on the female rat hypothalamo-pituitary axis J Steroid Biochem 1990 36 485 491 2145477 10.1016/0022-4731(90)90092-7
Cos S Sanchez-Barcelo EJ Melatonin and mammary pathological growth Front Neuroendocrinol 2000 21 133 170 10764528 10.1006/frne.1999.0194
Reiter RJ Tan DX Sainz RM Mayo JC Lopez-Burillo S Melatonin: reducing the toxicity and increasing the efficacy of drugs J Pharm Pharmacol 2002 54 1299 1321 12396291 10.1211/002235702760345374
Liebmann PM Wolfler A Felsner P Hofer D Schauenstein K Melatonin and the immune system Int Arch Allergy Immunol 1997 112 203 211 9066504
Maestroni GJ The immunoneuroendocrine role of melatonin J Pineal Res 1993 14 1 10 8483103
Hill SM Kiefer T Teplitzky S Spriggs LL Ram P Bartsch C, Bartsch H, Blask DE, Cardinali DP, Hushesky WJM, Mecke D Modulation of the estrogen response pathway in human breast cancer cells by melatonin The Pineal Gland and Cancer 2001 Berlin: Springer-Verlag 343 358
Blask DE Hill SM Effects of melatonin on cancer: studies on MCF-7 human breast cancer cells in culture J Neural Transm Suppl 1986 21 433 449 3462341
Blask DE Sauer LA Dauchy RT Melatonin as a chronobiotic/anticancer agent: cellular, biochemical, and molecular mechanism of action and their implications for circadian-based cancer therapy Curr Top Med Chem 2002 2 113 132 11899096 10.2174/1568026023394407
Todorovic R Ariese F Devanesan P Jankowiak R Small G Rogan E Cavalieri E Determination of benz(a)pyrene- and 7,12-dimethylbenz[a]anthracene DNA adducts formed in rat mammary gland Chem Res Toxicol 1997 10 941 947 9305574 10.1021/tx970003y
Garner R The role of DNA adducts in chemical carcinogenesis Mutat Research 1998 402 67 75
Subramanian A Kothari LS Suppressive effect by melatonin on different phases of 9,10-dimethyl-1,2-benzanthracene (DMBA)-induced rat mammary gland carcinogenesis Anti cancer Drugs 1991 2 297 303 1802026
Sanchez-Barcelo EJ Cos S Fernandez R Mediavella MD Melatonin and mammary cancer: a short review Endocrine Related Cancer 2003 10 153 159 12790777 10.1677/erc.0.0100153
Blask DE Pelletier DB Hill SM Lemus-Wilson A Grosso DS Wilson D Wise ME Pineal melatonin inhibition of tumor promotion in the N-nitrosomethylurea model of mammary carcinogenesis: potential involvement of antiestrogenic mechanisms in vivo J Cancer Res Clin Oncol 1991 117 526 532 1744157
Beau Yon De Jonage-Canonico M Lenoir V Scholler R Kerdelhué B Long term dysregulation of circadian and 17β estradiol-induced LH, prolactin and corticosterone secretion after dimethylbenz (a) anthracene administration in the Sprague-Dawley female rat Breast Cancer Res Treat
Fu Lee CC The circadian clock: pacemaker and tumor suppressor Nat Rev of cancer 2003 3 350 361 12724733 10.1038/nrc1072
| 15987452 | PMC1175060 | CC BY | 2021-01-04 16:54:50 | no | Breast Cancer Res. 2005 Apr 29; 7(4):R470-R476 | utf-8 | Breast Cancer Res | 2,005 | 10.1186/bcr1031 | oa_comm |
==== Front
Breast Cancer ResBreast Cancer Research1465-54111465-542XBioMed Central London bcr10321598745310.1186/bcr1032Research ArticleEffects of milk fermented by Lactobacillus helveticus R389 on a murine breast cancer model de Moreno de LeBlanc Alejandra [email protected] Chantal [email protected] Nicole [email protected]ón Gabriela [email protected] Départment de Chimie-Biochimie, Université de Moncton, NB, Canada2 Centro de Referencia para Lactobacilos (CERELA-CONICET), Tucumán, Argentina3 Cátedra de Inmunología, Facultad de Bioquimíca, Química y Farmacia, Universidad Nacional de Tucumán, Argentina2005 26 4 2005 7 4 R477 R486 30 11 2004 2 3 2005 21 3 2005 31 3 2005 Copyright © 2005 de Moreno de LeBlanc et al, licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is cited.
Introduction
Antitumour activity is one of the health-promoting effects attributed to the lactic acid bacteria and their products of fermentation. Previous studies in mice demonstrated that bioactive compounds released in milk fermented by Lactobacillus helveticus R389 contribute to its immunoenhancing and antitumour properties. The aim of the present work was to study the effects of the consumption of milk fermented by L. helveticus R389 or its proteolytic-deficient variant, L. helveticus L89, on a murine hormone-dependent breast cancer model.
Methods
Mice were fed with milk fermented by L. helveticus R389 or L. helveticus L89, during 2 or 7 days. The tumour control group received no special feeding. At the end of the feeding period, the mice were challenged by a subcutaneous injection of tumour cells in the mammary gland. Four days post-injection, the mice received fermented milk on a cyclical basis. The rate of tumour development and the cytokines in serum, mammary gland tissue and tumour-isolated cells were monitored. Bcl-2-positive cells in mammary glands and cellular apoptosis in tumour tissue were also studied.
Results
Seven days of cyclical administration of milk fermented by either bacterial strain delayed or stopped the tumour development. Cytokines demonstrated that L. helveticus R389 modulated the immune response challenged by the tumour. IL-10 and IL-4 were increased in all the samples from this group. In comparison with the tumour control, all test groups showed a decrease of IL-6, a cytokine involved in oestrogen synthesis. Seven days of cyclical feeding with milk fermented by L. helveticus R389 produced an increase in the number of apoptotic cells, compared with all other groups.
Conclusion
This study demonstrated that 7 days of cyclical administration of milk fermented by both strains of L. helveticus diminishes tumour growth, stimulating an antitumour immune response. Compounds released during milk fermentation with L. helveticus R389 would be implicated in its immunoregulatory capacity on the immune response in mammary glands and tumour, which were correlated with the cytokines found at the systemic level. The milk fermented by L. helveticus R389 was able to modulate the relationship between immune and endocrine systems (by IL-6 diminution), which is very important in oestrogen-dependent tumour and induced cellular apoptosis.
==== Body
Introduction
Considerable advances have been made in recent years towards an understanding of the molecular factors involved in breast cancer development, but for women in most Western countries breast cancer still remains a major cause of death. There are genetic and environmental factors that increase the chances of breast cancer, and the most common breast cancer types are oestrogen dependent. Some factors, such as diets rich in cultured dairy products, may inhibit the growth of many types of cancer, including breast tumours and the most investigated to date, colon cancer.
Live microbial feed supplements added to beneficially affect the host animal are known as probiotics [1]. Lactic acid bacteria (LAB) are the microorganisms most commonly used as probiotics to favour some biological functions in the host. LAB have been shown to exert effects on the immune system of the consumer and to increase the resistance to neoplasia and infections [2]. Consumption of LAB and milks fermented by them can increase the systemic immune response (macrophage function and number of immunoglobulin-secreting cells) [3,4] as well as increase the local immune responses in the mucosal areas (IgA-positive cells in the intestine, bronchus and mammary glands [5]). For these and other reasons, there is a steady increase in the consumption of fermented dairy products (i.e. yoghurt and other fermented milks) containing viable LAB.
Immunostimulation by fermented milks as a mean of keeping the host immune system in a permanent state of alert has been shown to successfully prevent different cancers [4,6,7]. Beneficial effects of fermented products in colon cancer prevention have been widely reported [8,9]. Studies carried out with an animal model of colon cancer showed inhibition of the tumour through yoghurt feeding, demonstrating that yoghurt modulated the immune system response and exerted its antitumour activity through its anti-inflammatory capacity [10,11]. This effect was observed by long-term cyclic yoghurt consumption, which inhibited promotion and progression of the experimental intestinal tumour [12].
In addition to LAB, fermented milks can possess other nonbacterial components produced during fermentation that contribute to immunogenicity and to other properties like their antitumour activities.
Matar and colleagues [13] have reported different roles and functions of biologically active peptides released from fermented milks. Peptides and free fatty acids released during fermentation were shown to increase the immune response. In this way, peptidic fractions liberated during milk fermentation with Lactobacillus helveticus R389 stimulated the immune system and inhibited the growth of an immunodependent fibrosarcoma in a mouse model [14]. The peptidic profiles of milk proteins were significantly different after fermentation by LAB, suggesting that microbial proteolysis could be a potential source of bioactive peptides [15]. Milk fermented with L. helveticus R389, a bacterium with high protease and peptidase activity, exerted an antimutagenic effect, while a mutant strain (L. helveticus L89) deficient in proteolytic activity did not [16]. In a similar way, milk fermented with the proteolytic strain increased the number of IgA-positive cells in the small intestine as well as in the bronchus of mice, but fermented milk obtained with the proteolytic-deficient mutant strain did not show the same in vivo results [6].
Fractions separated by dialysis from yoghurt showed tumour inhibition in in vivo murine assays [17]. Biffi and colleagues [18] studied the direct effect of milk fermented by five bacteria species (Bifidobacterium infantis, Bifidobacterium bifidum, Bifidobacterium animalis, Lactobacillus acidophilus, and Lactobacillus paracasei) on the growth of a breast cancer cell line, and reported that the antiproliferative effect was not related to the presence of bacteria in the fermented milk. This study suggested the potentiality offered by fermented milks as producers of compounds with antiproliferative activity useful in the prevention and therapy of solid tumours like breast cancer.
The aim of the present work was to study the effects of the consumption of milk fermented by L. helveticus R389 or its proteolytic deficient variant, L. helveticus L89, on a murine hormone-dependent breast cancer model, studying the systemic and local immune responses in the mammary glands and tumours.
Materials and methods
Animals and diets
BALB/c mice from Charles River Laboratories (Montreal, QC, Canada), weighing 19–21 g were separated into five experimental groups: the tumour control group, where the mice received an injection with the tumour cells; the P(+) 2d group, where the mice were fed with milk fermented by L. helveticus R389, proteolytic variant, for two consecutive days (basal 2 days), were injected with the tumour cells, and were then fed cyclically every 5 days with the fermented milk until day 28; the P(-) 2d group, which was the same as the P(+) 2d group except the mice were fed milk fermented with L. helveticus L89, the deficient proteolytic variant, instead of R389; the P(+) 7d group, where the mice were fed with milk fermented by L. helveticus R389 for seven consecutive days (basal 7 days), were injected with the tumour cells, and were then fed cyclically every 5 days, with the same fermented milk; and the P(-) 7d group, which was the same as the P(+) 7d group except that the mice were fed milk fermented with L. helveticus L89 instead of L. helveticus R389.
All groups contained 25–30 mice that were fed with a balanced diet ad libitum.
Milk fermentation
Nonfat, dried, low-heat-grade milk without added vitamins A and D (Dairytown Products Ltd, Sussex, NB, Canada) was rehydrated (12% wt/vol) and autoclaved (115°C for 15 min) to make the inoculums. The prepared milk was inoculated with L. helveticus R389 (2% vol/vol) and incubated at 37°C during 17 hours. Yeast extract (0.4%) was added to the milk used to grow L. helveticus L89 before autoclaving. This milk was then inoculated with L. helveticus L89 (2% vol/vol) and incubated at 37°C for 17 hours. Both fermented milks had a concentration of 1 × 109 colony-forming units/ml at the end of the fermentation period.
The inoculums were added to rehydrated milk prepared in the same manner (2% vol/vol) in 12% milk or 12% milk plus 0.4% yeast extract to start the milk fermentation.
The extent of milk protein proteolysis was evaluated using the o-phthaldialdehyde test [19].
Tumour induction and the feeding procedure
The ATCC tumoural cell line 4T1 was used to induce breast tumour growth. Each mouse was challenged by a single subcutaneous injection (0.5 ml) of tumour cells (1.4 × 104 cells/ml) in the upper right mammary gland.
The experimental groups – P(+) 2d, P(-) 2d, P(+) 7d and P(-) 7d – were given a diet supplemented with milks fermented by L. helveticus R389 or L. helveticus L89 for two or seven consecutive days. At the end of each feeding period the mice were injected with the tumour cells in the same way that the tumour control animals were. Four days after the tumour injection, fermented milks were added again to the diet during two or seven consecutive days (depending on the group), followed by a 5-day break, and then again fermented milk feeding for 2 or 7 days. Feeding was given in this manner cyclically until the end of the experimentation (28 days after tumour induction).
Obtaining the samples
The following samples were obtained from each group: basal sample (day 0), after 2 or 7 days of fermented milk feeding, and 12, 18, 22 or 28 days after tumour cell inoculation. Mice were anaesthetized intraperitoneally using a mix of ketamine hydrocholoride (Bioniche Animal Health Canada Inc, Ontario, Canada), 100 μg/g body weight, and xylazine hydrochloride (Sigma, St Louis, MO, USA), 5 μg/g body weight. Blood samples were obtained by cardiac punction. For the basal sample, and 12 days after tumour cells injection, mammary glands were removed. In the subsequent samples the tumour was also removed.
To obtain serum, blood was incubated at 37°C during 3 hours and was centrifuged at 1000 × g for 10 min. The serums were stored at -20°C until they were used for cytokine measurement.
ELISA assays of serum samples
To determine the concentration of the different cytokines (tumour necrosis factor alpha [TNF-α], interferon gamma [IFN-γ], IL-10, IL-4 and IL-6) in serum, the BD OptEIA™ mouse cytokine ELISA kits from BD Bioscience (San Diego, CA, USA) were used. The results are expressed as concentration of each cytokine in serum (pg/ml).
Cytokine-producing cell determination in histological sections
Mammary gland tissue sections (4 μm) from each group were used for immunofluorescence assays. Tissues were prepared for histological evaluation, were fixed in formaldehyde, were dehydrated using a graded series of ethanol and xylene substitute, and were embedded in paraffin.
Cytokines and Bcl-2-positive cells were detected by indirect immunofluorescence following the technique described by de Moreno de LeBlanc and colleagues [11]. Rabbit anti-mouse TNF-α, IFN-γ, IL-10, IL-6 and IL-4 (Peprotech, Inc., Rocky Hill, NJ, USA) polyclonal antibodies (diluted in saponin-PBS) were applied to the sections for 75 min at room temperature (21°C). The sections were then treated with a dilution of goat anti-rabbit antibody conjugated with fluorescein isothiocyanate (Jackson Immuno Research Labs Inc., West Grove, PA, USA).
Bcl-2 protein was measured using the same protocol with a diluted hamster anti-mouse Bcl-2 monoclonal antibody (PharMingen; Becton Dickinson Co., San Diego, CA, USA) and a dilution of the rabbit anti-Syrian hamster antibody conjugated with fluorescein isothiocyanate (Jackson Immuno Research Labs Inc.). The number of fluorescent cells was counted in 30 fields of vision as seen at 1000 × magnification using a fluorescence light microscope. The results were expressed as number of positive cells in 10 fields of vision as seen with 1000 × magnification using a fluorescence light microscope.
Isolation of mononuclear cells from the breast tumour
Tumours of three mice from each group were removed 18, 22 and 27 days after tumour inoculation and were washed with Hank's balanced saline solution (Sigma) with 4% foetal bovine serum. The cells were separated mechanically and incubated in 0.05% protease/collagenase (Sigma) solution in RPMI 1640 medium (Sigma) with added 10% foetal bovine serum at 37°C and agitated with a magnetic bar for 40 min. The cells collected from supernatant were washed with RPMI 1640 medium. The immune cells were concentrated using a percoll gradient (100% to 55% to 30%), were centrifuged at 800 × g for 30 min, and were recovered from the layer between 100% and 55%. Cells were adjusted at 4 × 106 to 5 × 106 cells/ml in RPMI 1640 medium. Cell suspensions (20 μl) were placed in each well of an immunofluorescence slide and were fixed with formalin (ICC fixation buffer, PharMingen; Becton Dickinson Biosciences, San Diego, CA, USA).
Cytokine determination in isolated cells
TNF-α, IL-4, IL-10, IL-6 or IFN-γ were determined in the fixed cells. They were incubated with 1% blocking solution of bovine serum albumin/PBS, were washed with PBS and were incubated with normal goat serum (diluted 1/10). The activity of the endogenous peroxidase was blocked with H2O2/methanol solution. The cells were then incubated with avidin-blocking and biotin-blocking solutions (avidin/biotin blocking kit; Vector Labs, Inc., Burlingame, CA, USA) to block endogenous avidin and biotin. The cells were incubated with rat anti-mouse TNF-α, IFN-γ, IL-10 or IL-4 polyclonal antibody (diluted in diluent ICC cytokine buffer; PharMingen, Becton Dickinson Biosciences), were washed with PBS, and were incubated with a biotin-conjugated goat anti-rat immunoglobulin-specific polyclonal antibody (PharMingen; Becton Dickinson Biosciences). Biotinylated anti-mouse IL-6 polyclonal antibody (PharMingen; Becton Dickinson Biosciences) was used to determine IL-6-positive cells. Vectastain Elite ABC solution (Vector Labs) was added to cells and they were incubated with a DAB kit (Vector Labs). The results were expressed as percentage (number of positive cells in 100 cells counted at 1000 × magnification).
Apoptosis determination
Apoptosis was evaluated by the presence of DNA breaks, detected in the paraffin cuts using the Apoptosis Detection System kit, Fluorescein (Promega, Madison, WI, USA). The fragmented DNA of apoptotic cells was measured by incorporation of fluorescein-12-dUTP at the 3'-OH ends of DNA using terminal deoxynucleotidyl transferase, which forms a polymeric tail via the principle of the Terminal deoxynucleotidyl Transferase Biotin-dUTP Nick End Labeling (TUNEL) assay. The fluorescein-12-dUTP nick end-labelled DNA was visualized directly by fluorescence microscopy. Cells were defined as apoptotic if the whole nuclear area of the cell was stained fluorescent.
Apoptosis was expressed as the number of apoptotic cells 10 ten fields with 1000 × magnification, using a fluorescence microscope with a standard fluorescent filter.
Statistical analysis
Comparisons were performed using the software package SigmaStat (SPSS, Chicago, IL, USA).
Comparisons of multiple means were accomplished by one-way analysis of variance followed by a 150 Tukey's post-hoc test. P < 0.05 was considered significant. Unless otherwise indicated, all values are the means of three independent trials ± standard deviation.
Results and discussion
The intestine is the first area of study to assay different properties of probiotics that enter the host by the oral route. This host contains the 'common' mucosal immune system, which ensures that all mucous membranes are furnished with a wide spectrum of secretory antibodies [20]. Both B cells and T cells can migrate from Peyer's patches, found in the small intestine, to mucosal membranes of the respiratory, gastrointestinal and genitourinary tract, as well as to exocrine glands such as the lacrimal glands, salivary glands, mammary glands and prostatic glands [21].
Lactobacillus casei CRL 431 administered orally was able to stimulate the IgA cycle, increasing IgA-positive cells not only in the intestine, but also in the bronchus and mammary gland tissues [5].
Previous studies performed in our laboratory showed different effects when mice were fed with milk fermented by L. helveticus R389 or its proteolytic deficient variant, L. helveticus L89 [6]. Mice fed with the L. helveticus R389 showed stimulation of the mucosal immune system. Also, it was observed that 2 and 7 days of feeding with this bacterial strain were optimal to stimulate the mucosal system of mice (data not shown). This strain was also able to prevent the growth of transplantable fibrosarcoma in mice [6]. These previous works led us to study the effect of L. helveticus R389 on the growth of breast cancer in an animal model, comparatively with the proteolytic-deficient strain L. helveticus L89.
Tumour growth
In the present work, mice were fed cyclically following a previous model of feeding with yoghurt to inhibit a colon tumour in mice, in which yoghurt feeding showed a modulation of the immune response in intestine [11]. Here we showed that mice receiving a 2-day cyclical fermented milk feeding did not show significant differences in tumour volume, compared with the tumour control group (Fig 1). Seven-day cyclical administration of both bacterial strains delayed or stopped tumour development, as compared with the control group (Fig. 1). There were no significant differences between both bacterial strains used in milk fermentation cyclical feeding for either 2 or 7 days (Fig. 1). Either the LAB themselves or some substances released during milk fermentation were responsible for this observed effect: mice fed only with milk (or with milk plus 0.4% yeast extract) did not show differences in the tumour size compared with the control (data not shown)
Determination of cytokine levels in blood serum
The influence of immune cells in breast cancer development has been reported in different models, but to our knowledge no published reports have studied the in vivo immunomodulatory effects of LAB and their relationship with mammary glands or breast cancer. Cytokines have been shown to regulate oestrogen synthesis in breast tumours, stimulating research into these important molecules. In this work, cytokines were assayed in different samples to have a spectrum at a systemic level, and to measure the local response in mammary glands or tumours to study the effect of our fermented milks on the immune response.
TNF-α is a cytokine with various functions such as proinflammatory pathway properties, tumour necrosis pathway properties, and apoptosis pathway properties [22,23]. TNF-α levels increased in the serum as a function of time as did the tumour volume in the control group (Fig. 2a). Mice from both 2-day groups showed increases of this cytokine in serum at day 12 (234 ± 14 pg/ml, 206 ± 38 pg/ml and 207 ± 24 pg/ml, for the P(+) 2d, P(-) 2d and tumour control groups, respectively). TNF-α levels remained constant after day 12. Mice that received 7 days of cyclical feeding with milk fermented with L. helveticus R389 or L. helveticus L89 showed a significant increase (P < 0.05) of TNF-α in the basal sample (207 ± 43 pg/ml and 256 ± 51 pg/ml for the P(+) 7d group and the P(-) 7d group, respectively), compared with the tumour control group (42 ± 2 pg/ml). This increase before tumour induction could be related to the decrease in the tumour growth in the mice from these groups. The P(+) 7d group maintained a TNF-α concentration near the basal level throughout the trial, showing a regulation of this cytokine, whereas the P(-) 7d group showed increased TNF-α in the final sample, similar to the control group (530 ± 71 pg/ml and 603 ± 106 pg/ml for 28 days, in the P(-) 7d group and control group, respectively). These results (the TNF-α increase) showed a typical immune response to the tumour [24].
IFN-γ is a cytokine related to the inflammatory response, but it was also reported as a key effector molecule in the immune response against solid cancers. Tumour infiltrating lymphocytes from ovarian tumours released this cytokine upon challenge with MICA-positive tumour cells [25]. In our study, IFN-γ levels varied in the different groups as a function of time (Fig. 2b).
IL-6 is a cytokine implicated in oestrogen synthesis [26], a hormone that the tumour needs for growth. It is also a proangiogenic factor [27], supporting the growth of new blood vessels that are essential for tumour growth. The three groups in which the tumour grew at a faster rate showed elevated levels of IL-6 (Fig. 2c). The P(+) 7d and P(-) 7d groups did not show increased levels of this cytokine throughout the time of the study, suggesting that this IL-6 decrease is involved in one of the mechanisms for the delay of tumour growth.
IL-10 and IL-4 are known as regulatory cytokines, associated with activated Th2 lymphocytes [22]; IL-10 can also be produced by other cell populations such as macrophages and dendritic cells. In different experimental models, TNF-α and IL-10 were demonstrated to have opposite effects [28]. The balance between TNF-α and IL-10 could modulate the effector function of macrophages and cellular apoptosis. IL-4 plays a significant role in controlling both cell growth and modulation of the immune response [29]. This cytokine has antagonist functions to IFN-γ and appears to possess certain anti-inflammatory properties; IL-4 can inhibit the production of several proinflammatory cytokines such as IL-1, IL-6, IL-8, and TNF-α [22]. There was a significant increase (P < 0.05) in IL-10 and IL-4 concentrations in the serum obtained from the P(+) 7d group in relation to the tumour control group, beginning at day 18 after tumour injection (Fig. 2e), which could explain the regulation of the immune response observed for TNF-α and IL-6 in these animals. This regulatory response was not observed in the P(-) 7d group. Mice from the P(+) 2d group showed increases for IL-10 concentrations in serum compared with the control, but the levels obtained were significantly lower (P < 0.05) than the P(+) 7d group. The P(-) 2d group showed increases of IL-10 compared with the tumour control group, but not compared with the other groups described. These results concur with the development of the tumour observed in these groups of mice.
IL-4 increased in all test groups in the basal sample, in comparison with the control group (18 ± 1 pg/ml). The P(+) 7d group showed the largest increases throughout the study (Fig. 2d).
Study of cytokine-positive cells in mammary gland tissues
Differences were observed with regard to the systemic levels of cytokines between the groups where the tumour grew and those where it did not; it was possible to observe a regulation of the immune response in the P(+) 7d group, but not in the P(-) 7d group.
The study of cytokine-positive cells in mammary glands allowed an understanding of the local cell response, after mice were fed with fermented milk as well as after tumour injection, in the tumour control group and different test groups.
TNF-α-positive cells in mammary glands showed very similar patterns to those obtained for this cytokine in serum. TNF-α-positive cells increased in the tumour control group throughout the tumour growth. The same observation was seen in the P(+) 2d and P(-) 2d groups (Fig. 3a). This cytokine increased in the P(+) 7d and P(-) 7d groups 12 days after tumour injection (13 ± 5 cells / 10 fields and 18 ± 5 cells / 10 fields, respectively), and the number of positive cells remained constant afterwards (Fig. 3a).
IFN-γ-positive cells increased in the tumour control group throughout the trial. The P(+) 2d and P(-) 2d groups showed similar values to the control group at the end of the experimental period (27 ± 3 cells / 10 fields, 22 ± 4 cells / 10 fields and 24 ± 5 cells / 10 fields for the P(+) 2d, P(-) 2d and control groups, respectively). Seven days of cyclical feeding, independent of the bacterial strain used, maintained the number of positive cells for this cytokine; however, a significant decrease was observed in the final samples compared with the other groups (14 ± 3 cells / 10 fields and 13 ± 4 cells / 10 fields for the P(+) 7d group and the P(-) 7d group, respectively) (Fig. 3b).
The IL-6-positive cell number was constant and similar in all groups until 18 days after tumour injection. This observation can be explained because this cytokine is related to the synthesis of oestrogen in the mammary gland, a hormone that this tumour cell line needs for proper growth. Eighteen days after tumour cell injection, these cytokine-positive cells increased in the control, P(+) 2d and P(-) 2d groups, whereas both the P(+) 7d and P(-) 7d groups showed no differences in the number of IL-6-positive cells, showing significantly lower numbers in relation to the other groups (Fig. 3c).
IL-10-positive cells increased in the control group after 12 days of tumour injection (11 ± 3 cells / 10 fields) compared with the basal number (6 ± 2 cells / 10 fields), and remained constant throughout the study (Fig. 3e). Mice fed with L. helveticus R389 had increased numbers of IL-10-positive cells throughout the time of the entire study, but only the P(+) 7d group showed significantly higher numbers (P < 0.05) compared with the tumour control group on days 18 and 22 (25 ± 5 cells / 10 fields and 22 ± 3 cells / 10 fields at 18 days, and 14 ± 6 cells / 10 fields and 11 ± 3 cells / 10 fields at 22 days for the P(+) 7d group and the control group, respectively; see Fig. 3e). This cytokine can be related to the regulation of the other cytokines observed in this group, where increases in TNF-α-positive and IFN-γ-positive cells in mammary glands were observed. This was not the case with the P(-) 7d group, which did not show a significant increase of IL-10-positive cells compared with the tumour control group.
The number of IL-4-positive cells followed the same pattern as the IL-10-positive cells (Fig. 3d): significant increases (P < 0.05) were observed in the final samples of the P(+) 2d and P(+) 7d groups (21 ± 5 cells / 10 fields, 22 ± 4 cells / 10 fields and 15 ± 5 cells / 10 fields for the P(+) 2d, P(+) 7d and control groups, respectively) compared with the control.
Determination of cytokines in tumour-infiltrative cells
Breast tumour tissue contains malignant epithelial cells, stromal cells, adipocytes, lymphocytes and macrophages. The role of tumour-infiltrating immune cells in antitumour immunity, as well as their potential for cancer immunotherapy, has been investigated extensively [30,31]. Lymphocytes and macrophages invade the tumour in response to cytokines such as IL-8 and macrophage chemoattractant protein 1. These lymphocytes and macrophages produce IL-6, IL-6 soluble receptor, and TNF-α. IL-6 is also produced by stromal cells, and TNF-α is also produced by adipocytes. IL-6 is able to stimulate activity of aromatase, an enzyme related to the synthesis of oestrogen from androgens in malignant cells and stromal cells. IL-6 acts through its receptor on malignant cells.
TNF-α showed differences in the isolated cells compared with the other samples (serum and mammary glands). It has previously been reported that cytokines produced by tumour-infiltrating immune cells play an important role in the antitumour response [32]. The positive cells for this cytokine increased in the groups fed with fermented milk where the tumour growth was delayed (the P(+) 7d and P(-) 7d groups), showing that an induction of the production of this cytokine by fermented milk may be playing a biological role in the induction of cellular apoptosis. TNF-α-positive cells increased in the P(+) 7d and P(-) 7d groups 22 days after tumour injection (31 ± 4 cells / 100 and 47 ± 10 cells / 100), but on the last sample (day 27) the number of positive cells did not differ significantly with the control group and the 2-day groups. The same observation was shown with the IFN-γ-positive cells (Table 1).
IL-10 is a regulatory cytokine that can be released by tumour-infiltrating immune cells such as macrophages and lymphocytes. IL-10(+) cell numbers decreased in both the control and the P(+) 2d groups throughout the time of the entire study. The IL-10(+) cells increased significantly (P < 0.05) in tumours from P(-) 2d and P(-) 7d groups at 22 days after tumour injection, as compared to the control (18 ± 2 cells / 100, 16 ± 6 cells / 100 and 6 ± 2 cells / 100 for P(-) 2d, P(-) 7d and control, respectively). P(+) 7d was the group with the highest number of IL-10(+) cells (29 ± 7 cells / 100) at 22 days. It is possible to observe another increases of these cells in mice from P(+) 7d group, such as were observed in the other samples (serum and mammary gland tissues), probably to regulate the proinflammatory cytokines (TNF-α and IFN-γ) produced. IL-4-positive cells were variable in all groups, and no significant increases were observed. All mice fed with fermented milk showed decreases in the number of IL-6 (+) cells compared to the tumour control group. This shows a protective effect of the LAB in this oestrogen-dependent tumour, due mainly to the decrease in IL-6. Inter-group differences were observed. After 18 days, P(-) 7d increased these positive cells, and the values were similar to P(+) 2d and P(-) 2d. Only P(+) 7d maintained the number of IL-6 (+) cells significantly lower than the other test groups in all the samples, showing once again the best antitumour response.
Apoptosis and Bcl-2-positive cell determination
The mechanisms of apoptosis (or programmed cell death) in the inhibition of tumour progression are well documented [33]. Apoptosis is a complex and active cellular process in which individual cells are triggered to undergo self-destruction in a manner that will neither injure neighbouring cells nor elicit an inflammatory reaction. The balance between cell proliferation and cell death is important to maintain equilibrium in different tissues, and a disturbance in this balance may lead to tumour development [34] since the disruption of this type of regulation is a characteristic of tumours.
Considering that cytokines such as TNF-α could be involved in certain apoptotic pathways [23], and that an enhancement of this cytokine was observed in our experimental model, apoptosis induction was studied (Table 2). The P(+) 2d and P(-) 2d groups showed a significant increase (P < 0.05) in the number of apoptotic cells at day 18, in relation to the control and other test groups. In mice from the P(+) 7d group, a significant increase was observed in the number of apoptotic cells (P < 0.05), compared with all the other groups, beginning at day 22. The increase in cellular apoptosis in the mice of the P(+) 7d group could play a role in the delay of the tumour growth.
Bcl-2 protein is a measure of cell survival due to its anti-apoptotic activity [23], which can be used to stimulate the growth of tumour cells. The increase of cellular apoptosis in mice from the group fed with fermented milk led us to study the Bcl-2 protein.
Significant differences between the groups were not observed when Bcl-2-positive cells were studied in tumour tissues (data not shown), but differences were seen in mammary gland tissues. Bcl-2-positive cells increased significantly (P < 0.05) in the final sample for the control group as compared with the basal sample for that group (33 ± 5 cells / 10 fields and 22 ± 4 cells / 10 fields at day 27 and the basal sample, respectively; Fig. 3f). The P(+) 7d and P(-) 7d groups showed significant decreases of Bcl-2-positive cells in mammary glands compared with the tumour control in the same period. These decreases were significant for both groups in the final sample (16 ± 3 cells / 10 fields, 16 ± 4 cells / 10 fields and 33 ± 5 cells / 10 fields for the P(+) 7d, P(-) 7d and control groups, respectively; Fig. 3f) and concur with the apoptosis results.
Conclusion
Numerous mechanisms in which the immune system plays a role can be involved in the antitumour activity of fermented milks, which could be mediated by the bioactive substances released during fermentation and by the microorganisms used as starter cultures.
Our studies using a model of breast cancer in mice demonstrated that 7 days of cyclical feeding with milk fermented by L. helveticus R389 or L. helveticus L89 delayed tumour development. This effect was related principally to a decrease in IL-6, a cytokine implicated in the synthesis of oestrogen in both normal and tumour-invaded breast. L. helveticus R389, a strain with high proteolytic activity, has been selected for future studies because of its capacity to modulate the immune response. Milk fermented by this LAB induced not only a decrease of IL-6, but also an increase of regulatory cytokines, principally IL-10, and also induced cell apoptosis in the tumour. This observation allows us suggest that substances released in this fermented milk, possibly peptides due to the high proteolytic activity of the bacterial strain, could be related to the regulatory response observed with this fermented milk, which was not observed with the milk fermented by L. helveticus L89 (proteolytic-deficient variant).
This is the first report of an in vivo study demonstrating the possible mechanisms by which LAB and fermented milks can influence the activity of the infiltrative immune cells in mammary glands, and also to delay or even stop a breast tumour. This study has demonstrated the immunoregulatory capacity of milk fermented by L. helveticus R389 on the immune response in mammary glands and tumour, as well as the correlations with the cytokines found at a systemic level. The milk fermented by L. helveticus R389 was able to delay tumour growth by its immunoregulatory capacity, and we demonstrated that this fermented milk was able to modulate the relationship between immune and endocrine systems by IL-6 decrease, which is very important in oestrogen-dependent tumours, and by induction of cellular apoptosis.
Research is currently in progress into the antitumour/immunomodulating properties of substances released during milk fermentation by this highly proteolytic LAB, and their effects on breast cancer development.
Abbreviations
ELISA = enzyme-linked immunosorbent assay; IFN-γ = interferon gamma; IL = interleukin; LAB = lactic acid bacteria; PBS = phosphate-buffered saline; TNF-α = tumour necrosis factor alpha.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
AML carried out the study design, animal feeding, data collection (ELISA, immunohistochemistry, immunocytochemistry), statistical analysis, data interpretation, manuscript preparation, and literature search. CM participated in the design of the study, in the data interpretation and manuscript preparation, and in funding the collection. NL participated in the animal feeding, data collection (ELISA, immunohistochemistry, immunocytochemistry), data interpretation, and manuscript preparation. GP participated in the design and coordination of the study, interpretation of data, manuscript preparation, and funding the collection. All authors read and approved the final manuscript.
Acknowledgements
The authors thank Mr Jairo Duarte for his help with animal care and sampling. This work was financially supported by the Atlantic Innovation Fund, by the Natural Sciences and Engineering Council of Canada, and by Consejo Nacional de Investigaciones Científicas y Técnicas, Argentina. All experiments comply with the current laws of Canada. All protocols were approved by the Animal Protection Committee of the Université de Moncton and followed the Guide for the Care and Use of Laboratory Animals of the National Institute of Health.
Figures and Tables
Figure 1 Rate of tumour growth. Results are expressed as the volume (cm3) of the tumour for each mouse of different groups (tumour control group, black circle and line; P(+) 2d group, grey square and dark grey line; P(-) 2d group, white triangle and black broken line; P(+) 7d group, black diamond and lines of points; P(-) 7d group, grey triangle and light grey broken lines) with the tendency line for each group.
Figure 2 Effect of tumour injection and fermented milk feeding on the serum cytokines. (a) Tumour necrosis factor alpha (TNFα), (b) interferon gamma (IFNγ), (c) IL-6, (d) IL-4 and (e) IL-10. Results are expressed as the mean concentration of each cytokine (pg/ml) ± standard deviation. Means for each cytokine without a common letter differ significantly (P < 0.05).
Figure 3 Cytokine-positive cells in mammary glands. Positive cells for each cytokine were counted in histological sections from mammary glands of the tumour control group (black bars), the P(+) 2d group (white bars), the P(-) 2d group (diagonal lined bars), the P(+) 7d group (grey bars) and the P(-)7d group (horizontal lined bars). (a) Tumour necrosis factor alpha (TNFα), (b) interferon gamma (IFNγ), (c) IL-6, (d) IL-4, (e) IL-10 and (f) Bcl-2. Values are means ± standard deviation for n = 5. Means for each cytokine without a common letter differ significantly (P < 0.05).
Table 1 Cytokine-positive cells isolated from the tumour
Experimental group Sample (days) Cytokine
Tumour necrosis factor alpha Interferon gamma IL-6 IL-4 IL-10
Tumour control 18 24 ± 2ae 22 ± 6ac 29 ± 2a 25 ± 5a 18 ± 2ae
22 12 ± 4bcd 17 ± 3ad 25 ± 3a 17 ± 4ad 6 ± 2b
27 13 ± 1bcd 12 ± 4b 23 ± 4abd 10 ± 3ce 8 ± 2b
P(+) 2d 18 21 ± 3ce 26 ± 8cde 12 ± 3d 11 ± 1bcde 19 ± 4a
22 9 ± 1d 10 ± 2b 8 ± 1cd 12 ± 2bcde 7 ± 3bc
27 11 ± 3bd 12 ± 4bd 17 ± 3d 10 ± 1ce 9 ± 2bc
P(-) 2d 18 14 ± 6bcd 14 ± 5abd 10 ± 3c 7 ± 2e 9 ± 4bc
22 19 ± 4bce 13 ± 4ab 16 ± 1bd 13 ± 5bcde 18 ± 6ae
27 12 ± 2bd 12 ± 2b 17 ± 2bd 7 ± 2e 8 ± 2bd
P(+) 7d 18 22 ± 7ac 20 ± 2ae 4 ± 2c 15 ± 5dbc 14 ± 3ace
22 31 ± 4a 30 ± 5c 9 ± 2c 8 ± 1e 29 ± 7
27 13 ± 1bcd 17 ± 5abe 4 ± 1c 6 ± 1e 12 ± 2eb
P(-) 7d 18 12 ± 2bcd 14 ± 1b 4 ± 1c 12 ± 3bcde 13 ± 6acde
22 53 ± 15 44 ± 6 17 ± 3bd 12 ± 3bcde 16 ± 4ace
27 11 ± 4bd 24 ± 5ac 13 ± 4d 9 ± 1ce 9 ± 3bcd
Results are expressed as means ± standard deviation of cytokine-positive cells for each 100 counted cells (cells/100). Means for each cytokine without a common letter differ significantly (P < 0.05).
Table 2 Study of the cellular apoptosis in the tumour tissue
Experimental group Sample
18 days 22 days 27 days
Tumour control 8 ± 2a 17 ± 5ab 13 ± 4ac
P(+) 2d 27 ± 6bd 38 ± 5d 19 ± 6bc
P(-) 2d 34 ± 6d 10 ± 2ac 8 ± 2a
P(+) 7d 8 ± 2a 57 ± 7e 61 ± 10e
P(-) 7d 9 ± 2a 20 ± 3bd 11 ± 3a
Results are expressed as the mean ± standard deviation of the number of apoptotic cells counted in 10 fields at 100 × magnification (cells/10 fields). Means for each cytokine (IL-10 or IL-6) without a common letter differ significantly (P < 0.05).
==== Refs
Fuller R Probiotics in man and animals J Appl Bacteriol 1989 66 365 378 2666378
Kato I Fuller R, Perdigón G Antitumor activity of lactic acid bacteria Probiotics 3: Immunomodulation by the Gut Microflora and Probiotics 2000 London: Kluwer Academic Publishers 115 138
Perdigón G Vintini E Alvarez S Medina M Medici M Study of the possible mechanisms involved in the mucosal immune system activation by lactic acid bacteria J Dairy Sci 1999 82 1108 1114 10386296
Perdigón G Fuller R Raya R Lactic acid bacteria and their effect on the immune system Curr Issues Intest Microbiol 2001 2 27 42 11709854
de Moreno de LeBlanc A Maldonado Galdeano C Chaves S Perdigón G Oral administration of L. casei CRL 431 increases immunity in bronchus and mammary glands European J Inflamm 2005 3 1 23 28
Matar C Valdez JC Medina M Rachid M Perdigón G Immunomodulating effects of milks fermented by Lactobacillus helveticus and its non-proteolytic variant J Dairy Res 2001 68 601 609 11928956 10.1017/S0022029901005143
Valdéz JC Rachid M Bru E Perdigón G The effect of yoghurt on the cytotoxic and phagocytic activity of macrophages in tumour-bearing mice Food Agric Immunol 1997 9 299 308
Wollowski I Rechkemmer G Pool-Zobel BL Protective role of probiotics and prebiotics in colon cancer Am J Clin Nutr 2001 73 451S 455S 11157356
Brady LJ Gallaher DD Busta FF The role of probiotic cultures in the prevention of colon cancer J Nutr 2000 130 410S 414S 10721916
Perdigón G de Moreno de LeBlanc A Valdez J Rachid M Role of yoghurt in the prevention of colon cancer Eur J Clin Nutr 2002 56 Suppl 3 S65 S68 12142967 10.1038/sj.ejcn.1601490
de Moreno de LeBlanc A Valdéz J Perdigón G Regulatory effect of yoghurt on intestinal inflammatory immune response Eur J Inflamm 2004 2 21 61
de Moreno de LeBlanc A Perdigón G Yoghurt feeding inhibits promotion and progression of experimental colorectal cancer Med Sci Monit 2004 10 Br96 Br104 15039638
Matar C LeBlanc JG Martin L Perdigón G Farnworth T Biologically active peptides released from fermented milk: role and functions Handbook of Fermented Functional Foods 2003 Boca Raton, FL: CRC Press 177 201
LeBlanc JG Matar C Valdez JC LeBlanc J Perdigón G Immunomodulatory effects of peptidic fractions issued from milk fermented with Lactobacillus helveticus J Dairy Sci 2002 85 2733 2742 12487440
Matar C Amiot J Savoie L Goulet J The effect of milk fermentation by Lactobacillus helveticus on the release of peptides during in vitro digestion J Dairy Sci 1996 79 971 979 8827460
Matar C Nadathur SS Bakalinsky AT Goulet J Antimutagenic effects of milk fermented by Lactobacillus helveticus and its non-proteolytic variant J Dairy Sci 1997 80 1965 1970 9313136
Ayebo AD Shahani KM Dam R Friend BA Ion exchange separation of the antitumour component(s) of yogurt dialyzate J Dairy Sci 1982 65 2388 2390 7161436
Biffi A Coradini D Larsen R Riva L Di Fronzo G Antiproliferative effect of fermented milk on the growth of a human breast cancer cell line Nutr Cancer 1997 28 93 99 9200156
Church FC Swaisgood HE Porter DH Catignani GL Spectrophotometric assay using o-phtaldialdehyde for determination of proteolysis in milk and isolated milk proteins J Dairy Sci 1983 66 1219 1227
Brandtzaeg P Baekkevold ES Farstad IN Jahnsen FL Johansen FE Nilsen EM Yamanaka T Regional specialization in the mucosal immune system: what happens in the microcompartments? Immunol Today 1999 20 141 151 10203706 10.1016/S0167-5699(98)01413-3
Brandtzaeg P Pabst R Let's go mucosal: communication on slippery ground Trends Immunol 2004 25 570 577 15489184 10.1016/j.it.2004.09.005
Feghali CA Wright TM Cytokines in acute and chronic inflammation Front Biosci 1997 2 d12 d26 9159205
Sellers WR Fisher DE Apoptosis and cancer drug targeting J Clin Invest 1999 104 1655 1661 10606616
Urban JL Shepard HM Rothstein JL Sugarman BJ Tumour necrosis factor: a potent effector molecule for tumour cell killing by activated macrophages Proc Natl Acad Sci USA 1986 83 5233 5237 3487788
Groh V Rhinehart R Secrist H Bauer S Grabstein KH Broad tumor-associated expression and recognition by tumor-derived-γδ T cells of MICA and MICB Proc Natl Acad Sci USA 1999 96 6879 6884 10359807 10.1073/pnas.96.12.6879
Purohit A Newman SP Reed MJ The role of cytokines in regulating estrogen synthesis: implications for the etiology of breast cancer Breast Cancer Res 2002 4 65 69 11879566 10.1186/bcr425
Benny M Itin A Sachs L Keshet E Pattern of interleukin 6 gene expression in vivo suggests a role for this cytokine in angiogenesis Proc Natl Acad Sci USA 1990 87 3092 3096 1691500
Bogdab C Park J Yodovotz Y Nathan C Contrasting mechanisms for suppression of macrophage cytokine release by transforming growth factor and interleukin-10 J Biol Chem 1992 267 23301 23309 1429677
Chang TH Peng X Fu X Interleukin-4 mediates cell growth inhibition through activation of Stat1 J Biol Chem 2000 275 1012 1017
Reome JB Hylind JC Dutton RW Dobrzanski MJ Type 1 and type 2 tumour infiltrating effector cell subpopulations in progressive breast cancer Clin Immunol 2004 111 69 81 15093554 10.1016/j.clim.2003.11.013
Ferrarini M Ferrero E Dagna LPA Zocchi MR Human-γδ T cells: a nonredundant system in the immune-surveillance against cancer Trends Immunol 2002 23 14 18 11801449 10.1016/S1471-4906(01)02110-X
Bingle L Brown NJ Lewis CE The role of tumour-associated macrophages in tumour progression: implications for new anticancer therapies J Pathol 2002 196 254 265 11857487 10.1002/path.1027
Butler LM Hewett PJ Fitridge RA Cowled PA Deregulation of apoptosis in colorectal carcinoma: theoretical and therapeutic implications Aust N Z J Surg 1999 69 88 94 10030808 10.1046/j.1440-1622.1999.01498.x
Hao X Du M Bishop A Talbot I Imbalance between proliferation and apoptosis in the development of colorectal carcinoma Virchows Arch 1998 433 523 527 9870685 10.1007/s004280050284
| 15987453 | PMC1175061 | CC BY | 2021-01-04 16:54:50 | no | Breast Cancer Res. 2005 Apr 26; 7(4):R477-R486 | utf-8 | Breast Cancer Res | 2,005 | 10.1186/bcr1032 | oa_comm |
==== Front
Breast Cancer ResBreast Cancer Research1465-54111465-542XBioMed Central London bcr10331598745710.1186/bcr1033Research ArticleGenetic polymorphisms in the matrix metalloproteinase 12 gene (MMP12) and breast cancer risk and survival: the Shanghai Breast Cancer Study Shin Aesun [email protected] Qiuyin [email protected] Xiao-Ou [email protected] Yu-Tang [email protected] Wei [email protected] Department of Medicine and Vanderbilt-Ingram Cancer Center, Vanderbilt University School of Medicine, Nashville, TN, USA2 Department of Epidemiology, Shanghai Cancer Institute, Shanghai, China2005 10 5 2005 7 4 R506 R512 19 1 2005 25 2 2005 31 3 2005 4 4 2005 Copyright © 2005 Shin et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Introduction
Matrix metalloproteinase 12 (MMP12) is a proteolytic enzyme responsible for cleavage of plasminogen to angiotensin, which has an angiostatic effect. Using data from a population-based case–control study conducted among Chinese women in Shanghai, we evaluated the association of breast cancer risk and survival with two common polymorphisms in the MMP12 gene: A-82G in the promoter region and A1082G in exon, resulting in an amino acid change of asparagine to serine.
Methods
Included in the study were 1,129 cases and 1,229 age-frequency-matched population controls. Breast cancer patients were followed up to determine the intervals of overall survival and disease-free survival.
Results
The frequencies of the G allele in the A-82G and A1082G polymorphism among controls were 0.029 and 0.107, respectively. There were no associations between MMP12 polymorphisms and breast cancer risk. Patients with the AG or GG genotype of the A1082G polymorphism showed poorer overall survival (though the difference was not statistically significant) than patients with the AA genotype (hazard ratio 1.36, 95% CI 0.92 to 2.00).
Conclusion
This result suggests that MMP12 A1082G polymorphism may be related to prognosis in breast cancer patients. Additional studies with larger sample sizes are warranted.
==== Body
Introduction
Matrix metalloproteinases (MMPs) are a family of zinc-dependent proteolytic enzymes that are involved in tumor angiogenesis, migration, and invasion as well as the regulation of immune surveillance [1,2]. With a few exceptions, the expression and activity of MMPs are increased in almost every type of human cancer and are correlated with advanced tumor stage, increased invasion and metastasis, and shortened survival [2,3]. In transplantation assays, relatively benign cancer cells acquire malignant properties when the expression of certain MMPs is up-regulated. Conversely, highly malignant cells become less aggressive when the expression or activity of certain MMPs is reduced [2].
Substrates of MMP12 are various extracellular matrix and non-extracellular-matrix proteins [4]. MMP12 may inhibit angiogenesis through cleavage of plasminogen and collagen XVIII, resulting in the generation of angiostatin and endostatin, which have an angiostatic effect [2,5,6]. On the other hand, MMP12 may promote angiogenesis by cleaving structural components of the extracellular matrix, such as collagen type IV and fibrin [2]. It has been shown that increased expression of MMP12 may reflect a favorable prognosis in a few cancers [2].
A-82G polymorphism is located on the promoter region of the MMP12 gene where the transcription factor activator protein 1 (AP1) binds. The A allele is associated with a higher binding affinity for AP1, resulting in higher MMP12 promoter activity in vitro [7]. A study showed that the A allele was associated with smaller coronary artery luminal diameter in diabetic patients treated with percutaneous transluminal coronary angiography and stent implantation [7]. In another study, however, no association was found with risk of coronary aneurysm [8]. A1082G polymorphism is located on the coding region of the hemopexin domain that is responsible for MMP12 activity. The substitution of the G allele for the A allele results in an amino acid change from asparagine (Asn) to serine (Ser) in codon 357. The functional significance of this single nucleotide polymorphism, however, has not been clearly determined. In this study, we evaluated the association of these two common polymorphisms of the MMP12 gene with breast cancer risk and survival in the Shanghai Breast Cancer Study.
Materials and methods
Study participants and design
The Shanghai Breast Cancer Study is a population-based case-control study conducted in urban Shanghai. Detailed study design and data collection procedures have been described elsewhere [9]. Briefly, cases were permanent Shanghai residents between the ages of 25 and 64 years who were newly diagnosed with breast cancer between August 1996 and March 1998. Through a rapid case ascertainment system, supplemented by the population-based Shanghai Cancer Registry, 1,602 eligible breast cancer patients were identified, and 1,459 (91.1%) completed in-person interviews using a structured questionnaire. The initial cancer diagnoses for all patients were confirmed by two senior pathologists through a review of pathological slides. Information about clinical cancer characteristics, including TNM (tumor, node, metastasis) stage, treatment for cancer, and estrogen receptor (ER) and progesterone receptor (PR) status, was obtained by medical record review using a standard protocol. The major reasons for nonparticipation were refusal (109 cases; 6.8%), death before the patient could be interviewed (17 cases; 1.1%), and our inability to locate the patient (17 cases; 1.1%)
Eligible controls were randomly selected from the Shanghai Resident Registry, which contains demographic information for all residents of urban Shanghai, and were frequency-matched on age by 5-year intervals to the predetermined age distribution of the cases reported to the Shanghai Cancer Registry from 1990 to 1993. Of the 1,734 eligible controls, 1,556 (90.3%) completed interviews. The major reasons for nonparticipation of the eligible controls were refusal (166 controls, 9.6%) or death or a prior cancer diagnosis (2 controls, 0.1%).
The structured questionnaire used for this study included information on demographic factors, menstrual and reproductive history, hormone use, previous disease history, family history of cancer, physical activity, tobacco and alcohol use, and a quantitative food-frequency questionnaire. All participants were measured for current weight, circumferences of the waist and hips, and sitting and standing height. In addition to the in-person interviews and anthropometric measurement, 10 ml blood samples were collected from 1,193 (82%) cases and 1,310 (84%) controls. These samples were processed on the same day and stored at -70°C.
The methodology for the follow-up of cancer cases was described previously [10]. All 1,459 cancer patients were followed through January 2003 with active follow-up and record linkage to the death certificates of the Vital Statistics Unit of the Shanghai Center for Disease Control and Prevention. In all, 1,290 (88.4%) patients successfully completed the follow-up interview either in person (n = 1,241; 85%) or by telephone (n = 49; 3.4%) between March 2000 and December 2002. Among them, 197 patients were deceased. Through interviews with patients – or, for deceased patients, next of kin – information on disease progress, recurrence of cancer, quality of life, and cause of death (if the patient had died) was obtained. For the remaining 169 participants, who could not be contacted in person or by phone, linkage to the death certificates was completed in June 2003. Forty deaths were identified through the linkage, and information on the date of death and cause of death was obtained. The remaining 126 subjects who had no match in the death registry were assumed to be alive on December 30, 2002, 6 months before the linkage in order to allow for a possible delay of entry of the death certificates into the registry. Four subjects had insufficient information for the record linkage and were excluded from survival analysis. Finally, 1,129 cases and 1,229 controls were included in the case-control comparison and and 1,125 cases were included in the survival analysis
Genotyping methods
Genomic DNA was extracted from buffy coat fractions using a Puregene® DNA Purification kit (Gentra Systems, Minneapolis, MN, USA) following the manufacturer's protocol. DNA concentration was measured by PicoGreen® dsDNA Quantitation Kit (Molecular Probes, Eugene, OR, USA). The allelic discrimination of the MMP12 gene A-82G and A1082G polymorphisms were assessed with the ABI PRISM 7900 Sequence Detection Systems (Applied Biosystems, Foster City, CA, USA), using the fluorogenic 5' nuclease assay with primers and probes obtained from ABI (Assay ID: C_15880589_10 and C_785907_10). PCR was performed in a total volume of 5 μl, which contained 2.5 ng DNA, 1 × TaqMan Universal PCR Master Mix, each primer at 900 nM, and each probe at 200 nM. The thermal cycling conditions were as follows: 95°C for 10 min to activate the AmpliTaq Gold enzyme, followed by 40 cycles of 92°C for 15s and 60°C for 1 min. The fluorescence level was measured with an ABI PRISM 7900HT Sequence Detector (Applied Biosystems), resulting in clear identification of three genotypes.
The laboratory staff was blind to the identity of the subjects. Quality control samples were included in the genotyping assays. Each 384-well plate contained four water, eight CEPH 1347-02 DNA, eight blinded quality control samples, and eight unblinded quality control samples. The concordances for the blinded samples were 98% for A-82G and 100% for A1082G polymorphisms, respectively. Genotypes for polymorphisms of A-82G in the MMP12 gene were successfully determined for 1,118 cases and 1,223 controls and those of A1082G for 992 cases and 976 controls.
Statistical methods
The χ2 test and t-test were used for comparing characteristics of cases and controls. Minor genotypes AG or GG of A-82G and A1082G were combined in stratified analysis because of the small number of subjects in each category. Odds ratios and 95% confidence intervals (CIs) were derived using unconditional logistic regression models. To evaluate the association of MMP12 with survival, Cox proportional hazard models were applied after adjusting for age, TNM stages, and ER/PR status. The proportional hazard assumption of the Cox model was examined by graphic evaluation of Schoenfeld's residual plot. All P values presented in this paper are two-sided. SAS software was used for statistical analysis (version 9.1; SAS Institute, Cary, NC, USA).
Results
The distribution of demographic characteristics and known breast cancer risk factors of the cases and controls are presented in Table 1. Consistent with our previous reports [9,11], reproductive risk factors such as early menarche, late menopause, and late age at the first live birth were related to increased breast cancer risk. Cases were also more likely than controls to have higher body mass index (BMI), waist-to-hip ratio, or history of breast fibroadenomas, and were less likely to have exercised regularly during the preceding 10 years. The case-control difference was not statistically significant in age and education.
The distributions of MMP12 A-82G and A1082G genotypes are shown in Table 2. In the controls, the genotype frequency of the A-82G polymorphism did not deviate from the Hardy–Weinberg equilibrium, but that of the A1082G genotype deviated marginally (P = 0.05). In the cases, the genotype frequencies of both polymorphisms deviated from the Hardy–Weinberg equilibrium; this deviation was not likely to have been due to a laboratory error, because the concordances for the quality-control samples were more than 98%. Small numbers of subjects in the GG genotypes of both polymorphisms would be a possible explanation for this deviation. The frequencies of the minor G allele of A-82G (0.029 for controls and 0.026 for cases) were substantially lower than those previously reported for Caucasian populations, which ranged from 0.11 to 0.19 [7,8,12,13], whereas the minor allele frequencies of A1082G (0.107 for controls and 0.112 for cases) were higher than in one previous report of 0.05 [12]. In agreement with an earlier report [12], we found that these two polymorphisms are not in linkage disequilibrium [14].
Overall, there were no associations of breast cancer risk with either A-82G or A1082G polymorphisms alone or in combination. The genotype association did not differ by age (<45 years vs ≥ 45 years old at the time of diagnosis), menopausal status, or family history of breast cancer (data not shown).
The frequencies of minor genotypes of both polymorphisms were not significantly higher among patients with an advanced stage of breast cancer, nor did they differ by ER/PR status (Table 3).
The association of two polymorphisms of the MMP12 gene with breast cancer survival and disease-free survival are presented in Table 4 and Fig. 1. Patients who had the AG or GG genotypes of A1082G showed poorer overall survival than patients who had the AA genotype (hazard ratio (HR) 1.36, 95% CI 0.92 to 2.00). Compared with those who had only the AA genotypes in both A-82G and A1082G polymorphisms, patients who had one or more of the minor genotypes in these polymorphisms showed a poorer overall survival (HR 1.42, 95% CI 0.99 to 2.04). Our data did not suggest an association between MMP12 gene polymorphisms and disease-free survival.
Discussion
This study suggests that two common polymorphisms (A-82G and A1082G) of the MMP12 gene may not be related to breast cancer risk. The A1082G polymorphism, however, may be associated with the prognosis for breast cancer patients. The association with survival seems to be independent of other clinical prognostic factors such as cancer stage or ER/PR status.
Yang and colleagues reported that overexpression of MMP12 in tumors correlated with increased survival and decreased tumor neovascularization in colorectal cancer patients [15]. Similarly, Kerkelä and colleagues reported that MMP12 expressed in macrophages in the tumor site correlated with well-differentiated cancer cells [16]. MMP12 is expressed in breast tissue and may exert its protective effect through the cleavage of plasminogen to angiostatin and of collagen XVIII to endostatin [17-19]. In addition, MMP12 is also involved in the cleavage of domain D1 of urokinase-type plasminogen activator cellular receptor, which is responsible for cell migration during tumor invasion and angiogenesis [20]. The A-82G polymorphic site is the binding site of AP1, and the A allele is related to increased MMP12 activity [3,7]. Given the functional significance of this single nucleotide polymorphism and the role of MMP12 in breast carcinogenesis, we hypothesized that this single nucleotide polymorphism may be related to breast cancer risk and survival. Our findings, however, do not support this hypothesis. The much lower frequency of the minor G allele in our study population than in Caucasian populations [7,8,12,13] substantially reduces the statistical power. Indeed, we had only 51% power to detect 30% decreased risk of AG or GG genotypes, assuming a type I error of 0.05 [21].
A1082G polymorphism of the MMP12 gene results in a substitution of amino acid Ser for Asn in codon 357. The function of this polymorphism has not yet been determined; however, the substitution of a hydroxylic amino acid (Ser) for an acidic amino acid (Asn) may affect the activity of the enzyme [12]. In our study, the AG or GG genotypes of A1082G polymorphism were associated with poor prognosis of breast cancer patients. This result was prominent only in overall survival, but not in disease-free survival. The information on overall survival, however, is likely to be more accurate than that on disease-free survival, because information on disease progress and recurrence was collected by interviews with patients, or kin of deceased patients, rather than by reviewing medical records. Further evaluation of this association in other populations is required.
Our study has several strengths. First, the population-based study design and the high participation rate minimize potential selection bias. Second, the homogeneous ethnicity of this population (Han Chinese) minimizes possible population stratification [22]. Third, including comprehensive lifestyle and clinical information makes it possible to consider potential confounding and interactive effects in data analysis.
Conclusion
Our study suggests breast cancer risk may not be associated with the A-82G and A1082G polymorphisms in the MMP12 gene. The minor G allele in the A1082G polymorphism, however, may be related to poorer prognosis for breast cancer patients. This is the first report on the association of the MMP12 gene polymorphism with breast cancer risk and survival, and the results need to be confirmed in other large-scale studies.
Abbreviations
AP = activator protein; CI = confidence interval; ER = estrogen receptor; HR = hazard ratio; MMP = matrix metalloproteinase; PR = progesterone receptor; TNM = tumor, node, metastasis.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
AS and WZ conducted data analysis and drafted the manuscript. All authors contributed to result interpretation and manuscript revision. QC performed lab assays. X-OS, Y-TG, and WZ designed the study, recruited subjects, and collected data and biological samples. WZ was the principal investigator of the study and secured the research funding. All authors read and approved the final manuscript.
Acknowledgements
The authors wish to thank Drs Qi Dai, Fan Jin, and Jia-Rong Cheng for their contributions in coordinating data and specimen collection in Shanghai and Ms Bethanie Hull for technical assistance in the preparation of this manuscript. This study would not have been possible without the support of all of the study participants and research staff of the Shanghai Breast Cancer Study. This study was supported by research grants RO1CA64277 and RO1CA90899 from the National Cancer Institute.
Figures and Tables
Figure 1 Overall survival among breast cancer patients analyzed according to A1082G polymorphism and combined A-82G and A1082G polymorphisms of the MMP12 gene and P values for survival curves derived using a log-rank test. (a) Patients carrying the AG or GG genotypes at the A1082G polymorphism had a lower overall survival rate than those who had the AA genotype. (b) Patients carrying either of the minor genotypes (AG or GG) in either A-82G or A1082G polymorphisms had a lower overall survival rate than those who had the AA genotype for both A-82G and A1082G polymorphisms.
Table 1 Comparisons of participants with MMP12 genotype information.
Participant characteristics Casesa (N = 1,129) Controlsa (N = 1,229) Pb
Demographic factors
Age (years) 47.6 ± 8.0 47.2 ± 8.7 0.20
Education (% high school or more) 43.4% 43.1% 0.89
Reproductive risk factors
Age (years) at menarche 14.5 ± 1.6 14.7 ± 1.7 <0.01
Age (years) at menopausec 48.2 ± 4.7 47.5 ± 4.9 0.04
Age (years) at first live birthd 26 ± 4.1 26.0 ± 3.8 <0.01
Other risk factors
Breast cancer among first-degree relatives (%) 3.4 2.4 0.11
Ever had breast fibroadenoma (%) 9.7 5.1 <0.01
Body mass index (kg/m2) 23.5 ± 3.4 23.2 ± 3.4 0.03
Waist-to-hip ratio 0.81 ± 0.06 0.80 ± 0.06 <0.01
Physically active during past 10 years (%) 19.3 25.7 <0.01
aValues are presented as means ± standard deviations unless otherwise noted.
bCalculated from the t-test for continuous variables and the χ2 test for categorical variables. cAmong postmenopausal women. dAmong parous women.
Table 2 Association of MMP12 A-82G and A1082G polymorphisms with breast cancer risk
MMP12 genotype Cases Controls Crude OR (95% CI) Adjusted ORa (95% CI)
No. (%) No. (%)
A-82G
AA 1063 (95.1) 1153 (94.3) 1.0 1.0
AG 51 (4.6) 69 (5.6) 0.9 (0.6–1.2) 0.9 (0.6–1.3)
GG 4 (0.3) 1 (0.1)
A1082G (Asn357Ser)
AA 739 (80.2) 784 (80.3) 1.0 1.0
AG 159 (17.2) 175 (17.9) 1.0 (0.8–1.3) 1.0 (0.8–1.3)
GG 24 (2.6) 17 (1.8)
Presence of any minor genotypesb
No 687 (75.4) 731 (75.4) 1.0 1.0
Yes 224 (24.6) 239 (24.6) 1.0 (0.8–1.2) 1.0 (0.8–1.2)
aAdjusted for age, education, age at menarche, menopausal status, age at menopause, age at first live birth, waist-to-hip ratio, and physical activity during past 10 years. bAG/GG genotypes for A-82G and AG/GG genotype for A1082G polymorphisms. CI, confidence interval; OR, odds ratio.
Table 3 Association of MMP12 polymorphisms with clinical stage and ER/PR status in breast cancer patients
MMP12 genotype
TNM stage No. (%)
Analyzed according to stage of cancer 0, I, or II III or IV P
A-82G
AA 876 (95.1) 113 (93.4) 0.42
AG/GG 45 (4.9) 8 (6.6)
A1082G (Asn357Ser)
AA 606 (80.5) 81 (79.4) 0.80
AG/GG 147 (19.5) 21 (20.6)
Presence of any minor genotypesa
No 563 (75.6) 74 (74.0) 0.73
Yes 182 (24.4) 26 (26.0)
Analyzed according to ER/PR status ER+/PR+ No. (%) ER-/PR- No. (%) ER+/PR- or ER-/PR+ No. (%)
A-82G
AA 401 (97.1) 188 (93.5) 144 (95.4) 0.11
AG/GG 12 (2.9) 13 (6.5) 7 (4.6)
A1082G (Asn357Ser)
AA 251 (77.0) 145 (84.8) 102 (83.6) 0.07
AG/GG 75 (23.0) 26 (15.2) 20 (16.4)
Presence of any minor genotypesa
No 238 (74.1) 133 (78.2) 96 (79.3) 0.41
Yes 83 (25.9) 37 (21.8) 25 (20.7)
aAG/GG genotypes for A-82G and AG/GG genotype for A1082G polymorphisms. ER, estrogen receptor; PR, progesterone receptor; TNM, tumor, node, metastasis.
Table 4 Association of MMP12 A-82G and A1082G polymorphism with survival in 1,125 breast cancer patients
MMP12 genotype Events/subjects 5-year survival rate (%) Crude HR (and 95% CI) Adjusted HRa (and 95% CI)
Overall survival
A-82G
AA 173/1059 84.4 1.0 1.0
AG/GG 10/55 81.4 1.14 (0.60–2.16) 1.06 (0.56–2.00)
AG 9/51 82.0
GG 1/4 75.0
A1082G (Asn357Ser)
AA 108/738 86.0 1.0 1.0
AG/GG 34/181 81.0 1.33 (0.90–1.95) 1.36 (0.92–2.00)
AG 32/157 79.4
GG 2/24 91.7
Presence of any minor genotypesb
No 99/686 86.3 1.0 1.0
Yes 43/222 80.4 1.40 (0.98–2.00) 1.42 (0.99–2.04)
Disease-free survival
A-82G
AA 224/1059 78.8 1.0 1.0
AG/GG 10/55 81.7 0.87 (0.46–1.64) 0.80 (0.42–1.51)
AG 9/51 82.3
GG 1/4 75.0
A1082G (Asn357Ser)
AA 151/738 79.5 1.0 1.0
AG/GG 38/181 78.8 1.04 (0.73–1.49) 1.07 (0.74–1.52)
AG 35/157 77.5
GG 3/24 87.5
Presence of any minor genotypesb
No 142/686 79.3 1.0 1.0
Yes 47/222 78.7 1.04 (0.75–1.45) 1.06 (0.76–1.47)
aAdjusted for age, TNM stage, and ER/PR status. bAG/GG genotypes for A-82G and AG/GG genotype for A1082G polymorphisms. CI, confidence interval; ER, estrogen receptor; HR = hazard ratio; PR, progesterone receptor; TNM, tumor, node, metastasis.
==== Refs
Vu TH Werb Z Matrix metalloproteinases: effectors of development and normal physiology Genes Dev 2000 14 2123 2133 10970876 10.1101/gad.815400
Egeblad M Werb Z New functions for the matrix metalloproteinases in cancer progression Nat Rev Cancer 2002 2 161 174 11990853 10.1038/nrc745
Ye S Polymorphism in matrix metalloproteinase gene promoters: implication in regulation of gene expression and susceptibility of various diseases Matrix Biol 2000 19 623 629 11102751 10.1016/S0945-053X(00)00102-5
Chakraborti S Mandal M Das S Mandal A Chakraborti T Regulation of matrix metalloproteinases: an overview Mol Cell Biochem 2003 253 269 285 14619979 10.1023/A:1026028303196
Dong Z Kumar R Yang X Fidler IJ Macrophage-derived metalloelastase is responsible for the generation of angiostatin in Lewis lung carcinoma Cell 1997 88 801 810 9118223 10.1016/S0092-8674(00)81926-1
Cornelius LA Nehring LC Harding E Bolanowski M Welgus HG Kobayashi DK Pierce RA Shapiro SD Matrix metalloproteinases generate angiostatin: effects on neovascularization J Immunol 1998 161 6845 6852 9862716
Jormsjo S Ye S Moritz J Walter DH Dimmeler S Zeiher AM Henney A Hamsten A Eriksson P Allele-specific regulation of matrix metalloproteinase-12 gene activity is associated with coronary artery luminal dimensions in diabetic patients with manifest coronary artery disease Circ Res 2000 86 998 1003 10807873
Lamblin N Bauters C Hermant X Lablanche JM Helbecque N Amouyel P Polymorphisms in the promoter regions of MMP-2, MMP-3, MMP-9 and MMP-12 genes as determinants of aneurysmal coronary artery disease J Am Coll Cardiol 2002 40 43 48 12103254 10.1016/S0735-1097(02)01909-5
Gao YT Shu XO Dai Q Potter JD Brinton LA Wen W Sellers TA Kushi LH Ruan Z Bostick RM Association of menstrual and reproductive factors with breast cancer risk: results from the Shanghai Breast Cancer Study Int J Cancer 2000 87 295 300 10861490 10.1002/1097-0215(20000715)87:2<295::AID-IJC23>3.0.CO;2-7
Shu XO Gao YT Cai Q Pierce L Cai H Ruan ZX Yang G Jin F Zheng W Genetic polymorphisms in the TGF-beta 1 gene and breast cancer survival: a report from the Shanghai Breast Cancer Study Cancer Res 2004 64 836 839 14871809
Shu XO Cai Q Gao YT Wen W Jin F Zheng W A population-based case-control study of the Arg399Gln polymorphism in DNA repair gene XRCC1 and risk of breast cancer Cancer Epidemiol Biomarkers Prev 2003 12 1462 1467 14693738
Joos L He JQ Shepherdson MB Connett JE Anthonisen NR Pare PD Sandford AJ The role of matrix metalloproteinase polymorphisms in the rate of decline in lung function Hum Mol Genet 2002 11 569 576 11875051 10.1093/hmg/11.5.569
Zhang B Dhillon S Geary I Howell WM Iannotti F Day IN Ye S Polymorphisms in matrix metalloproteinase-1, -3, -9, and -12 genes in relation to subarachnoid hemorrhage Stroke 2001 32 2198 2202 11546917
Devlin B Risch N A comparison of linkage disequilibrium measures for fine-scale mapping Genomics 1995 29 311 322 8666377 10.1006/geno.1995.9003
Yang W Arii S Gorrin-Rivas MJ Mori A Onodera H Imamura M Human macrophage metalloelastase gene expression in colorectal carcinoma and its clinicopathologic significance Cancer 2001 91 1277 1283 11283927 10.1002/1097-0142(20010401)91:7<1277::AID-CNCR1129>3.0.CO;2-H
Kerkelä E Ala-aho R Klemi P Grenman S Shapiro SD Kahari VM Saarialho-Kere U Metalloelastase (MMP-12) expression by tumour cells in squamous cell carcinoma of the vulva correlates with invasiveness, while that by macrophages predicts better outcome J Pathol 2002 198 258 269 12237887 10.1002/path.1198
Heppner KJ Matrisian LM Jensen RA Rodgers WH Expression of most matrix metalloproteinase family members in breast cancer represents a tumor-induced host response Am J Pathol 1996 149 273 282 8686751
Zucker S Vacirca J Role of matrix metalloproteinases (MMPs) in colorectal cancer Cancer Metastasis Rev 2004 23 101 117 15000152 10.1023/A:1025867130437
Ferreras M Felbor U Lenhard T Olsen BR Delaissé JM Generation and degradation of human endostatin proteins by various proteinases FEBS Lett 2000 486 247 251 11119712 10.1016/S0014-5793(00)02249-3
Koolwijk P Sidenius N Peters E Sier CF Hanemaaijer R Blasi F van Hinsbergh VW Proteolysis of the urokinase-type plasminogen activator receptor by metalloproteinase-12: implication for angiogenesis in fibrin matrices Blood 2001 97 3123 3131 11342439 10.1182/blood.V97.10.3123
Schlesselman JJ Sample size Case-Control Studies 1982 New York: Oxford University Press 144 150
Freedman ML Reich D Penney KL McDonald GJ Mignault AA Patterson N Gabriel SB Topol EJ Smoller JW Pato CN Assessing the impact of population stratification on genetic association studies Nat Genet 2004 36 388 393 15052270 10.1038/ng1333
| 15987457 | PMC1175062 | CC BY | 2021-01-04 16:54:50 | no | Breast Cancer Res. 2005 May 10; 7(4):R506-R512 | utf-8 | Breast Cancer Res | 2,005 | 10.1186/bcr1033 | oa_comm |
==== Front
Breast Cancer ResBreast Cancer Research1465-54111465-542XBioMed Central London bcr10341598745410.1186/bcr1034Research ArticleDesigning a HER2/neu promoter to drive α1,3galactosyltransferase expression for targeted anti-αGal antibody-mediated tumor cell killing Lanteri Marion [email protected] Laurence [email protected] Valérie [email protected] Jean-Claude [email protected] INSERM U526, Laboratoire de Virologie, Faculté de Médecine, avenue de Valombrose, 06107, Nice cedex 2, France2005 3 5 2005 7 4 R487 R494 6 12 2004 24 2 2005 10 3 2005 5 4 2005 Copyright © 2005 Lanteri et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Introduction
Our goal was to specifically render tumor cells susceptible to natural cytolytic anti-αGal antibodies by using a murine α1,3galactosyltransferase (mαGalT) transgene driven by a designed form of HER2/neu promoter (pNeu), the transcription of which is frequently observed to be above basal in breast tumors. Indeed, the αGalT activity that promotes Galα1,3Galβ1,4GlcNAc-R (αGal) epitope expression has been mutationally disrupted during the course of evolution, starting from Old World primates, and this has led to the counter-production of large amounts of cytotoxic anti-αGal antibodies in recent primates, including man.
Method
Expression of the endogenous c-erbB-2 gene was investigated in various cell lines by northern blotting. A mαGalT cDNA was constructed into pcDNA3 vector downstream of the original CMV promoter (pCMV/mαGalT) and various forms of pNeu were prepared by PCR amplification and inserted in the pCMV/mαGalT construct upstream of the mαGalT cDNA, in the place of the CMV promoter. These constructs were transferred into HEK-293 control and breast tumor cell lines. Stably transfected cells were analyzed by northern blotting for their expression of αGalT and c-erbB-2, and by flow cytometry for their binding with fluorescein isothiocyanate-conjugated Griffonia simplicifolia/isolectin B4.
Results
We show that expression of the mαGalT was up- or down-modulated according to the level of endogenous pNeu activity and the particular form of constructed pNeu. Among several constructs, two particular forms of the promoter, pNeu250 containing the CCAAT box and the PEA3 motif adjacent to the TATAA box, and pNeu664, which has three additional PEA3 motifs upstream of the CCAAT box, were found to promote differential αGalT expression.
Conclusion
Our results strengthen current concepts about the crucial role played by the proximal PEA3 motif of pNeu, and may represent a novel therapeutic approach for the development of targeted transgene expression.
==== Body
Introduction
The enzyme α1,3galactosyltransferase (αGalT) is responsible for the synthesis of galactose-α1,3galactose-β1,4N-acetylglucosamine-R (αGal) epitopes in all mammals except Old World primates [1]. Highly expressed in nonprimate mammals, prosimians and New World monkeys, this glycosyltransferase has been mutationally inactivated during the course of evolution, starting from Old World primates [2]. We have previously shown that, in human cells, transcription of the αGalT gene is interrupted by the presence of a strong stop signal in exon 7, which leads to a chimeric mRNA comprising the first four coding exons and part of intron VII, but lacking the last two exons corresponding to the catalytic domain of the enzyme [3]. As a consequence, and given the broad circulation of αGal carbohydrate antigens, humans, apes and Old World monkeys produce large amounts of anti-αGal antibodies, which represent approximately 1% of total IgG in humans [4]. These antibodies are responsible for the hyperacute rejection of xenografts and thus prevent trials on transplantation of pig organs to humans [5,6]. Conversely, they represent a potential constitutive tool for therapeutic applications because their highly efficient cytolytic activity could be directed against pathological cells transgenically modified to express αGal epitopes [7-10].
Gene therapy based on the induction of cytotoxicity generally makes use of transgenes that encode prodrug activating enzymes [11]. In the case of anti-αGal-induced cytotoxicity, no chemical drug is needed to obtain the effect of αGalT because natural circulating anti-αGal antibodies are sufficient to promote cell lysis via complement activation. One common problem with gene therapy is target cell selectivity. A frequent solution to this is the use of tissue-specific or tumor-activated promoters to drive expression of the transgene [12,13]. Human c-erbB-2 (synonyms erbB2, HER2/neu), a member of the erbB family that is overexpressed in about one third of breast tumors and in a variety of other tumors, is often correlated with a poor prognosis [14-16]. This gene is normally expressed at a low level in a variety of human embryonic and adult epithelial and hematopoietic cells [17,18]. The high overexpression of c-erbB-2 in tumor cells [19] results from multiple mechanisms, including gene amplification and transcription modulation [20-22]. c-erbB-2 is a 185 kDa transmembrane tyrosine kinase receptor related to the epidermal growth factor receptor that functions as a growth factor receptor to regulate cell growth and transformation [23-25]. Regulation of the c-erbB-2 promoter (pNeu) has been extensively investigated in a domain located within the 700 bp region upstream of its transcription start site. A -213/-87 fragment relative to the gene's transcription start site contains the minimal promoter region able to drive preferential transgene expression in breast cancer cells [26].
The present study was designed to obtain targeted expression of αGal epitopes by human breast cancer cells in order to render them susceptible to anti-αGal antibody-induced cell death. For this purpose, we used a murine αGalT (mαGalT) transgene driven by a form of pNeu designed to be significantly activated in breast tumor cells.
Materials and methods
Cells and reagents
The cell line HEK-293 (ATCC CRL-1573) and the human breast cancer cell lines MCF-7 (ATCC HTB-22), SK-BR-3 (ATCC HTB-30), MDA-MB-231 (ATCC HTB-26), and MDA-MB-453 (ATCC HTB-131) were cultured in Dulbecco's modified Eagle's medium (Gibco, Invitrogen, Rockeville, MD, USA) supplemented with 10% fetal calf serum (FCS; BioWhittaker, Rockland, ME, USA). Fluorescein isothiocyanate-conjugated Griffonia simplicifolia / isolectin B4 (FITC-GS-I-B4), which recognizes a terminal galactosyl residue in an α linkage, was purchased from EY Laboratories (San Matteo, CA, USA). The rabbit complement ORAX 07 was from Dade Behring (Paris, France).
Murine αGalT constructions in a plasmid vector and transfection
The mαGalT cDNA was kindly provided by Uri Galili and cloned within HindIII/XbaI sites into pcDNA3 vector (Invitrogen, Cergy-Pontoise, France), downstream of the CMV early promoter (pCMV) or various truncated forms of the HER2/neu promoter (pNeu) obtained by PCR on genomic DNA extracted from human CEM cells (ATCC CCL-119) using the following primer sets: 5'-GGGGGTCCTGGAAGCCACAAG-3' and 5'-GTGCTCACTGCGGCTCCGGCC-3' for pNeu664 (-488/+176); 5'-TCGCGAGCAGGCAACCCAGGCGTCCCG-3' and 5'-AAGCTTCTCCCCTGGTTTCTCCGGTCCCAA-3' for pNeu250 (-216/+34); 5'-TCGCGAGCAGGCAACCCAGGCGTCCCG-3' and 5'-CCAAAAAGCTTGTGCTCACTGCGGCTCCGGCC-3' for pNeu392 (-216/+176); 5'-GGAAATCGCGAAGGAAGTATAAGAATGAAG-3' and 5'-CCAAAAAGCTTGTGCTCACTGCGGCTCCGGCC-3' for pNeu209 (-33/+176). All pNeu derivative forms were constructed within NruI/HindIII sites upstream of mGalT cDNA in the place of pCMV.
Cells were transfected in six-well plates using FuGENE 6 transfection reagent (Roche Diagnostics, Meylan, France), as recommended by the manufacturer. Stably transfected cells were selected by G418 resistance.
Murine αGalT constructions in a retroviral vector
To overcome the poor efficiency of classic methods of transfection in MDA-MB-231 cells, a retroviral vector system was developed. The undesirable promoting activity of the 5' long terminal repeat (LTR) was avoided by constructing the cassette pNeu250/mαGalT in a self-inactivating murine retroviral vector (pcPMΔU3) that had been prepared by removing nearly the entire U3 region of the 3' LTR (Lefebvre JC and March D, manuscript in preparation). Making use of this strategy, described in [27], the U3 deletion is transferred to the 5' LTR during the first retrotranscription of the retroviral construct, and further results in the transcriptional inactivation of the provirus in the infected cells. In addition, the cassette pNeu250/mαGalT was oriented in the opposite direction (3' to 5') to the LTR so as to completely rule out any residual viral promoter activity. To obtain retroviral particles pseudotyped with a vesicular stomatitis viral G glycoprotein (VSV-G), the plasmid construct was co-transfected in GP2-293 packaging cells with a pVSV-G vector (both from Clontech, BD Biosciences, Le Pont de Claix, France). Supernatants were harvested 48 h post-transfection and filtered (membrane pore size = 0.45 μm). VSV-G pseudotyped particles were concentrated by ultracentrifugation. Infected cells were seeded in 24-well plates (BD Falcon, Le Pont de Claix, France) and stably transduced subclones were selected by antibiotic resistance. Expression of m αGalT was evaluated using GS-I-B4 reactivity.
Flow cytometry analysis
Phenotypic analyses were performed using FITC-GS-I-B4, as previously described [28]. Stained cells were analyzed on a FACScan cytometer (Becton Dickinson, San Jose, CA, USA).
Northern blot analysis
Total RNAs were isolated using RNA Now reagent (Biogentex Inc., Seabrook, TX, USA), according to the manufacturer's instructions, and poly(A)-rich RNAs were selected as described elsewhere [29]. Poly(A)-rich RNAs (3 μg) were electrophoresed on denaturing 1.2% agarose gel and transfered in 20 × NaCl/Citrate onto a Hybond-N+ nylon membrane (Amersham-Biosciences, Saclay, France). Membranes were probed overnight at 42°C with [α-32P]-random-labeled mαGalT or c-erbB-2 cDNA and washed according to standard procedures. The mαGalT probe was excised from the pcDNA3 construct and the c-erbB-2 probe was obtained by PCR amplification on genomic DNA from SK-BR-3 cells with the primer set 5'-CCAGGAGGTGCAGGGCTACG-3' and 5'-ATCCTCAGAACTCTCTCCCC-3'. Membranes were exposed to Kodak XAR film (Eastman Kodak, Rochester, NY, USA). Detection of glyceraldehyde 3-phosphate-dehydrogenase (GAPDH) was used as an internal control.
Cytotoxicity assay
Parental and transfected HEK-293, MCF-7, or SK-BR-3 cells were distributed in 96-well plates (5.104 cells/well) and incubated for 1 h after addition of 10 μl of various human sera, in triplicate. Rabbit complement (20 μl) was then added and plates were incubated for 1 h. Cell death was evaluated by the trypan blue vital dye exclusion method. The percentage of killed cells was evaluated by comparison of the number of blue-stained cells in the reaction and control wells.
Cell proliferation assay
After cell incubation with human sera and rabbit complement as above, XTT reagent from the Cell Proliferation Kit II (Roche Diagnostics, Meylan, France) was added, according to the manufacturer's instructions, and the cells were re-incubated at 37°C for 2 h. Formazan formation was measured at 490 nm.
Statistical analysis
Statistical comparison of mean values was performed with a one-way analysis of variance (ANOVA).
Results
Expression of the c-erbB-2 gene in various cell lines
Expression of the endogenous c-erbB-2 gene was investigated in various cell lines by northern blotting. c-erbB-2 was very weakly expressed in HEK-293 cells and was differentially transcribed in human breast cancer cell lines (Figs 1 and 2C): it was absent in MDA-MB-231, moderately expressed in MCF-7 and MDA-MB-453, and strongly expressed in SK-BR-3 cells. These results are consistent with data published by others [21,30,31]. SK-BR-3 cells are known to overexpress c-erbB-2 as the result of gene amplification in proportions estimated at up to 13:1 [21]. MCF-7 is a known HER2/neu non-amplified cell line [20], but various degrees of gene expression have been reported [30,32-34]. The HER2/neu transcription level of the MCF-7 cell line used in our laboratory was notably superior to that of the HEK-293 cells (Fig. 1). Serially passaged in different laboratories, MCF-7 cell lines probably exhibit variable levels of HER2/neu transcription. Immortalized by adenovirus type 5 (ad5) [35], HEK-293 cells require continuous ad5 E1A expression to proliferate and avoid senescence [36], and can thus be considered subnormal because they are not tumorigenic [37]. Interestingly, their very weak HER2/neu expression done them useful as controls to investigate the activation of various forms of pNeu as a function of HER2/neu expression in breast tumor cells. Because our goal was to take advantage of differential up-regulation of endogenous pNeu to overexpress a suicide transgene that is itself driven by an exogenous pNeu, MCF-7 cells without HER2/neu gene amplification appeared more suitable than HEK-293 cells. SK-BR-3 cells might provide other information, as detailed hereafter.
Variable activity of mαGalT driven by various forms of c-erbB-2 promoter in human breast cancer cell lines
Human cells do not express any αGalT activity responsible for αGal epitope expression that is recognized by the GS-I-B4 lectin. GS-I-B4 binding might thus specifically reveal expression of exogenous mαGalT in transgenically modified human cells. A mαGalT cDNA was constructed into pcDNA3 vector downstream of the original pCMV (pCMV/mαGalT). In addition, various forms of pNeu were prepared by PCR amplification (Fig. 2a): pNeu664 (nucleotides -488/+176, relative to the transcription start site of HER2/neu), pNeu392 (nucleotides -216/+176) and pNeu250 (nucleotides -216/+34). These forms were inserted in the pCMV/mαGalT construct upstream of the mαGalT cDNA, in the place of the original pCMV (pNeu664/mαGalT, pNeu392/mαGalT and pNeu250/mαGalT). These constructs were transferred into HEK-293 control and breast tumor cell lines. Stably transfected cells were analyzed by northern blotting for their expression of αGalT (Fig. 2b) and c-erbB-2 (Fig. 2c), and by flow cytometry for their binding with FITC-GS-I-B4 (Fig. 3a–c). pCMV and pNeu664 promoted noticeable expression of αGalT in both HEK-293 and breast cancer cell lines (Fig. 2b, lanes 2–3, 7–8 and 12–13). In contrast, the shortest form, pNeu250, raised αGal expression to a more specific level in MCF-7 than in HEK-293 cells (Fig. 2b, compare lane 10 to 5). Similar differential results were observed with pNeu392 (Fig. 2b, compare lanes 9 and 14 to 4) whereas a complete switch-off was observed with pNeu209 (data not shown). These results were verified phenotypically. Elevated αGal expression was observed in breast tumor and HEK-293 cells with mαGalT driven by pCMV or pNeu664, while pNeu250 promoted αGal expression only in the breast tumor cells SK-BR-3 and MCF-7 (Fig. 3a). To confirm these results, expression of the cassette pNeu250/mαGalT was compared for breast tumor cells expressing and not expressing c-erbB-2. The non-expressing MDA-MB-231 cell line, although appropriate for this purpose, was unfortunately resistant to classic transfection methods. A self-inactivating retroviral vector pseudotyped by a VSV-G glycoprotein was thus used to transfer pNeu250/mαGalT. Stably transduced cell lines were subcloned during the course of antibiotic selection. Four to five subclones of each type were analyzed for the expression of mαGalT. The MDA-MB-231 cells, which did not express c-erbB-2, showed a very low level of GS-I-B4 reactivity compared to breast tumor cell lines SK-BR-3 and MCF-7, which express c-erbB-2 (Fig. 3c).
Apparently conflicting results show that pNeu250 promoted clearly higher αGal expression in MCF-7 than in SK-BR-3 cells (Fig. 3a,b), while c-erbB-2 was inversely expressed in these two cell lines (Fig. 1, lanes 2 and 3; Fig. 2c, lanes 4 and 7). These data could be explained by the differential mechanisms that sustain HER2/neu overexpression, which is regulated at the transcriptional level in MCF-7 cells and is dependent on the existence of multiple gene copies in SK-BR-3 cells. Interestingly, pNeu250 promoted a much more specific expression as a function of the differential level of endogenous c-erbB-2 transcription in breast tumor cells (Fig. 1, lanes 2–4) and HEK-293 cells (Fig. 1, lane 1).
Cytolytic activity of anti-αGal antibodies to transgenically modified breast tumor cells
A high fraction of antibodies from human sera bind αGal epitopes and can efficiently induce the death of cells that exhibit these epitopes via complement activation [4]. Cytotoxicity assays were carried out on HEK-293, MCF-7 and SK-BR-3 cells stably transfected by mαGalT cDNA under the control of pNeu664 and pNeu250, or pCMV used as a control. The susceptibility of transfected cells to natural human anti-αGal antibodies was verified using a complement-dependent cytotoxicity test (data not shown). Cell death averages were confirmed by an XTT proliferation assay (Fig. 4). All of the cell types transfected with pCMV/mαGalT or pNeu664/mαGalT were killed, but to varying degrees (Fig. 4). Interestingly, differential cytolytic activity of antibodies was observed with pNeu250/mαGalT, being much greater in MCF-7 and SK-BR-3 cells than in HEK-293 cells (Fig. 4). Here again, mαGalT was much more efficiently driven by pCMV or pNeu664 in HEK-293 cells than in breast tumor cells. pNeu250 gave differential results that were in favor of the tumor cells; it promoted a more significant proportion of death in MCF-7 than in SK-BR-3 cells (Fig. 4) that correlated with their respective levels of αGal epitope expression (Fig. 3).
Discussion
Efforts to develop anticancer therapies based on suicide transgenes generally focus on prodrug activating enzymes [38] combined with effective targeting of pathological cells. We have been studying αGalT gene expression and the very high efficiency of natural anti-αGal antibodies in inducing complement-mediated cell killing [39,40] in the field of hyperacute xenograft rejection. We attempted to take advantage of this constitutive immune system to target tumor cells. Several authors have demonstrated the efficacy of natural anti-αGal antibodies for destruction of tumor cells [7-10]. Indeed, the high numbers of circulating oligosaccharides bearing αGal epitopes are responsible for constant booster immunizations. This may explain the high plasma level (1% of IgG) of anti-αGal antibodies in humans [4] and their constant de novo synthesis in αGalT knockout mice [41,42]. Natural anti- αGal antibodies are highly cytotoxic and cytolytic as the result of highly efficient complement activation, and this results in hyperacute rejection of xenografts [2,40]. They have also been shown capable of protecting αGalT-deficient mice against engrafted αGal+ colon cancer cells [9].
The purpose of this study was the targeting of tumor cells known to overexpress c-erbB-2 using a selected form of its promoter pNeu to drive an active αGalT. This approach was designed to take advantage of the effective antibodies preexisting in all humans. In a similar study, a derived form of the human telomerase promoter was shown to render human pancreatic carcinoma cells susceptible to αGal/complement-mediated cell killing [43]. We selected pNeu because it has been well characterized and is overexpressed in a variety of tumors [16,44]. Because definition of a precise pNeu sequence with well-restricted activation in tumor cells remains uncertain, however, we analyzed several forms of this promoter. The shortest form, pNeu209, which comprises the only PEA3 motif adjacent to the TATA box plus two SP1 sites and one AP-2 site downstream of the transcription start site, promoted very weak αGal expression. The minimal forms, pNeu392 and pNeu250, were equally capable of selectively inducing αGalT in breast tumor cells compared with HEK-293 cells. We thus concluded that the Ap-2 and SP1 motifs downstream of the transcription start site (Fig. 2) were not essential. The noticeable absence of a CCAAT box in pNeu209 probably explains its disrupted activity because in cells over-expressing c-erbB-2, the CCAAT box is up-regulated rather than the TATAA box [45]. Further studies thus focused on comparing pNeu250 with the longest form, pNeu664. This last form contains several PEA3, NF-kB, HER2 transcription factor (HTF) and SP1 sites upstream of the minimal pNeu250. The role played by the Ets family and activator protein-2 (AP-2) factors has been extensively studied in breast tumor cells. While the AP-2 binding site was present in both pNeu250 and pNeu664, the main difference between these forms was the presence of three additional PEA3 motifs in pNeu664. Activation of pNeu664 was virtually the same in MCF-7 and SK-BR-3 tumor cells, whereas a marked decrease in pNeu250 activity was observed only in SK-BR-3 cells (Fig. 3), in complete contrast to their high c-erbB-2 expression (Fig. 2c). As discussed above, the striking overexpression of c-erbB-2 in SK-BR-3 cells can be explained by their multiple gene copies. In MCF-7 cells, the differential promoting activity of pNeu250 could be explained by the relative increase in the transcription level compared with HEK-293 cells. In other aspects, in association with c-erbB-2 gene amplification, up-regulation of transcriptional factors that control endogenous pNeu remains possible. Conflicting results have been published on Ets regulation of c-erbB-2, with activation and repression of pNeu by PEA3 factors having been reported [46,47]. The observation that Ets binding leads to a severe bend in DNA could be further support for our findings [46]. When the number of PEA3 binding sites is reduced from four in pNeu664 to one in pNeu250, over-occupation of the single remaining site in pNeu250 might hinder formation of the required DNA conformation rather than favor its reading.
The differential promoting activity of pNeu664 and pNeu250 in HEK-293 cells (Fig. 3) does not appear to be relevant to the transcriptional regulation of c-erbB-2 because this gene is only weakly expressed in these cells (Fig. 1, lane 1). In contrast, HEK-293 cells continuously express ad5 E1A, which has been shown to target pNeu [48] as a repressor of HER2/neu overexpression [49], and has been proposed for use in cancer gene therapy [50]. Like the Ets factors expressed in tumor cells, an equal level of E1A in HEK-293 cells might activate pNeu664, which contains four PEA3 motifs, and repress pNeu250, which has only one.
Our efforts to take advantage of natural cytotoxic anti-αGal antibodies as a means of destroying breast tumor cells, and to design a promoter specific for these undesirable cells, have to be considered as a preliminary contribution to the field of cancer gene therapy, given that our results have been obtained in cell line culture models. It has been shown that human primary breast tumors can be successfully engrafted into NOD/SCID mice and maintained in a growing state for more than 100 days [51]. Moreover, αGalT(-/-) KO mice have been fortunately generated by others [52,53]. So to progress towards a gene therapy application, we are developing a two step procedure in mice. First, the distribution and expression of the transgene pNeu250/mαGalT, cloned into the retroviral vector pcPMΔU3 (see Materials and methods), will be studied in a human breast cancer xenograft model. Various types of human breast tumor differentially expressing HER2/neu will be implanted in NOD/SCID mice, and thereafter the transgene will be injected by a variety of methods. Second, αGalT-transduced tumor pieces will be transplanted into immunocompromised αGalT KO mice to evaluate the tumor destroying activity of purified human anti-αGal antibodies.
Conclusion
Our results show that the association pNeu250/mαGalT could be used to target tumor cells overexpressing c-erbB-2, and thus expose them to the cytolytic activity of natural anti-αGal antibodies. Development of a discriminating in vivo system capable of targeting tumor cells according to their level of c-erbB-2 expression could prove beneficial.
Abbreviations
ad5 = adenovirus type 5; AP-2 = activator protein-2; bp = base pair; FCS = fetal calf serum; FITC-GS-I-B4 = fluorescein isothiocyanate-conjugated Griffonia simplicifolia/isolectin B4; αGal = galactose-α1,3galactose-β1,4N-acetylglucosamine-R; αGalT = α1,3galactosyltransferase; GAPDH = glyceraldehyde 3-phosphate-dehydrogenase; LTR = long terminal repeat; mαGalT = murine αGalT; NF-kB = nuclear factor kB; pCMV = CMV promoter; PCR = polymerase chain reaction; pNeu = HER2/neu promoter; VSV-G = vesicular stomatitis viral G glycoprotein.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
ML performed the various pNeu constructs, transduction assays, Northern blot analyses and participated in literature search and critical reading of the manuscript. LO conducted the statistical analysis and participated in cell proliferation assays. VG carried out flow cytometry analysis and contributed in Northern blot analyses. JCL conceived the study and further developments, looked after data interpretation, and wrote the manuscript. All authors read and approved the final manuscript.
Acknowledgements
We thank Dr. Uri Galili for the kind gift of murine α1,3galactosyltransferase cDNA, and Nancy Reed for editing the manuscript. This work was supported by institutional grants from the Institut National de la Santé et de la Recherche Médicale (INSERM).
Figures and Tables
Figure 1 Expression of the c-erbB-2 gene in various cell lines. Total RNAs from HEK-293 and human breast cancer cell lines MCF-7, SK-BR-3, MDA-MB-231 and MDA-MB-453 were probed with c-erbB-2 cDNA and thereafter with glyceraldehyde 3-phosphate-dehydrogenase (GAPDH) cDNA as a control.
Figure 2 Promoting activity of newly designed forms of pNeu in various cell lines differentially expressing c-erbB-2 (a) Schematic representation of various forms of pNeu constructed upstream of the murine α1,3galactosyltransferase (mαGalT) cDNA. The transcription start site is indicated by arrows, and the 5' end points of the various forms of pNeu by bold bars. (b) Northern blot analysis of expression of mαGalT driven by the CMV promoter (pCMV) and the various forms of constructed pNeu. (c) Northern blot analysis of c-erbB-2 expression in stably transfected HEK-293, MCF-7 and SK-BR-3 cells. Total RNAs were electrophoresed and probed with [α-32P]-random-labeled mαGalT or c-erbB-2 cDNA, and thereafter with glyceraldehyde 3-phosphate-dehydrogenase (GAPDH) cDNA as a control.
Figure 3 Expression of Galα1,3Gal residues at the surface of transduced cells. Cells stably transduced or not with CMV promoter (pCMV)/murine α1,3galactosyltransferase (mαGalT), pNeu664/mαGalT, or pNeu250/mαGalT were labeled with fluorescein isothiocyanate-conjugated Griffonia simplicifolia/isolectin B4 (FITC-GS-I-B4) lectin (Galα1,3Gal) and analyzed on a FACScan cytometer (counts, cells numbers; FL1-H, fluorescence intensity). (a) Binding of FITC-GS-I-B4 to parental and pNeu250/mαGalT transduced HEK-293, MCF-7 and SK-BR-3 cells. Mean values of fluorescence detected by flow cytometry at the surface of stably transduced cells with (b) the constructs pNeu664/mαGalT or pNeu250/mαGalT in a plasmid vector (experiments in triplicates), and (c) the construct pNeu250/mαGalT in a self-inactivating retroviral vector (experiments were conducted using four to five subclones).
Figure 4 Cytotoxicity assays on HEK-293, MCF-7 and SK-BR-3 cells stably transfected or not with CMV promoter (pCMV)/murine α1,3galactosyltransferase (mαGalT), pNeu664/m αGalT, or pNeu250/mαGalT. Cells were distributed in 96-well plates (5.104/well) and incubated with human sera, after which rabbit complement was added. Cell death was analyzed using an XTT reagent proliferation assay. Results (means of triplicates) are expressed as percentages of cell death, relative to the amount of formazan formation (absorbance evaluated at 490 nm; NS, not significant).
==== Refs
Galili U Galili U, Avila JL Evolution of a 1,3Galactosyltransferase and of the a-Gal epitope a-Gal and Anti-Gal: a-1,3-galactosyltransferase, a-Gal epitopes, and the natural anti-Gal antibody 1999 New York, NY: Kluwer Academic/Plenum Publishers 1 18
Galili U Shohet SB Kobrin E Stults CL Macher BA Man, apes, and Old World monkeys differ from other mammals in the expression of alpha-galactosyl epitopes on nucleated cells J Biol Chem 1988 263 17755 17762 2460463
Lanteri M Giordanengo V Hiraoka N Fuzibet JG Auberger P Fukuda M Baum LG Lefebvre JC Altered T cell surface glycosylation in HIV type 1 infection results in increased susceptibility to galectin-1-induced cell death Glycobiology 2003 13 909 918 12925577 10.1093/glycob/cwg110
Galili U Avila JL a-Gal and Anti-Gal: a-1,3-galactosyltransferase, a-Gal epitopes, and the natural anti-Gal antibody 1999 New York, NY: Kluwer Academic/Plenum Publishers
Hortobagyi GN Hung MC Lopez-Berestein G A Phase I multicenter study of E1A gene therapy for patients with metastatic breast cancer and epithelial ovarian cancer that overexpresses HER-2/neu or epithelial ovarian cancer Hum Gene Ther 1998 9 1775 1798 9721088
Buhler L Friedman T Iacomini J Cooper DK Xenotransplantation – state of the art – update 1999 Front Biosci 1999 4 D416 432 10209058
Jager U Takeuchi Y Porter CD Sensitization of human cells to lysis by human complement as an approach for cancer gene therapy Adv Exp Med Biol 1998 451 359 363 10026897
Link CJ JrSeregina T Atchison R Hall A Muldoon R Levy JP Eliciting hyperacute xenograft response to treat human cancer: alpha(1,3) galactosyltransferase gene therapy Anticancer Res 1998 18 2301 2308 9703870
Unfer RC Hellrung D Link CJ Jr Immunity to the alpha(1,3)galactosyl epitope provides protection in mice challenged with colon cancer cells expressing alpha(1,3)galactosyl-transferase: a novel suicide gene for cancer gene therapy Cancer Res 2003 63 987 993 12615713
Aubert M Crotte C Bernard JP Lombardo D Sadoulet MO Mas E Decrease of human pancreatic cancer cell tumorigenicity by alpha1,3galactosyltransferase gene transfer Int J Cancer 2003 107 910 918 14601050 10.1002/ijc.11470
Dachs GU Dougherty GJ Stratford IJ Chaplin DJ Targeting gene therapy to cancer: a review Oncol Res 1997 9 313 325 9406237
Ring CJ Blouin P Martin LA Hurst HC Lemoine NR Use of transcriptional regulatory elements of the MUC1 and ERBB2 genes to drive tumour-selective expression of a prodrug activating enzyme Gene Ther 1997 4 1045 1052 9415310 10.1038/sj.gt.3300510
Manley PW Martiny-Baron G Schlaeppi JM Wood JM Therapies directed at vascular endothelial growth factor Expert Opin Investig Drugs 2002 11 1715 1736 12457433
Slamon DJ Clark GM Wong SG Levin WJ Ullrich A McGuire WL Human breast cancer: correlation of relapse and survival with amplification of the HER-2/neu oncogene Science 1987 235 177 182 3798106
Kraus MH Popescu NC Amsbaugh SC King CR Overexpression of the EGF receptor-related proto-oncogene erbB-2 in human mammary tumor cell lines by different molecular mechanisms Embo J 1987 6 605 610 3034598
Menard S Pupa SM Campiglio M Tagliabue E Biologic and therapeutic role of HER2 in cancer Oncogene 2003 22 6570 6578 14528282 10.1038/sj.onc.1206779
Brennan PJ Kumagai T Berezov A Murali R Greene MI Kumogai T HER2/neu: mechanisms of dimerization/oligomerization Oncogene 2000 19 6093 6101 11156522 10.1038/sj.onc.1203967
Leone F Perissinotto E Cavalloni G Fonsato V Bruno S Surrenti N Hong D Capaldi A Geuna M Piacibello W Expression of the c-ErbB-2/HER2 proto-oncogene in normal hematopoietic cells J Leukoc Biol 2003 74 593 601 12960261 10.1189/jlb.0203068
Slamon DJ Godolphin W Jones LA Holt JA Wong SG Keith DE Levin WJ Stuart SG Udove J Ullrich A Studies of the HER-2/neu proto-oncogene in human breast and ovarian cancer Science 1989 244 707 712 2470152
Li BD Harlow SP Budnick RM Sheedy DL Stewart CC Detection of HER-2/neu oncogene amplification in flow cytometry-sorted breast ductal cells by competitive polymerase chain reaction Cancer 1994 73 2771 2778 7514954
Lyon E Millson A Lowery MC Woods R Wittwer CT Quantification of HER2/neu gene amplification by competitive pcr using fluorescent melting curve analysis Clin Chem 2001 47 844 851 11325887
Wang SC Hung MC Transcriptional targeting of the HER-2/neu oncogene Drugs Today (Barc) 2000 36 835 843 12845342
Coussens L Yang-Feng TL Liao YC Chen E Gray A McGrath J Seeburg PH Libermann TA Schlessinger J Francke U Tyrosine kinase receptor with extensive homology to EGF receptor shares chromosomal location with neu oncogene Science 1985 230 1132 1139 2999974
Klapper LN Glathe S Vaisman N Hynes NE Andrews GC Sela M Yarden Y The ErbB-2/HER2 oncoprotein of human carcinomas may function solely as a shared coreceptor for multiple stroma-derived growth factors Proc Natl Acad Sci USA 1999 96 4995 5000 10220407 10.1073/pnas.96.9.4995
Holbro T Civenni G Hynes NE The ErbB receptors and their role in cancer progression Exp Cell Res 2003 284 99 110 12648469 10.1016/S0014-4827(02)00099-X
Yu L Kamo S Tagawa M Identification of a minimal c-erbB-2 promoter region that mediates preferential expression of a linked foreign gene in human breast cancer cells Int J Oncol 2002 20 607 610 11836576
Yu SF von Ruden T Kantoff PW Garber C Seiberg M Ruther U Anderson WF Wagner EF Gilboa E Self-inactivating retroviral vectors designed for transfer of whole genes into mammalian cells Proc Natl Acad Sci USA 1986 83 3194 3198 3458176
Lefebvre JC Giordanengo V Limouse M Doglio A Cucchiarini M Monpoux F Mariani R Peyron JF Altered glycosylation of leukosialin, CD43, in HIV-1-infected cells of the CEM line J Exp Med 1994 180 1609 1617 7964449 10.1084/jem.180.5.1609
Giordanengo V Bannwarth S Laffont C Van-Miegem V Harduin-Lepers A Delannoy P Lefebvre JC Cloning and expression of cDNA for a human Gal(b1-3)GalNAc a2,3-sialyltransferase from the CEM-T cell line Eur J Biochem 1997 247 558 566 9266697 10.1111/j.1432-1033.1997.00558.x
Wright M Grim J Deshane J Kim M Strong TV Siegal GP Curiel DT An intracellular anti-erbB-2 single-chain antibody is specifically cytotoxic to human breast carcinoma cells overexpressing erbB-2 Gene Ther 1997 4 317 322 9176517 10.1038/sj.gt.3300372
Brodowicz T Wiltschke C Budinsky AC Krainer M Steger GG Zielinski CC Soluble HER-2/neu neutralizes biologic effects of anti-HER-2/neu antibody on breast cancer cells in vitro Int J Cancer 1997 73 875 879 9399669 10.1002/(SICI)1097-0215(19971210)73:6<875::AID-IJC19>3.0.CO;2-3
Kumar R Mandal M Lipton A Harvey H Thompson CB Overexpression of HER2 modulates bcl-2, bcl-XL, and tamoxifen-induced apoptosis in human MCF-7 breast cancer cells Clin Cancer Res 1996 2 1215 1219 9816290
Kurokawa H Lenferink AE Simpson JF Pisacane PI Sliwkowski MX Forbes JT Arteaga CL Inhibition of HER2/neu (erbB-2) and mitogen-activated protein kinases enhances tamoxifen action against HER2-overexpressing, tamoxifen-resistant breast cancer cells Cancer Res 2000 60 5887 5894 11059787
Child SJ Miller MK Geballe AP Cell type-dependent and -independent control of HER-2/neu translation Int J Biochem Cell Biol 1999 31 201 213 10216954 10.1016/S1357-2725(98)00068-5
Graham FL Smiley J Russell WC Nairn R Characteristics of a human cell line transformed by DNA from human adenovirus type 5 J Gen Virol 1977 36 59 74 886304
Quinlan MP Expression of antisense E1A in 293 cells results in altered cell morphologies and cessation of proliferation Oncogene 1993 8 257 265 8426736
Kamei D Murakami M Nakatani Y Ishikawa Y Ishii T Kudo I Potential role of microsomal prostaglandin E synthase-1 in tumorigenesis J Biol Chem 2003 278 19396 19405 12626523 10.1074/jbc.M213290200
Springer CJ Niculescu-Duvaz I Prodrug-activating systems in suicide gene therapy J Clin Invest 2000 105 1161 1167 10791987
Galili U Rachmilewitz EA Peleg A Flechner I A unique natural human IgG antibody with anti-alpha-galactosyl specificity J Exp Med 1984 160 1519 1531 6491603 10.1084/jem.160.5.1519
Good AH Cooper DK Malcolm AJ Ippolito RM Koren E Neethling FA Ye Y Zuhdi N Lamontagne LR Identification of carbohydrate structures that bind human antiporcine antibodies: implications for discordant xenografting in humans Transplant Proc 1992 24 559 562 1566430
Thall AD Murphy HS Lowe JB alpha 1,3-Galactosyltransferase-deficient mice produce naturally occurring cytotoxic anti-Gal antibodies Transplant Proc 1996 28 556 557 8623267
Salvaris E Gock H Han W Murray-Segal L Barlow H Mottram P Pearse M Cowan P Goodman D d'Apice AJ Naturally acquired anti-alpha Gal antibodies in a murine allograft model similar to delayed xenograft rejection Xenotransplantation 2000 7 42 47 10809056 10.1034/j.1399-3089.2000.00040.x
Sawada T Yamada O Yoshimura N Hatori K Fuchinoue S Teraoka S Xenoantigen, an alphaGal epitope-expression construct driven by the hTERT-promoter, specifically kills human pancreatic cancer cell line Cancer Cell Int 2002 2 14 20 12392598 10.1186/1475-2867-2-14
Arteaga CL Chinratanalab W Carter MB Inhibitors of HER2/neu (erbB-2) signal transduction Semin Oncol 2001 28 30 35 11774203 10.1053/sonc.2001.29722
Hurst HC Update on HER-2 as a target for cancer therapy: the ERBB2 promoter and its exploitation for cancer treatment Breast Cancer Res 2001 3 395 398 11737892 10.1186/bcr329
Scott GK Chang CH Erny KM Xu F Fredericks WJ Rauscher FJ 3rdThor AD Benz CC Ets regulation of the erbB2 promoter Oncogene 2000 19 6490 6502 11175365 10.1038/sj.onc.1204041
Xing X Wang SC Xia W Zou Y Shao R Kwong KY Yu Z Zhang S Miller S Huang L The ets protein PEA3 suppresses HER-2/neu overexpression and inhibits tumorigenesis Nat Med 2000 6 189 195 10655108 10.1038/72294
Yu D Suen TC Yan DH Chang LS Hung MC Transcriptional repression of the neu protooncogene by the adenovirus 5 E1A gene products Proc Natl Acad Sci USA 1990 87 4499 4503 1972274
Chang JY Xia W Shao R Sorgi F Hortobagyi GN Huang L Hung MC The tumor suppression activity of E1A in HER-2/neu-overexpressing breast cancer Oncogene 1997 14 561 568 9053854 10.1038/sj.onc.1200861
Yu D Hung MC Overexpression of ErbB2 in cancer and ErbB2-targeting strategies Oncogene 2000 19 6115 6121 11156524 10.1038/sj.onc.1203972
Beckhove P Schutz F Diel IJ Solomayer EF Bastert G Foerster J Feuerer M Bai L Sinn HP Umansky V Efficient engraftment of human primary breast cancer transplants in nonconditioned NOD/Scid mice Int J Cancer 2003 105 444 453 12712433 10.1002/ijc.11125
Thall AD Maly P Lowe JB Oocyte Gal alpha 1,3Gal epitopes implicated in sperm adhesion to the zona pellucida glycoprotein ZP3 are not required for fertilization in the mouse J Biol Chem 1995 270 21437 21440 7545161 10.1074/jbc.270.37.21437
Tearle RG Tange MJ Zannettino ZL Katerelos M Shinkel TA Van Denderen BJ Lonie AJ Lyons I Nottle MB Cox T The alpha-1,3-galactosyltransferase knockout mouse. Implications for xenotransplantation Transplantation 1996 61 13 19 8560551 10.1097/00007890-199601150-00004
| 15987454 | PMC1175063 | CC BY | 2021-01-04 16:04:33 | no | Breast Cancer Res. 2005 May 3; 7(4):R487-R494 | utf-8 | Breast Cancer Res | 2,005 | 10.1186/bcr1034 | oa_comm |
==== Front
Breast Cancer ResBreast Cancer Research1465-54111465-542XBioMed Central London bcr10351598746010.1186/bcr1035Research ArticleMenopausal hormone therapy after breast cancer: a meta-analysis and critical appraisal of the evidence Col Nananda F [email protected] Jung A [email protected] Rowan T [email protected] Brown Medical School and Harvard University, Providence, Rhode Island, USA2 Department of Nursing, Hanyang University, Seoul, Korea3 Harbor-UCLA Research and Education Institute, Torrance, California, USA2005 19 5 2005 7 4 R535 R540 3 2 2005 18 3 2005 6 4 2005 8 4 2005 Copyright © 2005 Col et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Introduction
Menopausal hormone therapy (HT) is typically withheld from breast cancer survivors because of concerns about risk for recurrence. Our objectives were to estimate the effects of HT on recurrence in breast cancer survivors and to examine the reliability of these estimates.
Methods
In a systematic review of the literature we identified all reports of HT use in breast cancer survivors that included comparison groups. Study design features that might affect selection of participants, detection of recurrence, and manuscript publication were assessed. The relative risks for breast cancer recurrence associated with HT were combined with random effects models.
Results
Two randomized and eight observational studies included 1,316 breast cancer survivors who used HT and 2,839 nonusers. In the observational studies, HT users were younger and more commonly node negative; only two reported balanced restaging for HT and control groups. Randomized trials suggest that HT increased the risk for recurrence (relative risk 3.41, 95% confidence interval 1.59–7.33), whereas observational studies suggest that HT decreased this risk (relative risk 0.64, 95% confidence interval 0.50–0.82).
Conclusion
Results from observational studies of HT conducted in breast cancer survivors are discrepant with results from randomized trials. Observational studies of HT use in breast cancer survivors have design limitations that cannot be controlled for using standard statistical methods. Therefore, the randomized clinical trial data provide the only reliable estimates of the effect of HT use on recurrence risks in breast cancer survivors.
See related commentary by Colditz:
==== Body
Introduction
Most breast cancer survivors are menopausal either at diagnosis or as a result of premature therapy-induced menopause, and they frequently experience climacteric symptoms [1]. Menopausal hormone therapy (HT), either with estrogen alone or with combined estrogen and progestin, relieves estrogen deficiency symptoms [2] but it is commonly withheld from women with diagnosed breast cancer because of concerns regarding an increased risk for recurrence [3].
The available data from observational studies indicate that use of HT is associated with increased risk for breast cancer [4]. In postmenopausal women, the randomized Women's Health Initiative HT trials found an increased risk for breast cancer with estrogen plus progestin [5] but not with unopposed estrogen [6]. An apparent reduction in risk seen during the first 2 years of combination HT was attributed to a masking of breast cancer detection, with a higher risk for more advanced breast cancers subsequently [5]. In breast cancer survivors, observational studies have consistently reported similar or lower risks for recurrence among women using HT as compared with nonusers [7], albeit with methodological weaknesses [8]; this has been interpreted as evidence of the safety or perhaps benefit of HT in women with breast cancer. However, the first large randomized trial in this population reported that HT significantly increased the risk for recurrence [9].
The objectives of this meta-analysis were to estimate the impact HT has on recurrence risk among observational and randomized studies, and to examine the reliability of these estimates.
Materials and methods
A previous Medline search from 1966 to 1999 [7] was updated to February 2004 using the medical subject headings 'breast neoplasm', 'neoplasm recurrence', 'estrogens', 'estrogen replacement therapy', 'hormone replacement therapy', and 'estradiol', and reference lists of abstracted manuscript and protocols were reviewed. Only studies that included women with invasive breast cancer who received oral HT, that had an explicitly defined comparison group, and that reported breast cancer recurrences were included. Studies that reported overlapping or redundant data were excluded [10-16], as were those that did not adequately describe the selection or composition of control groups [17,18] or that included only topical hormones [19].
Two of the authors (NFC and JAK) independently abstracted data on the following variables: sample size, age at diagnosis and at trial induction, tumor stage, nodal status, estrogen and progesterone receptor status, disease-free interval (DFI) between initial breast cancer diagnosis and initiation of HT, type and duration of HT used, follow up after initiation of HT, and number and timing of breast cancer recurrences.
Each study was systematically reviewed for features that could introduce bias, including procedures for identifying participants, whether institutional review board approval and/or informed consent was obtained, whether risk factors for recurrence were similar at diagnosis, and whether restaging before entry (to exclude metastatic disease) and duration of follow up were similar for HT users and nonusers. Observational studies were classified as 'clinical experiences' if one or more study authors provided health care to the cohort with potential participation in the decision to use HT.
When not reported, the follow up after HT initiation was assumed to equal the duration of HT use. Any second breast cancer event (local, regional, or distant recurrence or invasive cancer in either breast) was treated as a recurrence because studies did not consistently make these distinctions.
Relative risk (RR) and 95% confidence interval (CI) were calculated for each study for the recurrence rate and mortality rate among HT users and nonusers. A random effects model was used to estimate the combined RR for randomized and observational studies using Meta-Analyst [13].
Results
Ten studies were identified, including a total of 1,316 breast cancer survivors who used HT and 2,839 who did not. Of these 10 studies, two were unblinded randomized controlled trials without placebo arms [9,20], one began as a randomized trial but was reported as an observational study and is considered as such here, and seven were observational studies.
Summary of randomized trials
Both randomized trials were conducted in Europe (one in England and one in Sweden). They involved a total of 445 patients with a mean age of 55.5 years, a mean DFI of 33.2 months, a duration of HT use of 19.9 months, and a mean follow-up period after HT initiation of 25.2 months (Table 1). A total of 36 recurrences and nine deaths occurred during this time in these trials; the pooled RR for the two randomized trials was 3.41 (95% CI 1.59–7.33).
Summary of observational studies
Of the eight observational studies, six were clinical experiences [22-27]. The eight studies involved a total of 3710 patients with a mean age of 59.7 years, a mean DFI of 49.2 months, a duration of HT use of 28 months, and a mean follow-up period after HT initiation of 57.1 months (Table 1). A combined total of 552 recurrences (109 among HT users) and 460 deaths (51 among HT users) occurred in these trials. The pooled RR for the observational studies was 0.64 (95% CI 0.50–0.82).
All studies
Most studies included both combination HT and unopposed estrogens without stratifying risk estimates according to preparation. Three of the observational studies [22,24,25] reported obtaining informed consent but only from women who used HT. Three studies [20,24,26] reported similar restaging for treatment and control groups at the beginning of the observation period, although one of these [26] did not report whether those found to have occult metastasis were excluded. Not all studies reported the DFI for the control groups, but several reported matching control individuals according to DFI [22,27]. Prognostic factors for HT users and nonusers differed in most studies (Table 1). On average, HT users were more than 3 years younger than nonusers and were more likely to be node negative. The average duration of HT use was 26.6 months, with an average duration of follow up after initiation of HT of 53 months. The mean DFI was 36.9 months for HT users and 55.6 for nonusers.
Among the 1,191 HT users in nine studies reporting recurrences, 137 (11.7%) experienced a recurrence of their breast cancer during follow up. Among the 2,477 nonusers in these studies, 451 (18.2%) had a recurrence. The average annual recurrence rate was 3.3% (range 0.6–7.1%), with substantially higher rates in the randomized trials. Combining all studies yielded a RR for recurrence of 0.84 (95% CI 0.54–1.3; Fig. 1), with statistically significant heterogeneity (Q = 25.3).
Discussion
Estimates from observational studies of HT among breast cancer survivors suggest that HT prevents breast cancer recurrence, whereas estimates from randomized trials suggest the opposite. Because of statistically significant heterogeneity, these estimates should not be combined. Although all of the trials included in our analyses contained methodological weaknesses, the nonrandomized studies had design features that could introduce selection, reporting, and/or publication biases. The selection of healthier women to begin HT, the benefit of restaging before initiation of HT, the short duration of HT exposure and follow up, the potential effects of HT on mammograms that could obscure the diagnosis of recurrent or new breast cancers, and publication bias favoring publication and/or completion of studies reporting a protective effect of HT could explain the apparent protective effect of short-term HT on recurrence among breast cancer survivors in these studies.
Systematic serial restaging with blood tests and imaging during follow up is no longer generally recommended. However, their use detects breast cancer recurrence earlier. Balanced restaging was defined in only two out of seven observational studies. If breast cancer survivors contemplating HT use were more likely to have restaging, then the imbalance could account for the apparent protective effect of HT in observational studies. Although the description of prognostic factors was rarely complete, HT users in observational studies were younger and had more favorable prognostic profiles than did control individuals. This process also selected women with severe vasomotor symptoms, who have lower estradiol and testosterone levels; higher levels of these hormones have been associated with increased breast cancer risk. As a result, it is possible that women who were more likely to be offered HT [20] had lower recurrence risks. It is important to note that the majority of observational studies included in these analyses were not designed as observational studies from the start but rather as clinical experiences. Had these observational studies been more rigorously designed, using modern epidemiological techniques, many of these biases could have been minimized.
The adverse effect of combined HT on mammographic breast cancer detection [5] might have affected recurrence detection. Both recurrent and new breast cancers, which account for 10–20% of cancer events in women with prior lumpectomy, could have falsely appeared lower in HT users because of HT-related interference with mammographic diagnosis. However, this factor is probably not large, given the sharp increase in risk observed even after short-term HT use in randomized trials [36] and that the increase in risk pertained to distant as well as local recurrences.
The randomized trial reported by Holmberg and colleagues [9] overcomes many of the shortcomings of observational studies and provides the best available data on the impact of HT in breast cancer survivors. Although their unblinded design and lack of a placebo group could result in selective attrition, follow-up rates were comparable among HT users and nonusers. These investigators also reported summary interim analyses of a similar randomized trial, the Stockholm trial, with a relative hazard ratio of 0.82 (95% CI 0.35–1.9). This trial was not included in this analysis because its findings have not yet been reported in full; the reasons for its discrepant findings are unclear at this time.
Conclusion
Observational studies of HT use in breast cancer survivors have design limitations that cannot be controlled for using standard statistical methods and hence should be considered essentially uninformative with respect to the safety of HT use in breast cancer survivors. Only randomized clinical trials are likely to provide reliable estimates of the effect of HT use in this setting.
Abbreviations
CI = confidence interval; DFI = disease-free interval; HT = hormone therapy; RR = relative risk.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
NC conceived the study (with RC), designed the study, reviewed the source studies, abstracted data, drafted the paper, and supervised the statistical analyses. JK participated in the design of the study and reviewed the source studies, abstracted data, carried out the meta-analysis, and helped to draft the manuscript. RC conceived of the study (with NC), designed the analysis, participated in its coordination, and helped to draft the manuscript. All authors read and approved the final manuscript.
Acknowledgements
This work was supported, in part, by the Agency for Healthcare Quality (RO1 HS01332901), American Cancer Society Breast Cancer Prevention Forum, and the Robert Wood Johnson Foundation Generalist Physician Faculty Scholars Award (#033958).
This work was presented at the 24th Annual Meeting of the Society for Medical Decision Making in Baltimore (MD, USA) in October 2002.
Figures and Tables
Figure 1 Relative risks for recurrent breast cancer associated with hormone therapy (HT). Each black circle indicates the relative risk for recurrent breast cancer; the horizontal lines indicate the 95% confidence interval (CI). The top portion of the figure describes randomized controlled trials, the middle portion describes observational studies, and the bottom portion describes all trials combined.
Table 1 Characteristics of 1316 users and 2839 nonusers of hormone therapy
Study Treatment n Mean age (years) Stage Nodal status ER status PgR status Mean DFI before HT (months) Estrogen alone (%) Mean duration of HT (months) Mean follow-up after HT (months) Recurrences (n) Deaths, all cause (n) Deaths, primary tumor (n)
Randomized trials
Marsden et al. (2000; n = 100) [20] HT 51a 58b NR NR NR NR 40b NR 6 NR 2 NR NR
No HT 49a 55b NR NR NR NR 36b NR 1 NR NR
Holmberg et al. (2004; n = 345) [9] HT 174 55.5 NR 25.9% (38) positive 86 positivec NR 31.2b NR 24 25.2b 26 5 3
No HT 171 55.0 NR 21.4% (31) positive 73 positivec NR 32.4b 25.2b 7 4 4
Observational studies
Ursic-Vrscaj and Bebar (1999; n = 63) [27] HT 21d 47b 1 G1
10 G2
7 G3 14 negative, 7 positive 5 positive, 16 negative 8 positive, 13 negative 62 4.8 28 38g 4 0g 0
No HT 42d 48.2 7 G1
17 G2
11 G3 28 negative 14 positive 18 positive, 22 negative 22 positive, 18 negative NR 38g 5 1g 1
DiSaia et al. (2000; n = 487) [22] HT 125 55.7 17 DCIS
52 stage I
27 stage II
10 stage III
1 stage IV NR NR NR 46b 28 22b 92.1g NR 4g NR
No HT 362 55.9 NR NR NR NR NR 90.6g NR 57g NR
O'Meara et al. (2001; n = 869) [36] HT 174d 63.6e 91 stage I
51 stage II
20 stage I/II
10 stage III
2 stage II/III 128 negative, 31 positive 84 positive, 39 negative 71 positive, 45 negative 47.7e 79 15b 44.4b,f 16 17 5
No HT 695d 63.6e 403 stage I
246 stage II
3 stage I/II
42 stage III
1 stage II/III 470 negative, 175 positive 409 positive, 137 negative 311 positive, 206 negative 47.7e 44.4b,f 101 115 59
Beckmann et al. (2001; n = 185) [24] HT 64 NA 37 T1
19 T2
8 T3/4 44 negative, 20 positive 31 positive, 33 negative 34 positive, 30 negative 0 NA 33b 37b 6 4 NR
No HT 121 NA 62 T1
42 T2
17 T3/4 76 negative, 45 positive 48 positive, 73 negative 48 positive, 73 negative 0 42b 17 15 NR
Marttunnen et al. (2001; n = 131) [26] HT 88 53.4 3 DCIS
67 T1
17 T2
1 T3 72 negative, 10 positive 57 positive, 15 negative 54 positiveg, 13 negativeg 50.4 38.6 30 30 7 2 2
No HT 43 52.8 1 DCIS
29 T1
11 T2
2 T3 30 negative, 13 positive 29 positive, 9 negative 30 positiveg, 7 negativeg 50.4 31.2 5 3 3
Durna et al. (2002; n = 1122) [23] HT 286 56.8b 180 stage I
64 stage II
22 stage III/IV NA NR NR 12b 5.9 21b 69.6b 44 16 13
No HT 836 64.7b 470 stage I
191 stage II
120 stage III/IV NA NR NR NR 61.2b 247 199 122
Vassilopoulou-Sellin et al. (2002; n = 299) [21] HT 56h 56b 9 <1 cm
30 1–2.5 cm
15 >2.5 cm 35 negative, 13 1–3, 6 >3 37 negative NR 105.6 100 30 >5 years, 20 2–5 years, 6 2 years 71 2 1 0
No HT 243h 53b 38 <1 cm
134 1–2.5 cm
67 >2.5 cm 133 negative, 70 1–3, 33 >3 164 negative NR 99.6 NR 33 2 1
Decker et al. (2003; n = 554) [25] HT 277 57.4b 84 DCIS
124 stage I
47 stage IIA
19 stage IIB
3 stage IIIA NR 100 positive, 54 negative 63 positive, 46 negative 43.3 48.7 44.4 49.7 30 7 5
No HT 277 59.0b 84 DCIS
124 stage I
47 stage IIA
19 stage IIB
3 stage IIIA NR 121 positive, 35 negative 73 positive, 42 negative NR 45.6 35 17 9
Summary
Randomized trials HT 225 56.07 38 positive 86 positive 33.19 19.92 25.20 28 5 3
No HT 220 55.00 31 positive 73 positive 33.20 25.20 8 4 4
Observational studies HT 1091 56.98 293 negative, 87 positive 277 positive, 194 negative 230 positive, 147 negative 37.70 40.4 28.02 57.46 109 51 25
No HT 2619 60.87 737 negative, 350 positive 625 positive, 440 negative 484 positive, 346 negative 54.01 57.02 443 409 195
All combined HT 1316 56.82 293 negative, 125 positive 363 positive, 194 negative 230 positive, 147 negative 36.93 40.4 26.58 53.03 137 56 28
No HT 2839 60.39 737 negative, 381 positive 698 positive, 440 negative 484 positive, 346 negative 50.55 54.88 451 413 199
aExcluding stage III/IV patients. bMedian value. cRefers to hormone receptor status; specific data concerning estrogen receptor (ER) and progesterone receptor (PgR) status were not reported. dExcluding patients with ductal carcinoma in situ (DCIS). eWeighted mean. fFor recurrence only; follow-up for mortality was 55.2 months. gPersonal communication. hExcluding DCIS, stages III and IV, and ER-positive patients. DFI, disease-free interval; HT, hormone therapy; NA, not able to calculate; NR, not reported.
==== Refs
Carpenter J Johnson D Wagner L Andrykowski M Hot flashes and related outcomes in breast cancer survivors and matched comparison women Nurs Forum 2002 29 E16 E25
MacLennan A Lester S Moore V Oral estrogen replacement therapy versus placebo for hot flushes: a systematic review Climacteric 2001 4 58 74 11379379
Pritchard K Khan H Levine M Steering Committee on Clinical Practice Guidelines for the Care and Treatment of Breast Cancer Clinical practice guidelines for the care and treatment of breast cancer: The role of hormone replacement therapy in women with a previous diagnosis of breast cancer CMAJ 2002 166 1017 1022 12002977
Collaborative Group on Hormonal Factors in Breast Cancer Breast cancer and hormone replacement therapy: collaborative reanalysis of data from 51 epidemiological studies of 52,705 women with breast cancer and 108,411 without breast cancer Lancet 1997 350 1047 1059 10213546 10.1016/S0140-6736(97)08233-0
Chlebowski RT Hendrix SL Langer RD Stefanick ML Gass M Lane D Rodabough RJ Gilligan MA Cyr MG Thomson CA Influence of estrogen plus progestin on breast cancer and mammography in healthy postmenopausal women: the Women's Health Initiative Randomized Trial JAMA 2003 289 3243 3253 12824205 10.1001/jama.289.24.3243
Alving B NIH asks participants in Women's Health Initiative estrogen-alone study to stop study pills, begin follow-up phase NIH News 2004 (last accessed 19 April 2005).
Col NF Hirota LK Orr RK Erban JK Wong JB Lau J Hormone replacement therapy after breast cancer: a systematic review and quantitative assessment of risk J Clin Oncol 2001 19 2357 2363 11304788
Chlebowski R Col N Menopausal hormone therapy after breast cancer Lancet 2004 363 410 411 14962515 10.1016/S0140-6736(04)15519-0
Holmberg L Anderson H for the HABITS steering and data monitoring committees HABITS (hormonal replacement therapy after breast cancer – is it safe?), a randomised comparison: trial stopped Lancet 2004 363 453 455 14962527 10.1016/S0140-6736(04)15493-7
Eden JA Wren BG Hormone replacement therapy after breast cancer: a review Cancer Treat Rev 1996 22 335 343 9118119 10.1016/S0305-7372(96)90006-7
Eden JA Bush T Nand S Wren BG A case–control study of combined continuous estrogen-progestin replacement therapy among women with a personal history of breast cancer Menopause 1995 2 67 72
Dew J Eden J Beller E Magarey C Schwartz P Crea P Wren B A cohort study of hormone replacement therapy given to women previously treated for breast cancer Climacteric 1998 1 137 142 11907916
Lau J Metaanalyst Boston, MA, USA
Vassilopoulou-Sellin R Asmar L Hortobagyi GN Klein MJ McNeese M Singletary SE Theriault RL Estrogen replacement therapy after localized breast cancer: clinical outcome of 319 women followed prospectively J Clin Oncol 1999 17 1482 1487 10334534
Beckmann MW Jap D Djahansouzi S Nestle-Kramling C Kuschel B Dall P Brumm C Bender HG Hormone replacement therapy after treatment of breast cancer: effects on postmenopausal symptoms, bone mineral density and recurrence rates Oncology 2001 60 199 206 11340370 10.1159/000055319
Durna EM Crowe SM Leader LR Eden JA Quality of life of breast cancer survivors: the impact of hormonal replacement therapy Climacteric 2002 5 266 276 12419085
Natrajan P Soumakis K Gambrell R Jr Estrogen replacement therapy in women with previous breast cancer Am J Obstet Gynecol 1999 181 288 295 10454671
Natrajan P Gambrell R Estrogen replacement therapy in patients with early breast cancer Am J Obstet Gynecol 2002 187 289 294 12193914 10.1067/mob.2002.125999
Dew J Wren B Eden J A cohort of topical vaginal estrogen therapy in women previously treated for breast cancer Climacteric 2003 6 45 52 12725664
Marsden J Whitehead M A'Hern R Baum M Sacks N Are randomized trials of hormone replacement therapy in symptomatic women with breast cancer feasible? Fertil Steril 2000 73 292 299 10685532 10.1016/S0015-0282(99)00510-5
Vassilopoulou-Sellin R Cohen DS Hortobagyi GN Klein MJ McNeese M Singletary SE Smith TL Theriault RL Estrogen replacement therapy for menopausal women with a history of breast carcinoma: results of a 5-year, prospective study Cancer 2002 95 1817 1826 12404273 10.1002/cncr.10913
DiSaia P Brewster W Ziogas A Anton-Culver H Breast cancer survival and hormone replacement therapy: a cohort analysis Am J Clin Oncol 2000 23 541 545 11202792 10.1097/00000421-200012000-00001
Durna E Wren B Heller G Leader L Sjoblom P Eden J Hormone replacement therapy after a diagnosis of breast cancer: cancer recurrence and mortality Med J Aust 2002 177 347 351 12358575
Beckmann MW Jap D Djahansouzi S Nestle-Kramling C Kuschel B Dall P Brumm C Bender HG Hormone replacement therapy after treatment of breast cancer: effects on postmenopausal symptoms, bone mineral density and recurrence rates Oncology 2001 60 199 206 11340370 10.1159/000055319
Decker DA Pettinga JE VanderVelde N Huang RR Kestin L Burdakin JH Estrogen replacement therapy in breast cancer survivors: a matched-controlled series Menopause 2003 10 277 285 12851510 10.1097/01.GME.0000061806.76067.E9
Marttunnen M Hietanen P Pyrhonen S Tiitinen A Ylikorkala O A prospective study on women with a history of breast cancer and with or without estrogen replacement therapy Maturitas 2001 39 217 225 11574181 10.1016/S0378-5122(01)00211-0
Ursic-Vrscaj M Bebar S A case-control study of hormone replacement therapy after primary surgical breast cancer treatment Eur J Surg Oncol 1999 25 146 151 10218456 10.1053/ejso.1998.0617
Smith TJ Davidson NE Schapira DV Grunfeld E Muss HB Vogel VG IIISomerfield MR American Society of Clinical Oncology 1998 update of recommended breast cancer surveillance guidelines J Clin Oncol 1999 17 1080 1082 10071303
Temple LK Wang EE McLeod RS Preventive health care, 1999 update: 3. Follow-up after breast cancer. Canadian Task Force on Preventive Health Care CMAJ 1999 161 1001 1008 10551200
American Society of Clinical Oncology Clinical practice guidelines for the use of tumor markers in breast and colorectal cancer J Clin Oncol 1996 14 2843 2877 8874347
Guthrie J Dennerstein L Hopper J Burger H Hot flushes, menstrual status, and hormone levels in a population-based sample of midlife women Obstet Gynecol 1996 88 437 442 8752255 10.1016/0029-7844(96)00196-2
Overlie I Moen M Holte A Finset A Androgens and estrogens in relation to hot flashes during the menopausal transition Maturitas 2002 41 69 77 11809345 10.1016/S0378-5122(01)00256-0
Cauley JA Lucas FL Kuller LH Stone K Browner W Cummings SR Elevated serum estradiol and testosterone concentrations are associated with a high risk for breast cancer. Study of Osteoporotic Fractures Research Group Ann Intern Med 1999 130 270 277 10068384
Endogenous Hormones and Breast Cancer Collaborative Group Endogenous sex hormones and breast cancer in postmenopausal women: reanalysis of nine prospective studies J Natl Cancer Inst 2002 94 606 616 11959894
Writing Group for the Women's Health Initiative Investigators Risks and benefits of estrogen plus progestin in healthy postmenopausal women: principal results from the Women's Health Initiative randomized controlled trial JAMA 2002 288 321 333 12117397 10.1001/jama.288.3.321
O'Meara ES Rossing MA Daling JR Elmore JG Barlow WE Weiss NS Hormone replacement therapy after a diagnosis of breast cancer in relation to recurrence and mortality J Natl Cancer Inst 2001 93 754 762 11353785 10.1093/jnci/93.10.754
| 15987460 | PMC1175064 | CC BY | 2021-01-04 16:54:50 | no | Breast Cancer Res. 2005 May 19; 7(4):R535-R540 | utf-8 | Breast Cancer Res | 2,005 | 10.1186/bcr1035 | oa_comm |
==== Front
Breast Cancer ResBreast Cancer Research1465-54111465-542XBioMed Central London bcr10371598745510.1186/bcr1037Research ArticleMutation analysis of the ATR gene in breast and ovarian cancer families Heikkinen Katri [email protected] Virpi [email protected] Sanna-Maria [email protected] Katrin [email protected] Robert [email protected] Department of Clinical Genetics, Oulu University Hospital/University of Oulu, Oulu, Finland2005 6 5 2005 7 4 R495 R501 24 2 2005 18 3 2005 8 4 2005 11 4 2005 Copyright © 2005 Heikkinen et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Introduction
Mutations in BRCA1, BRCA2, ATM, TP53, CHK2 and PTEN account for only 20–30% of the familial aggregation of breast cancer, which suggests the involvement of additional susceptibility genes. The ATR (ataxia-telangiectasia- and Rad3-related) kinase is essential for the maintenance of genomic integrity. It functions both in parallel and cooperatively with ATM, but whereas ATM is primarily activated by DNA double-strand breaks induced by ionizing radiation, ATR has been shown to respond to a much broader range of DNA damage. Upon activation, ATR phosphorylates several important tumor suppressors, including p53, BRCA1 and CHK1. Based on its central function in the DNA damage response, ATR is a plausible candidate gene for susceptibility to cancer.
Methods
We screened the entire coding region of the ATR gene for mutations in affected index cases from 126 Finnish families with breast and/or ovarian cancer, 75 of which were classified as high-risk and 51 as moderate-risk families, by using conformation sensitive gel electrophoresis and direct sequencing.
Results
A large number of novel sequence variants were identified, four of which – Glu254Gly, Ser1142Gly, IVS24-48G>A and IVS26+15C>T – were absent from the tested control individuals (n = 300). However, the segregation of these mutations with the cancer phenotype could not be confirmed, partly because of the lack of suitable DNA samples.
Conclusion
The present study does not support a major role for ATR mutations in hereditary susceptibility to breast and ovarian cancer.
==== Body
Introduction
Of all breast and ovarian cancers, 5–10% are due to genetic predisposition [1]. Mutations in the two high penetrance genes BRCA1 and BRCA2 are well known, but they account for only 20–30% of the familial aggregation of breast cancer. The remaining cases could be the result of a few additional, yet unidentified, high penetrance mutations, but the polygenic model may provide a more plausible explanation [2]. According to this model, genetic susceptibility to breast cancer is due to several loci, each conferring a modest independent risk [3]. Because the protein products of the genes thus far associated with breast and/or ovarian cancer predisposition are central players in the pathways involved in cell cycle checkpoint functions, and in the sensing, transduction and repair of DNA lesions [4], other similarly acting genes may represent new potential candidates.
The ATR (ataxia-telangiectasia- and Rad3-related) kinase is essential for the maintenance of genomic integrity. It is a key activator of the cellular responses to DNA lesions [5]. In response to DNA double-strand breaks induced by ionizing radiation ATR, along with ATR-interacting protein, acts in parallel with ATM (ataxia-telangiectasia mutated), which is defective in the neurodegenerative disorder ataxia-telangiectasia and is also associated with breast cancer susceptibility [6-9]. Whereas ATM is responsible for the immediate and rapid response to double-strand breaks, ATR joins in later and maintains the phosphorylated state of specific substrates. However, this is not the main role played by ATR; it also responds to ultraviolet-induced lesions, stalled replication forks and hypoxia [5]. In response to these events, ATR phosphorylates key proteins in various branches of the DNA damage response pathways, such as p53, BRCA1, CHK1 and Rad17, thereby activating DNA repair, cell cycle checkpoints, or apoptosis [10,11].
The cellular functions of ATR are indispensable, as demonstrated in mice, in which biallelic disruption of ATR leads to early embryonic lethality. In contrast, ATR+/- mice exhibit only a small decrease in survival but tumor incidence is increased [12]. In humans a connection between ATR defects and tumorigenesis has also been suggested, mainly by studies reporting somatic changes in ATR in gastric and endometrial cancers exhibiting microsatellite instability [13,14]. Consequently, it has been proposed that ATR serves as a haploinsufficient tumour suppressor on a mismatch repair deficient background [15].
Recently, inherited defects in ATR signalling were shown to associate with Seckel syndrome, because patients in two families were found to be homozygous for a hypomorphic ATR mutation. Seckel syndrome is a heterogenous recessive disorder that is characterized by dwarfism, developmental delay and severe microcephaly. It shares an overlap in clinical features with two recessive cancer susceptibility syndromes, Nijmegen breakage syndrome and Fanconi anemia (FA) [16,17]. Interestingly, the gene that is defective in two FA complementation groups, namely FA-B and FA-D1, has been identified as BRCA2 – a major breast cancer susceptibility gene [18,19]. In addition, carriers of the Nijmegen breakage syndrome Slavic founder mutation appear to be at increased risk for breast cancer [20]. Thus far, predisposition to cancer has not been reported in patients with Seckel syndrome. However, various cell lines in which ATR has been inactivated exhibit genetic instability, and this may predict proneness to cancer [17].
Based on this, we wanted to determine whether ATR germline mutations are involved in susceptibility to breast and/or ovarian cancer, and conducted a mutation analysis of all 47 coding exons and exon–intron boundaries in the affected index cases in 126 families.
Materials and methods
Subjects
The index cases of 126 breast and/or ovarian cancer families originating from northern Finland were screened for ATR germline mutations. Of the studied families, 94 were affected by breast, 29 by breast/ovarian and three by ovarian cancer. All index cases had been diagnosed with either breast or ovarian cancer. 75 of the families were classified as high-risk families and were defined as follows: three or more cases of breast and/or ovarian cancer in first- or second-degree relatives; or two cases of breast and/or ovarian cancer in first- or second-degree relatives, of which at least one had early disease onset (age ≤35 years), bilateral disease, or multiple primary tumours. Most of the high-risk families contained three or more cancer cases. The remaining 51 families contained two cases of breast and/or ovarian cancer in first- or second-degree relatives and were considered to be at moderate disease risk. All of the high-risk families had previously been screened for germline mutations in BRCA1, BRCA2, CHK2 and TP53 [21-23] and 10 families were known to have disease-related mutations in BRCA1 or BRCA2. The frequencies of all observed germline variants were assessed in either 100 or 300 control individuals, who were anonymous cancer-free blood donors originating from the same geographical region as the studied families.
All patients gave informed consent for acquisition of pedigree data and blood specimens for use in a study on cancer susceptibility gene mutations. Approval to perform the study was obtained from the Ethical Board of the Northern Ostrobothnia Health Care District and the Finnish Ministry of Social Affairs and Health.
Mutation analysis
DNA was extracted from blood lymphocytes using either the standard phenol–chloroform protocol or the Puregene D-50K purification kit (Gentra, Minneapolis, MN, USA). Screening of the protein encoding and exon–intron boundary regions of ATR was done by conformation sensitive gel electrophoresis (CSGE), which is a cost-efficient way to scan for mutations with high detection sensitivity and specificity [24,25], or by direct sequencing. Sequencing analysis was performed using the Li-Cor IR2 4200-S DNA Analysis system (Li-Cor Inc., Lincoln, NE, USA) and the SequiTherm EXCEL™II DNA Sequencing Kit-LC (Epicentre Technologies, Madison, WI, USA). Oligonucleotides for CSGE and sequencing (Table 1) were designed using Primer3 software [26], utilizing sequence information obtained from public databases. Polymerase chain reaction conditions for CSGE and sequencing are available upon request.
Statistical analyses
Fisher's exact test or χ2 test was used to determine statistical significance (SPSS version 12.0 for Windows; SPSS Inc., Chicago, IL, USA). All P values were two sided.
Results
Mutation analysis revealed several alterations in the ATR gene. Altogether, 23 nucleotide substitutions were observed: 17 in the exon and six in the intron regions (Tables 2 and 3). Eleven of the exonic changes resulted in amino acid substitutions, eight of which were novel and three were polymorphisms reported in the single nucleotide polymorphism database [27]. The location of the amino acid changes is summarized in Fig. 1. All observed nucleotide alterations were assessed for possible effects on splicing consensus sequences [28], and the coding sequence variants were tested using the ESEfinder program [29] to identify those that reduced the exonic splicing enhancer (ESE) score below the calculated threshold.
Of the observed amino acid substitutions, four were located in known functional sites: Arg2008Leu and Tyr2132Asp to the FAT (FRAP/ATM/TRRAP) domain, and Arg2425Gln and Ile2435Val to the PI3Kc (phosphoinositide 3-kinase related catalytic) domain. The novel Arg2008Leu, Tyr2132Asp and Ile2435Val alterations occurred in one index case each. Interestingly, Arg2008Leu appeared to have an effect on two ESEs (SF2/ASF and SC35), as predicted by the ESEfinder program.
Arg2008Leu was identified in a patient with both ovarian and colon cancer at age 51 years, and her sister, diagnosed with breast cancer at age 72 years, was found to be a carrier. However, because Arg2008Leu, Tyr2132Asp or Ile2435Val carriers were also observed in control individuals, these changes were all classified as rare variants. Arg2425Gln is a common polymorphism described in the single nucleotide polymorphism database, and its frequency was similar in cases (27.8%) and controls (24.0%).
The rest of the amino acid substitutions were all located outside the kinase and FAT/FATC domains, although two of these, Glu254Gly and Ser1142Gly, were absent from the tested controls. Glu254Gly affects a nonconserved residue, and was seen in one patient diagnosed with breast cancer at age 37 years. However, her maternal cousin, with bilateral breast cancer at ages 45 and 55 years, was not a carrier. The other change, Ser1142Gly, affects a residue that is also conserved in ATR of Xenopus laevis and in mei41 of Drosophila melanogaster [30]. Ser1142Gly was seen in two index cases with breast cancer (2/126; P = 0.09). In the first family the patient was diagnosed at age 64 years. Two of her daughters had breast cancer at ages 49 and 40 years, but both tested negative for Ser1142Gly. Also, two sisters of the index had breast cancer at unknown ages, but no samples were available for mutation testing. In the second family the index patient was diagnosed at age 59 years, but her sister, who had breast cancer at age 45 years, was not a carrier. Neither Glu254Gly nor Ser1142Gly had an affect on splicing consensus sequences or ESEs.
Of the six intronic changes, two were absent from the tested control individuals: IVS24-48G>A was observed in the index case of three families (3/126; P = 0.03) and IVS26+15C>T was observed in one case. Unfortunately, only one additional DNA sample from an affected relative was available for mutation testing, and this maternal cousin of the index case proved negative for IVS24-48G>A. Also IVS31-74G>A was found more frequently in cases (8.7%) than in control individuals (4.3%), but the difference was only marginally significant (odds ratio 2.1, 95% confidence interval 0.92–4.85; P = 0.07). None of the observed intron changes had an affect on consensus splice sites.
Discussion
ATR plays a critical role in the maintenance of genomic integrity. A number of tumour suppressor proteins act downstream of ATR, placing it high in the DNA damage response cascade [5]. Impaired ATR signalling has been shown to result in Seckel syndrome, but thus far predisposition to cancer in these patients has not been reported [16]. Nevertheless, cell lines with inactivated ATR exhibit genetic instability, which may suggest proneness to cancer [17].
To investigate the possible role played by ATR germline mutations in hereditary predisposition to breast and ovarian cancer, the whole coding region of the gene was screened for mutations in the index cases from 126 families. We found a number of novel sequence variants, but we did not identify any clearly pathogenic alterations. Only two of the observed missense changes, Glu254Gly and Ser1142Gly, were absent from control individuals. However, because the variants did not segregate with the cancer phenotype in these families, they are unlikely to be important cancer susceptibility alleles. Evaluation of the intronic variants IVS24-48G>A and IVS26+15C>T is more difficult because only one additional DNA sample was available for mutation testing, but neither had any affect on consensus splicing sequences. The possible association of the identified rare variants with predisposition to cancer must be demonstrated by more extensive case–control studies.
The performed mutation analysis is to our knowledge the first to investigate the possible association of germline ATR mutations with cancer predisposition. However, the results of the study suggest that ATR is not involved in hereditary susceptibility to breast and ovarian cancer. The lack of deleterious germline mutations could reflect a fundamental role for ATR in cell viability, including DNA replication [5]. Nonetheless, because ATR changes have thus far been reported only in gastric and endometrial tumours exhibiting microsatellite instability, it is also possible that breast and/or ovarian cancer is not the primary cancer phenotype associated with germline mutations in this gene [13,14]. These findings need confirmation by other studies.
Conclusion
Based on its central role in the maintenance of genomic integrity, we hypothesized that germline mutations in ATR may account for some breast and/or ovarian cancer families. However, analysis of 126 index cases suggests that ATR mutations do not play a major role in hereditary susceptibility to these cancers.
Abbreviations
ATM= ataxia-telangiectasia mutated; ATR = ataxia-telangiectasia- and Rad3-related; CSGE = conformation sensitive gel electrophoresis; ESE = exonic splicing enhancer; FA = Fanconi anemia.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
KH and VM conducted the screening studies. RW, KH and VM participated in the design of the study and performed the statistical analysis. S-MK and KR helped to draft the manuscript. All authors read and approved the final manuscript.
Acknowledgements
The authors thank Drs Guillermo Blanco, Ulla Puistola, Aki Mustonen and Jaakko Ignatius, and Nurse Outi Kajula for their help in patient contact; and the Academy of Finland, University of Oulu, Oulu University Hospital, Nordic Cancer Union and Cancer Foundation of Northern Finland for providing the means for the current investigation. The kind participation of all patients has been of utmost importance to this study.
Figures and Tables
Figure 1 Structure of ATR and the location of observed amino acid changes.
Table 1 Primers used to amplify exons and splice junctions of ATR
Exon Forwarda Reverse
1 ccgggtcctatgcagaaaag aggggagagcacgtgaaac
2 cattgacactgaacacatttgatg tctaaaactacatggagaaaatgct
3 ggcccacagtctggtttct gtaatatttcagaagagcagtaaaagg
4a tcgtcaaggatttagcaaatga acgagtaagaaccattaataaagtgac
4b atgtgatgggtcatgctgtg gctcttcatagagtttcaattggtc
4c tccaaaaattaaatccctagcaa tctcacatagaccttcctgacttg
4d aaactctgtgtcatgtttgaagac gccagactacactatgaaaatcatta
5 cattcttgctgcctatgaataa aaatcaaagcacttaactaaagctga
6 tctaaatatgtttcatgttttaaccaa tgagtcaagtgaataatgagtaaaca
7 tggctttactacaattttatgtttgac cacttaggcttcaggcaaaa
8b tttaaaagagatatgattaaggaaaag cacacattcttgtgagcactt
9 aaatgtattttaagtgttacttgactttt aaccctgcatacatagccaga
10 gtcccaaattaagcaagactattt caaggcttcagtctaattcttttac
11 tcatggcatattatttgttgac gaacaataaaattaactggttaaagaa
12 gaggttgataatttttgtttttaacat ccatttttaacagcaagcaaa
13 tgagtcaacatgaatttatttgtagat aaagaaaagcaagcaaaataaaac
14 ctctatggtggcttaaaaagtattagt caaagtcaaaatctagaatggaatg
15 ctccaaatatgtgtggcattt accctctttcctagaagaatgttac
16 ctcctgatgtactaatagcatgttaaa tgaccaaaaatatgatttcttcaat
17 gcttttggagaaacttaattaacca tgtttgtagctagatgcagaattt
18b tgtccttagggctcatctgc tgaacccaatttccctcaaa
19 gctgccttttaatctattgtttg cattaccatcagtaattttgagacat
20 ggccttagtttcaacttttactttaca caggaattagctatcagaataggact
21 gagaattcaggcctttggaa aatgtcattttgtcatcttttcttt
22 aactcatcaaaaactagctgaaaaa ggataagctgaatagttctttgtaaat
23 ccatggaaaaagcagtacacc aaaacaaaaaggagtttcacaagt
24 gcataaataaagcgaagtgcaa ggccaaaaaaatcgcatta
25 agtcaactgaaggagttgctg ttgtgtgtgctaggcattcag
26 ttatctcacatgctactctttgaca catttcctactaataggtagcctttc
27 ttagaatggttagctttagatgtcata agaactgataaagggaagagctaa
28 caattgttctgttgttagttacattct gcatagcatataaaacattcaataaaa
29 aaacaggtggttttatagttttatttc aaggtttccagagttcctattca
30 aagggcaataaggtaaatagtaat aaattacccaattcactaactaaaaa
31 gaacaaaatacaatataatgcaaattcaa accgcacccatcctaaaact
32 ttgatatttcagctgaccattttatc ccaaactcactatcaattatttactcaaa
33 caactgtgtattttaaattctttatttctg cacccccaaataatatccaa
34b attgggaacagaggctttca gacatttccctggccattac
35 caaaaacataatgaactaatacttttgc catgtgctttgccatatagactt
36 tcacatacttttgatccctaatca acctagaatatgctaagacatgtga
37 tttttgtgaaaacggtatgtgg agactgtccagccaaatctga
38 tgtgaaatgaactgatatactgattttt cgccctggaacttgtatcta
39 aactctcatcatgaatactttttaagtt aaaaactgctttattaagacaaatcat
40 ttgtaaaagtgaaatttttgttatagtgg ttgtgaaatacactttttatcttaatttga
41 tttacacagaaatttttggcccta caactctgaaataaaagcaatctgg
42 tttggttatgaaatgaacaatcttt aggaagggatggaaacactt
43 agtagatgtttcttgtccaattttaac catatgaggccaatataaatctaaaa
44 gttgttatggttgaatgtttattttta caaggaagatacagttgttgagaa
45 tggacatgaagttctttgagtaaa caaacatatgtaggggccaat
46 agcttctcatccttcacttaaa aactatagctgcatatcaagttca
47a gggtattggtcagtaaaatggta ccacagattcataccaaatgc
47b gaaggacatgtgcattaccttatac cttgcttgtttcttgcaaatatag
aAll primer sequences are shown in the 5' to 3' direction. bThe amplified fragment was analyzed by direct sequencing.
Table 2 Observed sequence variation in the protein-encoding regions of ATR
Exon Nucleotide change Effect on protein Carrier frequency P Statusa
Familial cases Controls
3 268C>T His90Tyr 12.7% (16/126) 10.3% (31/300) 0.50 Novel
4 632T>C Met211Thr 54.0% (68/126) 47.3% (142/300) 0.20 Reported
761A>G Glu254Gly 0.8% (1/126) - (0/300) 0.30 Novel
891G>C Lys297Asn 1.6% (2/126) 2.3% (7/300) 1.00 Reported
8 1776T>A Gly592Gly 50.8% (64/126) 41.0% (41/100) 0.18 Reported
1815T>C Asp605Asp 46.8% (59/126) 49.0% (49/100) 0.79 Reported
14 2946C>T Phe982Phe 0.8% (1/126) 2.0% (6/300) 0.68 Novel
17 3424A>G Ser1142Gly 1.6% (2/126) - (0/300) 0.09 Novel
21 3893A>T Asp1297Val 0.8% (1/126) 1.0% (3/300) 1.00 Novel
26 4576A>G Ile1526Val 4.0% (5/126) 2.0% (6/300) 0.31 Novel
30 5208T>C Tyr1736Tyr 23.8% (30/126) 37.0% (37/100) 0.03 Reported
32 5459T>C Tyr1820Tyr 26.2% (33/126) 31.0% (31/100) 0.43 Reported
35 6023G>T Arg2008Leu 0.8% (1/ 126) 0.3% (1/300) 0.51 Novel
38 6394T>G Tyr2132Asp 0.8% (1/126) 0.3% (1/300) 0.51 Novel
43 7274G>A Arg2425Gln 27.8% (35/126) 24.0% (72/300) 0.46 Reported
7303A>G Ile2435Val 0.8% (1/126) 0.3% (1/300) 0.51 Novel
47 7875A>G Gln2625Gln 27.0% (34/126) 35.0% (35/100) 0.19 Reported
aNovel or reported in the National Center for Biotechnology Information single nucleotide polymorphism database .
Table 3 Observed sequence variation in the intron regions of ATR
Location Nucleotide change Carrier frequency P Statusa
Familial cases Controls
IVS2-51 A>T 9.5% (12/126) 9.0% (9/100) 0.89 Novel
IVS16-26 T>A 2.4% (3/126) 4.3% (13/300) 0.41 Novel
IVS18-22 G>C 4.8 % (6/126) 4.0% (12/300) 0.72 Novel
IVS24-48 G>A 2.4% (3/126) - (0/300) 0.03 Novel
IVS26+15 C>T 0.8% (1 /126) - (0/300) 0.30 Novel
IVS31-74 G>A 8.7% (11/126) 4.3% (13/300) 0.07 Novel
aNovel or reported in the National Center for Biotechnology Information single nucleotide polymorphism database .
==== Refs
Claus EB Risch N Thompson WD Genetic analysis of breast cancer in the cancer and steroid hormone study Am J Hum Genet 1991 48 232 242 1990835
Houlston R Peto J The future of association studies of common cancers Hum Genet 2003 112 434 435 12574941
Pharoah PDP Antoniou A Bobrow M Zimmern RL Easton DF Ponder BAJ Polygenic susceptibility to breast cancer and implications for prevention Nat Genet 2002 31 33 36 11984562 10.1038/ng853
Nathanson KL Weber BL 'Other' breast cancer susceptibility genes: searching for more holy grail Hum Mol Genet 2001 10 715 720 11257104 10.1093/hmg/10.7.715
Abraham RT Cell cycle checkpoint signaling through the ATM and ATR kinases Genes Dev 2001 15 2177 2196 11544175 10.1101/gad.914401
Swift M Reitnauer PJ Morrell D Chase CL Breast and other cancers in families with ataxia telangiectasia N Engl J Med 1987 316 1289 1294 3574400
Teraoka SN Malone KE Doody DR Suter NM Ostrander EA Daling JR Concannon P Increased frequency of ATM mutations in breast carcinoma patients with early onset disease and positive family history Cancer 2001 92 479 487 11505391 10.1002/1097-0142(20010801)92:3<479::AID-CNCR1346>3.0.CO;2-G
Chenevix-Trench G Spurdle AB Gatei M Kelly H Marsh A Chen X Donn K Cummings M Nyholt D Jenkins MA Dominant negative ATM mutations in breast cancer families J Natl Cancer Inst 2002 94 205 215 11830610
Thorstenson YR Roxas A Kroiss R Jenkins MA Yu KM Bachrich T Muhr D Wayne TL Chu G Davis RW Contributions of ATM mutations to familial breast and ovarian cancer Cancer Res 2003 63 3325 3333 12810666
Shiloh Y ATM and related protein kinases: safeguarding genome integrity Nat Rev Cancer 2003 3 155 168 12612651 10.1038/nrc1011
Zou L Elledge SJ Sensing DNA damage through ATRIP recognition of RPA-ssDNA complexes Science 2003 300 1542 1548 12791985 10.1126/science.1083430
Brown EJ Baltimore D ATR disruption leads to chromosomal fragmentation and early embryonic lethality Genes Dev 2000 14 397 402 10691732
Menoyo A Alazzouzi H Espin E Armengol M Yamamoto H Schwartz S Jr Somatic mutations in the DNA damage-response genes ATR and CHK1 in sporadic stomach tumors with microsatellite instability Cancer Res 2001 61 7727 7730 11691784
Vassileva V Millar A Briollais L Chapman W Bapat B Genes involved in DNA repair are mutational targets in endometrial cancers with microsatellite instability Cancer Res 2002 62 4095 4099 12124347
Fang Y Tsao CC Goodman BK Furumai R Tirado CA Abraham RT Wang XF ATR functions as a gene dosage-dependent tumor suppressor on a mismatch repair-deficient background EMBO J 2004 23 3164 3174 15282542 10.1038/sj.emboj.7600315
O'Driscoll M Ruiz-Perez VL Woods CG Jeggo PA Goodship JA A splicing mutation affecting expression of ataxia-telangiectasia and Rad3-related protein (ATR) results in Seckel syndrome Nat Genet 2003 33 497 501 12640452 10.1038/ng1129
O'Driscoll M Jeggo PA Clinical impact of ATR checkpoint signalling failure in humans Cell Cycle 2003 2 194 195 12734423
Wooster R Bignell G Lancaster J Swift S Seal S Mangion J Collins N Gregory S Gumbs C Micklem G Identification of the breast cancer susceptibility gene BRCA2 Nature 1995 378 789 792 8524414 10.1038/378789a0
Howlett NG Taniguchi T Olson S Cox B Waisfisz Q De Die-Smulders C Persky N Grompe M Joenje H Pals G Biallelic inactivation of BRCA2 in Fanconi anemia Science 2002 297 606 609 12065746 10.1126/science.1073834
Górski B Debniak T Masojc B Mierzejewski M Medrek K Cybulski C Jakubowska A Kurzawski G Chosia M Scott R Germline 657del5 mutation in the NBS1 gene in breast cancer patients Int J Cancer 2003 106 379 381 12845677 10.1002/ijc.11231
Huusko P Pääkkönen K Launonen V Pöyhönen M Blanco G Kauppila A Puistola U Kiviniemi H Kujala M Leisti J Evidence of founder mutations in Finnish BRCA1 and BRCA2 families Am J Hum Genet 1998 62 1544 1548 9585608 10.1086/301880
Allinen M Huusko P Mäntyniemi S Launonen V Winqvist R Mutation analysis of the CHK2 gene in families with hereditary breast cancer Br J Cancer 2001 85 209 212 11461078 10.1054/bjoc.2001.1858
Rapakko K Allinen M Syrjäkoski K Vahteristo P Huusko P Vähäkangas K Eerola H Kainu T Kallioniemi OP Nevanlinna H Germline TP53 alterations in Finnish breast cancer families are rare and occur at conserved mutation-prone sites Br J Cancer 2001 84 116 119 11139324 10.1054/bjoc.2000.1530
Körkkö J Annunen S Pihlajamaa T Prockop DJ Ala-Kokko L Conformation sensitive gel electrophoresis for simple and accurate detection of mutations: comparison with denaturing gradient gel electrophoresis and nucleotide sequencing Proc Natl Acad Sci USA 1998 95 1681 1685 9465076 10.1073/pnas.95.4.1681
Ganguly A An update on conformation sensitive gel electrophoresis Hum Mutat 2002 19 334 342 11933188 10.1002/humu.10059
Primer 3 Input
db SNP
Berkeley Drosophila Genome Project
ESE finder
Perry J Kleckner N The ATRs, ATMs, and TORs are giant HEAT repeat proteins Cell 2003 112 151 155 12553904 10.1016/S0092-8674(03)00033-3
| 15987455 | PMC1175065 | CC BY | 2021-01-04 16:54:50 | no | Breast Cancer Res. 2005 May 6; 7(4):R495-R501 | utf-8 | Breast Cancer Res | 2,005 | 10.1186/bcr1037 | oa_comm |
==== Front
Breast Cancer ResBreast Cancer Research1465-54111465-542XBioMed Central London bcr10381598745610.1186/bcr1038Research ArticleTP53-binding protein variants and breast cancer risk: a case-control study Frank Bernd [email protected] Kari 12Bermejo Justo Lorenzo 1Klaes Rüdiger 3Bugert Peter 4Wappenschmidt Barbara 5Schmutzler Rita K 5Burwinkel Barbara 11 Division of Molecular Genetic Epidemiology, German Cancer Research Center (DKFZ), Heidelberg, Germany2 Department of Biosciences at Novum, Karolinska Institute, Huddinge, Sweden3 Institute of Human Genetics, University of Heidelberg, Heidelberg, Germany4 Institute of Transfusion Medicine and Immunology, Red Cross Blood Service of Baden-Württemberg-Hessia, University of Heidelberg, Faculty of Clinical Medicine, Mannheim, Germany5 Division of Molecular Gynaeco-Oncology, Department of Gynaecology and Obstetrics, Clinical Center University of Cologne, Germany2005 6 5 2005 7 4 R502 R505 25 2 2005 17 3 2005 23 3 2005 12 4 2005 Copyright © 2005 Frank et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Introduction
The TP53-binding protein (53BP1) has been shown to influence TP53-mediated transcriptional activation, thus playing a pivotal role in DNA damage signalling. Genetic aberrations in TP53 and in ATM and CHEK2 predispose to cancer. We have therefore examined the effects of 53BP1 single nucleotide polymorphisms (D353E, G412S, and K1136Q) and the novel 53BP1 6bp deletion (1347_1352delTATCCC) on breast cancer risk.
Methods
Allelic discrimination was performed to investigate the frequencies of 53BP1 D353E, G412S, and K1136Q and of 1347_1352delTATCCC in 353 patients with breast cancer and 960 control individuals.
Results
No significant association of 53BP1 D353E, G412S, or K1136Q with breast cancer risk was detected. 53BP1 1347_1352delTATCCC, leading to the loss of an isoleucine and a proline residue, showed a nonsignificant inverse association with breast cancer risk (odds ratio = 0.61, 95% confidence interval = 0.22 to 1.68, P = 0.34).
Conclusion
The lack of association casts doubt on the putative effects of D353E, G412S, and K1136Q on breast cancer risk. Investigating a larger study cohort might elucidate the influence of the 6bp deletion 1347_1352delTATCCC. Studying the functional effect and the impact of this variant on the risk of other cancers may be revealing.
==== Body
Introduction
The TP53-binding protein (53BP1), a conserved nuclear protein, was initially identified to interact with the DNA-binding domain of TP53, thus enhancing TP53-mediated transcriptional activation [1,2]. In response to exogenous exposure to ionising radiation, 53BP1 becomes hyperphosphorylated and rapidly localises to sites of DNA double-strand breaks, demonstrating its determining role in DNA damage signalling [3,4]. 53BP1-deficient mice exhibit growth retardation, high radiation sensitivity, and tumour development – features that are indicative of a defective DNA damage response [5]. 53BP1 is involved in the phosphorylation of various ataxia telangiectasia mutated protein (ATM) substrates such as cell cycle checkpoint kinase 2 (CHEK2) [3,6]. Mutations in ATM, CKEK2, and its substrate, TP53, have been shown to predispose to cancer [6-9]. Therefore, we selected 53BP1 as an attractive candidate gene for breast cancer susceptibility.
This is the first study to investigate the effects of the 53BP1 single nucleotide polymorphisms (SNPs) D353E (1059C>G), G412S (1234G>A), and K1136Q (3406A>C) on breast cancer risk, analysing 353 German patients with breast cancer and 960 controls. 53BP1 D353E, G412S, and K1136Q showed no association with breast cancer risk. In addition, we detected a novel, very rare 53BP1 6bp deletion (1347_1352delTATCCC) showing an inverse association with breast cancer risk (age-adjusted odds ratio (OR) = 0.61, 95% confidence interval (CI) = 0.22 to 1.68), lacking significance (P = 0.34).
Materials and methods
SNP verification
A randomly chosen set of 23 German patients with familial breast cancer was initially screened for annotated 53BP1 SNPs (dbSNP database; NCBI (National Center for Biotechnology Information)) by DNA sequencing. Sequencing primers are available upon request. The initial analysis included 53BP1 exons 9, 11, and 17, harbouring three reported nonsynonymous polymorphisms (D353E: rs560191; G412S: rs689647; and K1136Q: rs2602141). When sequencing exon 11, we additionally detected the 6bp deletion 1347_1352delTATCCC. All validated variants were chosen for further analyses using a large cohort of breast cancer patients.
Subjects
The breast cancer patients were 353 unrelated German women (mean age 44.8 years, range 21 to 80 years) who were negative for BRCA1 and BRCA2 mutations. In accordance with the German Consortium for Hereditary Breast and Ovarian Cancer, they were classified into six categories based on family history: (A1) families with two or more breast cancer cases including two cases with onset below the age of 50 (39.3% of analysed cases); (A2) families with at least one male breast cancer case (0.9%); (B) families with at least one breast cancer and one ovarian cancer case (16.2%); (C) families with at least two breast cancer cases including one case diagnosed before the age of 50 (33.5%); (D) families with at least two breast cancer cases comprising two cases diagnosed after the age of 50 (5.5%); (E) single cases of breast cancer diagnosed before the age of 35 (4.6%) [10]. They were collected during the years 1997 to 2004 through the Institute of Human Genetics (Heidelberg, Germany) and the Department of Gynaecology and Obstetrics (Cologne, Germany). The control series included 960 blood donors (mean age 30.5 years, range 18 to 67 years) collected by the Institute of Transfusion Medicine and Immunology (Mannheim, Germany) having the same ethnic background as the breast cancer patients. Both study populations have been described earlier [11]. The study was approved by the Ethics Committee of the University of Heidelberg (Heidelberg, Germany).
PCR amplification and sequencing were performed as previously described [11]. Conditions are available on request.
Genotyping
53BP1 polymorphisms D353E, G412S, and K1136Q were analysed using TaqMan allelic discrimination. TaqMan assays were performed in a reaction volume of 10 μl comprising 5ng of genomic DNA, each probe at 50 nM, each primer at 225 nM, and 1× Universal Master Mix with the following amplification conditions: 2 min at 50°C, 10 min at 95°C and 35 to 45 cycles at 92°C for 15 s and 60°C for 1 min. Amplification products were measured and analysed with the ABI Prism 7900 HT sequence detection system and the SDS software (version 1.2; Applied Biosystems, Foster City, CA, USA). TaqMan probes and primers were provided by the assay-on-demand and assay-by-design services, respectively (Applied Biosystems). 53BP1 1347_1352delTATCCC was analysed using the MGB Eclipse™ Probe System by Epoch Biosciences (Bothell, WA, USA). Allelic discrimination was carried out as recommended by the manufacturers using the following probes: D353E: assay-on-demand C_2944794_10; G412S: VIC-ACTTCAAAGTGGTGAACC, FAM-AACTTCAAAGTAGTGAACC; K1136Q: VIC-GGAGTACTAATAAGGAAA, FAM-CGGAGTACTAATCAGGAAA; 1347_1352delTATCCC: FAM-CACTTCATCCCAT; TET-CACTTCCTATCCCATC. Primers and probes were designed based on GenBank NM_005657 (NCBI) and are available on request. More than 5% of the genotyping results were confirmed by sequencing, and genotype distributions were consistent with Hardy–Weinberg equilibrium.
Statistical methods
Calculations of Hardy–Weinberg equilibrium, genotype-specific OR, and 95% CI were carried out using a tool offered by the Institute of Human Genetics, Technical University Munich, Munich, Germany [12]. Age-adjusted ORs and corresponding 95% CIs were computed by means of unconditional logistic regression using SAS (Version 8.2; SAS Institute Inc, Cary, NC, USA). Haplotypes were assigned to subjects using the SNPHAP software (see [13]), which also reports the posterior probability of the most likely assignment [14,15].
Results and discussion
Inactivation of ATM and ATM substrates such as CHEK2 have been shown to predispose to cancer in humans [7]. Along with ATM and CHEK2, 53BP1 is involved in DNA damage response and tumour suppression. Recent studies have shown that 53BP1 and ATM interact in irradiated cells, suggesting that ATM activation is the consequence of the recruitment of ATM to sites of DNA double-strand breaks by 53BP1 [7,9]. Thus, polymorphic variants in 53BP1 are excellent candidates for cancer susceptibility. We investigated the impact of three nonsynonymous amino acid exchanges in 53BP1 on breast cancer risk. 53BP1 G413S and K1136Q represented promising candidate SNPs, resulting in the replacement of a nonpolar by a polar amino acid. Genotype frequencies of the three 53BP1 polymorphisms between breast cancer cases and control samples were similar, showing no significant association with breast cancer risk (D353E: age-adusted OR = 1.07, 95% CI = 0.81 to 1.43, P = 0.62; G412S: age-adjusted OR = 1.22, 95% CI = 0.86 to 1.74, P = 0.26; K1136Q: age-adjusted OR = 1.10, 95% CI = 0.82 to 1.47, P = 0.53; Table 1). Aditionally, we detected a novel 53BP1 6 bp deletion, 1347_1352delTATCCC, leading to the loss of an isoleucine and a proline residue at positions 450 and 451, which has not been described previously. Comparing the occurrence of this rare, 6 bp deletion between cases and controls revealed an inverse association with breast cancer risk (OR = 0.61, 95% CI = 0.22 to 1.68, P = 0.34; Table 1), but lacking statistical significance.
The haplotype distribution and corresponding posterior probabilities are shown in Table 2. Since every mean posterior probability was higher than 0.9, only the most likely haplotypes were used to evaluate the association with breast cancer risk. Haplotype analysis showed a nonsignificant inverse association of the haplotype 1059C-1234G-1347_1352--3406A with breast cancer risk (age-adjusted OR = 0.63, 95% CI = 0.23 to 1.75, P = 0.38; Table 2). The distribution of the remaining haplotypes between breast cancer patients and controls was similar, indicating no significant effect with regard to breast cancer risk. Given our sample size, we had a 90% power to detect an odds ratio of 1.65 (D353E), 1.76 (G412S), and 1.66 (K1136Q), respectively [16]. Contrary to standard case-control association studies, this study comprised predominantly cases selected for family history of breast cancer. The use of unselected cases would have required at least twice the sample size to achieve the same power as in the present study [17,18]. The numbers of cases within the risk groups A1 to E were too low to be studied separately, as the power in these subgroups would have been limited. In addition to the results of this study, one cannot exclude the possibility that common 53BP1 SNPs may affect breast cancer risk. Regulatory polymorphisms, for example polymorphisms that reside in promotor or noncoding regions, have been shown to modify gene transcription, mRNA stability, and processing efficiency, as well as DNA methylation [19,20].
Conclusion
The three known 53BP1 SNPs – D353E, G412S, and K1136Q – lacked association with breast cancer risk. However, we detected a novel, very rare 6bp deletion, 1347_1352delTATCCC, that showed a statistically nonsignificant inverse association with breast cancer risk. Concerning the latter, a much larger study cohort is required to verify any putative significant effect. Additionally, it would be valuable to investigate a possible functional effect of this 53BP1 deletion and its impact on other cancers.
Abbreviations
53BP1 = TP53 binding protein; ATM = ataxia telangiectasia mutated protein; bp = base pairs; CHEK2 = cell cycle checkpoint kinase 2; CI = confidence interval; OR = odds ratio; PCR = polymerase chain reaction; SNP = single nucleotide polymorphism.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
All authors listed contributed to the production of this manuscript: RK, PB, BW, and RKS provided genomic DNAs of cases studied and helped to draft the manuscript. BB and KH participated in the design and coordination of the study and critically revised the manuscript. BF and JLB performed statistical analyses. BF carried out the SNP genotyping and wrote the manuscript. All authors read and approved the final manuscript.
Acknowledgements
The authors wish to thank Kerstin Wagner for critical comments on the manuscript and are grateful to Prof CR Bartram (Heidelberg) and Prof RK Schmutzler (Cologne) for providing the German breast cancer samples, which were collected within a project funded by the Deutsche Krebshilfe.
Figures and Tables
Table 1 Genotype frequencies of 53BP1 polymorphisms in breast cancer patients and controls
Polymorphism No. (%) of cases No. (%) of controls AOR (95% CI) P
D353E (1059C>G)
CC 165 (48.1) 453 (47.6)
GC 148 (43.1) 405 (42.5)
GG 30 (8.7) 94 (9.9)
Σ 343 952
CC vs GC+GG 1.07 (0.81–1.43) 0.62
G412S (1234G>A)
GG 269 (78.7) 760 (80.1)
AG 67 (19.6) 174 (18.3)
AA 6 (1.8) 15 (1.6)
Σ 342 949
AA+AG vs GG 1.22 (0.86–1.74) 0.26
1347_1352delTATCCC
wt/wt 334 (98.2) 931 (97.4)
delTATCCC/wt 6 (1.8) 25 (2.6)
delTATCCC/delTATCCC 0 (0.0) 0 (0.0)
Σ 340 956
delTATCCC/wt vs wt/wt 0.61 (0.22–1.68) 0.34
K1136Q (3406A>C)
AA 158 (47.4) 448 (47.8)
CA 144 (43.2) 396 (42.2)
CC 31 (9.3) 94 (10.0)
Σ 333 938
CC+CA vs AA 1.10 (0.82–1.47) 0.53
AOR, age-adjusted odds ratio; CI, confidence interval; wt, wild type.
Table 2 Haplotype distribution of 53BP1 polymorphismsa in breast cancer patients and control individuals
Cases Controls ORb 95 % CI P
Haplotype 1059C>G-1234G>A-1347_1352-/+-3406A>C No. (%) Mean posterior probability No. (%) Mean posterior probability
CG+A 444 (68.1) 1 1242 (67.5) 1 1 - -
CG+C 2 (0.3) 1 1 (0.1) 1 n.a. n.a. n.a.
CG–A 6 (0.9) 0.90 25 (1.4) 0.99 0.63 0.23–1.75 0.38
GG+A 1 (0.2) 1 1 (0.1) 1 n.a. n.a. n.a.
GG+C 125 (19.2) 1 376 (20.4) 1 1.02 0.79–1.33 0.86
GA+C 74 (11.3) 1 195 (10.6) 1 1.15 0.83–1.61 0.41
aPolymorphisms D353E (1059C>G), G412S (1234G>A), 1347_1352delTATCCC, and K1136Q (3406A>C). b1059C-1234G-1347_1352+-3406A as reference. CI, confidence interval; n.a., not available; OR, odds ratio.
==== Refs
Iwabuchi K Bartel PL Li B Marraccino R Fields S Two cellular proteins that bind to wild-type but not mutant p53 Proc Natl Acad Sci USA 1994 91 6098 6102 8016121
Iwabuchi K Li B Massa HF Trask BJ Date T Fields S Stimulation of p53-mediated transcriptional activation by the p53-binding proteins, 53BP1 and 53BP2 J Biol Chem 1998 273 26061 26068 9748285 10.1074/jbc.273.40.26061
Rappold I Iwabuchi K Date T Chen J Tumor suppressor p53 binding protein 1 (53BP1) is involved in DNA damage-signaling pathways J Cell Biol 2001 153 613 620 11331310 10.1083/jcb.153.3.613
Wang B Matsuoka S Carpenter PB Elledge SJ 53BP1, a mediator of the DNA damage checkpoint Science 2002 298 1435 1438 12364621 10.1126/science.1076182
Ward IM Minn K van Deursen J Chen J p53 Binding protein 53BP1 is required for DNA damage responses and tumor suppression in mice Mol Cell Biol 2003 23 2556 2563 12640136 10.1128/MCB.23.7.2556-2563.2003
CHEK2-Breast Cancer Consortium Low-penetrance susceptibility to breast cancer due to CHEK2(*)1100delC in noncarriers of BRCA1 or BRCA2 mutations Nat Genet 2002 31 55 59 11967536 10.1038/ng879
DiTullio RA JrMochan TA Venere M Bartkova J Sehested M Bartek J Halazonetis TD 53BP1 functions in an ATM-dependent checkpoint pathway that is constitutively activated in human cancer Nat Cell Biol 2002 4 998 1002 12447382 10.1038/ncb892
Vorechovsky I Luo L Lindblom A Negrini M Webster AD Croce CM Hammarstrom L ATM mutations in cancer families Cancer Res 1996 56 4130 4133 8797579
Mochan TA Venere M DiTullio RA JrHalazonetis TD 53BP1, an activator of ATM in response to DNA damage DNA Repair 2004 3 945 952 15279780 10.1016/j.dnarep.2004.03.017
Meindl A and German Consortium for Hereditary Breast and Ovarian Cancer Comprehensive analysis of 989 patients with breast or ovarian cancer provides BRCA1 and BRCA2 mutation profiles and frequencies for the German population Int J Cancer 2002 97 472 480 11802209 10.1002/ijc.1626
Frank B Hemminki K Wirtenberger M Bermejo JL Bugert P Klaes R Schmutzler RK Wappenschmidt B Bartram CR Burwinkel B The rare ERBB2 variant Ile654Val is associated with an increased familial breast cancer risk Carcinogenesis 2005 26 643 647 15550452 10.1093/carcin/bgh342
Hardy-Weinberg equilibrium
Software and course materials
Schaid DJ Evaluating associations of haplotypes with traits Genet Epidemiol 2004 27 348 64 15543638 10.1002/gepi.20037
Clayton D Chapman J Cooper J Use of unphased multilocus genotype data in indirect association studies Genet Epidemiol 2004 27 415 28 15481099 10.1002/gepi.20032
Dupont WD Plummer WD Jr Power and sample size calculations for studies involving linear regression Control Clin Trials 1998 19 589 601 9875838 10.1016/S0197-2456(98)00037-3
Antoniou AC Easton DF Polygenic inheritance of breast cancer: Implications for design of association studies Genet Epidemiol 2003 25 190 202 14557987 10.1002/gepi.10261
Houlston RS Peto J The future of association studies of common cancers Hum Genet 2003 112 434 435 12574941
Pastinen T Hudson TJ Cis-acting regulatory variation in the human genome Science 2004 306 647 650 15499010 10.1126/science.1101659
Loktionov A Common gene polymorphisms, cancer progression and prognosis Cancer Lett 2004 208 1 33 15105042 10.1016/j.canlet.2004.02.009
| 15987456 | PMC1175066 | CC BY | 2021-01-04 16:54:50 | no | Breast Cancer Res. 2005 May 6; 7(4):R502-R505 | utf-8 | Breast Cancer Res | 2,005 | 10.1186/bcr1038 | oa_comm |
==== Front
Breast Cancer ResBreast Cancer Research1465-54111465-542XBioMed Central London bcr10391598746210.1186/bcr1039Research ArticleBenign breast disease, recent alcohol consumption, and risk of breast cancer: a nested case–control study Tamimi Rulla M [email protected] Celia [email protected] Heather J [email protected] Bernie [email protected] Stuart J [email protected] James L [email protected] Graham A [email protected] Channing Laboratory, Department of Medicine, Brigham and Women's Hospital and Harvard Medical School, Boston, MA, USA2 Department of Epidemiology, Harvard School of Public Health, Boston, MA, USA3 Cancer Genetics and Epidemiology Program, Lombardi Comprehensive Cancer Center, Georgetown University, Washington, DC, USA4 Department of Biostatistics, Harvard School of Public Health, Boston, MA, USA5 Department of Pathology, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, MA, USA6 Harvard Center for Cancer Prevention, Boston, MA, USA2005 16 5 2005 7 4 R555 R562 7 1 2005 31 1 2005 8 3 2005 12 4 2005 Copyright © 2005 Tamimi et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Introduction
Alcohol consumption is a well-established risk factor for breast cancer. Some studies have suggested that the risk of breast cancer associated with alcohol consumption is greater for women with a history of benign breast disease (BBD). We hypothesized that among women with biopsy-confirmed BBD, recent alcohol consumption would increase the risk of breast cancer in women with proliferative breast disease to a greater extent than in women with nonproliferative breast disease.
Methods
We conducted a nested case–control study in the Nurses' Health Study I and II. The cases (n = 282) were women diagnosed with incident breast cancer, with a prior biopsy-confirmed breast disease. The controls (n = 1,223) were participants with a previous BBD biopsy, but without a diagnosis of breast cancer. Pathologists reviewed benign breast biopsy slides in a blinded fashion and classified the BBD as nonproliferative, proliferative without atypia, or atypical hyperplasia, according to standard criteria.
Results
Women with nonproliferative breast disease consuming ≥ 15 g of alcohol per day had a nonsignificant 67% increased risk of breast cancer (odds ratio = 1.67; 95% confidence interval 0.65 to 4.34) compared with nondrinkers. There was no evidence that recent alcohol consumption increased the risk of breast cancer to a greater extent in women with proliferative BBD than among women with nonproliferative BBD (P for interactio n = 0.20).
Conclusion
Contrary to our a priori hypothesis, there was no evidence that recent alcohol consumption increased the risk of breast cancer to a greater extent among women with proliferative BBD than among women with nonproliferative BBD.
==== Body
Introduction
Alcohol consumption is a well-established risk factor for breast cancer. Among those who drink, recent alcohol consumption of two or more drinks a day is associated with an approximately 30% increased risk of breast cancer compared with nondrinkers [1-3]. A pooled analysis of over 32,000 women participating in six prospective cohort studies [1], and at least three meta-analyses [2,4,5], confirmed that risk of breast cancer increases monotonically with increasing alcohol consumption [1,2].
Data from epidemiologic studies also suggest that alcohol may influence early stages as well as later stages of breast carcinogenesis. The primary evidence that alcohol may have an effect on early stages in the carcinogenic process comes largely from studies of alcohol consumption and mammographic density. Breast density is a strong predictor of breast cancer risk and is considered to be an early biomarker of breast cancer [6-9]. In most studies, recent alcohol consumption is positively associated with mammographic density [10-12]. Recent alcohol consumption has not been associated with increased incidence of proliferative benign breast disease (BBD) [13-15], although alcohol consumption between the ages of 18 and 22 has been associated with an increased risk of proliferative breast disease in one study [13].
A few studies have suggested that the increased risk of breast cancer associated with alcohol consumption is greater for women with a history of BBD [16,17], although the interactions were not statistically significant and this relation has not been observed in all studies [18]. BBD comprises multiple histologic subtypes. In prospective studies, where investigators have examined the risk associated with histologic subtypes of BBD, the proliferative lesions and in particular those with atypia were associated with the highest risk [19-21]. We hypothesized that among women with biopsy-confirmed BBD, alcohol consumption would increase the risk of breast cancer in those with proliferative breast disease to a greater extent than in women with nonproliferative breast disease.
Materials and methods
Study population
The Nurses' Health Study (NHS I) cohort was initiated in 1976, when 121,700 US registered nurses ages 30 to 55 returned an initial questionnaire. The Nurses' Health Study II (NHS II) is also an ongoing cohort study of over 116,000 US female nurses who were 25 to 42 years of age in 1989 when the study was initiated. Every 2 years, information on reproductive variables, body mass index, exogenous hormone use, and disease outcomes is collected in both cohorts. Semiquantitative food frequency questionnaires were included as part of the biennial questionnaire in 1980, 1984, 1986, and 1990 in NHS I, and in 1991 in NHS II. Alcohol consumption was also assessed on the baseline 1989 questionnaire in NHS II. The methods developed to follow up participants and confirm incident cancers and death in the Nurses' Health Study have been described previously in detail elsewhere [22] and have been applied to NHS II.
Benign breast disease
Beginning with the initial NHS I questionnaire in 1976, participants have been asked on every biennial questionnaire to report any diagnosis of fibrocystic disease or other BBD. Early questionnaires (1976, 1978, and 1980) asked whether the respondent had ever been diagnosed as having 'fibrocystic disease' or 'other BBD' and whether she had been hospitalized with this diagnosis. Beginning in 1982, the NHS I questionnaires sought specific details of a history of biopsy-confirmed BBD. The initial 1989 NHS II questionnaire and all subsequent biennial questionnaires also asked participants to report any diagnosis of BBD and to indicate if it was confirmed by biopsy or aspiration.
Selection of breast cancer cases and controls
We conducted a case–control study nested within the subcohort of women with a biopsy-confirmed BBD. Incident breast cancer cases in both cohorts were identified through the nurses' own reports and were confirmed by review of medical records. Cases in the study are cases of breast cancer diagnosed by 1 June 1994 (NHS I) or 1 June 1995 (NHS II) with a previous confirmed BBD biopsy and available pathology specimens (histologic sections and/or tissue blocks). Controls are women who did not have a diagnosis of breast cancer when the case was diagnosed and also had a previous biopsy-confirmed BBD and available pathology specimens. Controls were matched to cases on year of birth and year of biopsy. Attempts were made to identify four matched controls for each case, although this was not always possible. This study was approved by the Committee on Human Subjects at Brigham and Women's Hospital.
We identified incident confirmed breast cancer cases diagnosed after the return of the initial questionnaire through the 1994/1995 follow-up cycle and controls that also reported a previous biopsy-confirmed BBD. A total of 1,080 cases were originally identified for this study, and 4,353 controls matched on age and year of benign breast biopsy were selected. After selection, 2.1% were found to be no longer eligible and were excluded. Seventy-five percent of eligible participants confirmed their BBD biopsy and granted permission to review their pathology slides. In response to our slide requests, we received pathology specimens for 55% of the study subjects. The primary reason given by pathology departments for not sending slides or tissue blocks was that the specimens had been destroyed or were no longer available (33%). There were no significant differences in the success of obtaining slides for breast cancer cases and controls. A detailed description of the steps of specimen collection and exclusions of selected cases and controls during the collection process are included in Table 1. An additional 156 women were excluded for the following reasons: reported bilateral BBD, but tissue specimen did not identify breast side (n = 2), report of breast cancer either before study start or after the 1994/1995 follow-up cycle (n = 10), and date of benign breast diagnosis was within six months of the breast cancer diagnosis date (cases) or index date (controls) (n = 144).
Pathology review
Slides were coded by a research assistant and submitted to one of two collaborating pathologists (SJS, JLC) in a blinded fashion. The pathologists independently reviewed the BBD biopsy slides. Any slide identified as having either questionable atypia or atypia was jointly reviewed by the two pathologists. For each set of slides reviewed, a detailed work sheet was completed quantifying the extent of proliferative and atypical changes, morphologic features, and other details. BBDs were classified according to the Page classification system [23] into one of three categories: nonproliferative, proliferative without atypia, and atypical hyperplasia. Upon pathology review, it was determined that 29 women with an original BBD diagnosis had evidence of carcinoma in situ (n = 25) or invasive carcinoma (n = 4). These women were excluded from further analysis. There were 353 breast cancer cases and 1,495 controls with pathology slides available from their BBD biopsy.
Assessment of alcohol consumption
Information on alcohol consumption was obtained from semiquantitative food frequency questionnaires. In NHS I, questions regarding alcohol consumption were asked in 1980, 1984, 1986, and 1990. Women were asked about their average consumption of beer, wine, and liquor separately in the prior year. One drink was considered equal to one can or bottle of beer, a 4-ounce glass of wine, or one drink or shot of liquor. Participants were asked to select from the following categories: almost never, 1 to 3 per month, 1 per week, 2 to 4 per week, 5 to 6 per week, 1 per day, 2 to 3 per day, 4 to 6 per day, 6 or more per day. Similarly, women in NHS II answered questions on consumption of alcohol in the 1989 and 1991 questionnaires. In 1991, the NHS II alcohol questions were expanded to include red wine, white wine, light beer, regular beer, and liquor. Total alcohol consumption per questionnaire cycle was calculated by adding the alcohol contributions from beer, wine, and liquor. Grams of ethanol per day were calculated based on the following equivalents of 12.8 g for regular beer, 11.3 g for light beer, 11.0 g for wine, and 14.0 g for liquor. Alcohol consumption in NHS I has been shown to be valid and highly reproducible in repeated assessments [24].
Women were assigned the alcohol exposure from the cycle preceding the diagnosis of breast cancer. Cases and controls in NHS I with diagnosis (or index) dates preceding the initial food frequency questionnaire administered in 1980 were excluded from the analysis (n = 36 cases, 129 controls). If alcohol consumption was missing from the questionnaire before the index date, the exposure from the preceding cycle was assigned. The Spearman correlation between reported alcohol consumption in questionnaire cycles was 0.80 (P<0.0001) or greater for all consecutive cycles. Women with missing alcohol exposure data from two cycles preceding the index date were excluded from the analyses (n = 35 cases, 143 controls). Thus, there were 282 patients with breast cancer and 1,223 controls eligible for this study with both BBD pathology and detailed alcohol exposure preceding their index date.
Analytic methods
Odds ratios (ORs) and 95% confidence intervals (CIs) were determined using logistic regression models controlling for matching factors (age, year of BBD biopsy) and follow-up time using the SAS software package (version 8.0; SAS institute, Cary, NC). Follow-up time was defined as the time from BBD diagnosis to breast cancer diagnosis date (cases) or index date (controls). Information on potential confounding variables was obtained from responses on biennial questionnaires. For each 2-year cycle of case–control selection, covariate information was determined from the responses on questionnaires immediately preceding the cycle in which the breast cancer case was diagnosed. In addition to matching factors and follow-up interval, multivariate analyses were adjusted for the following confounders and breast cancer risk factors: first-degree family history of breast cancer (yes/no); quartiles of body mass index (≤ 21.6, 21.7 to 23.6, 23.7 to 26.6, ≥ 26.7 kg/m2); age at menarche (<12, 12, 13, ≥ 14 years); age at first birth/parity (nulliparous; one to four children, and age at first birth if ≤ 24 years of age; one to four children, and age at first birth if >24 years of age; five or more children, and age at first birth if ≤ 24 years of age; five or more children age at first birth >24); duration of postmenopausal hormone use (never, <5 years, ≥ 5 years); and menopausal status/type of menopause (premenopausal, natural, bilateral oopherectomy, other or unknown type of menopause).
Analyses examining the combined effect of benign breast histology and alcohol consumption were based on a cross-classified variable that used nonproliferative breast disease and consumption of 0 g of alcohol per day as the reference group. To examine whether the association between alcohol consumption and breast cancer was modified by benign breast histology, we conducted a likelihood ratio test to assess statistical significance of an interaction term using an ordered scale for benign breast histology and alcohol consumption.
Results
Among the controls selected in this nested case–control study, women with atypical hyperplasia were older at biopsy and had their biopsies slightly later (1983 versus 1979) than women with nonproliferative benign histology (Table 2). Women with atypical hyperplasia had a greater prevalence of family history of breast cancer and were less likely to be premenopausal at benign biopsy than women with nonproliferative breast disease. In addition, recent alcohol consumption was greatest for women with atypical hyperplasia and lowest for women with nonproliferative breast disease (Table 2). This result is unexpected, given that previous analysis in the NHS II cohort indicated that recent alcohol consumption is not associated with an increased risk of proliferative breast disease, including atypical hyperplasia [13]. Adjusting for age did not change the interpretation of the results in Table 2. For the other breast cancer risk factors presented in Table 2, there was no clear pattern of association with benign histology.
Overall, women with proliferative breast disease without atypia had a 50% greater risk of breast cancer than women with nonproliferative disease (OR = 1.54) (Table 3). Women with atypical hyperplasia had a more than fourfold increased risk of breast cancer in comparison with women with nonproliferative breast disease (OR = 4.43) (Table 3).
In this nested case–control study, there was no increased risk of breast cancer among women drinking ≥ 15 g of alcohol per day compared with nondrinkers (OR = 0.86; 95% CI 0.50 to 1.46).
The combined effects of alcohol consumption and category of BBD histology are presented in Table 4. Among women with nonproliferative BBD, recent alcohol consumption was positively associated with breast cancer risk: those consuming ≥ 15 g of alcohol per day had a nonsignificant 67% increased risk of breast cancer (OR = 1.67; 95% CI 0.65 to 4.34).
Among women with nonproliferative BBD, there was a 10% increased risk of breast cancer per 5 g increase in daily alcohol consumption (OR = 1.10; 95% CI 0.95 to 1.27). Among women with proliferative BBD without atypia, there was no increased risk of breast cancer associated with a 5-g increase in daily alcohol consumption (OR = 0.97; 95% CI 0.87 to 1.08). Similarly, among women with proliferative BBD with atypia, there was no evidence that increasing alcohol consumption increases risk: a 5-g increase in alcohol consumption was associated with a nonsignificant 12% reduction in breast cancer risk (OR = 0.88; 95% CI 0.73 to 1.06). The interaction between alcohol consumption and BBD category was not statistically significant (P = 0.20).
Analyses limited to invasive breast cancers (n = 232 cases) showed similar results. We were limited in power to adequately assess this relation according to menopausal status and estrogen receptor status of the cancer.
Discussion
To our knowledge, this is the first study to prospectively examine the effects of alcohol with proliferative BBD in relation to subsequent breast cancer risk. The results of this study indicate that recent alcohol consumption does not increase the risk of breast cancer to a greater extent among women with proliferative BBD than in women with nonproliferative BBD.
Previous epidemiologic studies have examined the effect of alcohol consumption among women with a self-report of BBD, with inconsistent results [1,16-18]. In a cohort study conducted in the Netherlands, van den Brandt and colleagues reported a nonsignificant 2.5-fold increased risk of breast cancer among women with BBD who drank ≥ 15 g of alcohol per day compared with nondrinkers (P for trend = 0.15) [16]. A pooled analysis of six cohort studies, including the Netherlands cohort just mentioned, reported no significant interaction between alcohol consumption and BBD with respect to breast cancer risk (P = 0.23) [1]. In the California Teachers Study, the joint effect of history of BBD and high alcohol consumption was associated with a twofold increased risk of breast cancer (relative risk (RR) = 1.97; 95% CI 1.39 to 2.79), while nondrinkers with biopsy-confirmed BBD had a 35% increased risk of breast cancer (RR = 1.35; 95% CI 1.05 to 1.73) in comparison with nondrinkers with no history of BBD [17].
Benign breast conditions are a heterogeneous group of diseases and therefore may not all respond to alcohol exposure in the same manner. Although all of the women in this study had a biopsy removing their benign lesion, we were operating under the assumption that the BBD is a marker of susceptibility and the remaining breast tissue may have a similar susceptibly to the effects of alcohol later in life. Nonproliferative breast diseases comprise a large proportion of reported BBD biopsies. The inconsistencies observed between our study and previous studies may be due to the heterogeneous nature of BBD and variability in distribution of BBD types between different studies.
One proposed mechanism by which alcohol may influence breast cancer risk is by increasing circulating estradiol levels, as has been observed in controlled feeding studies in both premenopausal and postmenopausal women [6,7]. A second potential mechanism is that alcohol may function as a cocarcinogen, inhibiting detoxification of carcinogens, or by impairing clearance of carcinogens [3,8-10]. There are data suggesting that alcohol may act early in the carcinogenic process [6-9], as well as later, functioning as a tumor promoter [3].
The current study failed to support our a priori hypothesis and suggests that recent alcohol consumption does not contribute additional risk to women with proliferative breast disease. One explanation may be that women with proliferative breast diseases are further along on the continuum to breast cancer, and are already at an elevated risk of breast cancer, which is no longer affected by alcohol. In contrast, women with nonproliferative breast disease are not as far along the pathway to breast cancer and specific exposures such as alcohol may exhibit harmful effects and influence the risk of breast cancer.
Byrne and colleagues conducted a similar study in the Nurses' Health Study to examine the effects of postmenopausal hormone use among women with BBD [25]. Similarly, the results were contrary to their a priori hypothesis and they observed no elevated risk of breast cancer among women with proliferative breast diseases according to current use of hormones or duration of use, again suggesting that these women may be at such an increased risk of breast cancer that the additional effect of exogenous hormone use is minimal or none.
In Figure 1, we have schematically described the evidence relating alcohol exposure to breast cancer. It is well accepted that recent alcohol consumption increases the risk of breast cancer [1,16-18,26-28] and there is little evidence that alcohol intake early in life affects breast cancer risk [17,26-28]. The three studies examining alcohol and proliferative breast disease suggest that recent alcohol intake does not increase the risk of proliferative breast disease [13-15] and may in fact be inversely related to it, with estimates ranging from a 10% [13] to a 77% [14] reduction in risk comparing the highest category of alcohol consumption with nondrinkers. These results, along with those from the current study, suggest that the association observed between recent alcohol consumption and breast cancer may not be mediated through proliferative breast disease. However, Byrne and colleagues examined alcohol consumption between the ages of 18 and 22 and proliferative breast disease and reported a 30% increased risk among women consuming ≥ 15 g of alcohol per day compared with nondrinkers (RR = 1.33, 95% CI 1.05 to 1.69) [13], suggesting that only very early alcohol consumption may affect proliferative BBD.
The progression from tumor initiation to breast cancer is not well defined. One hypothesis is that the pathway from normal tissue to breast cancer arises from a series of preinvasive lesions: benign proliferative changes, atypical hyperplasia, and carcinoma in situ. If all breast cancers arise from benign lesions, the results of this study, in conjunction with other data, imply that there may be a narrow window of time when alcohol consumption affects breast cancer risk. An alternative explanation may be that there are multiple pathways to breast cancer and pathways involving proliferative benign breast lesions may not be influenced by alcohol consumption. A third possibility is that the magnitude of risk associated with benign proliferative lesions, and in particular atypical hyperplasia, is so much greater than the effect of alcohol that we were underpowered to detect more subtle increases in risk attributable to alcohol in this study.
Although this is one of the largest studies of its kind, the number of women with proliferative breast disease and high alcohol consumption was small. We were underpowered to examine this relation in greater detail with regard to menopausal status and estrogen receptor status of tumors. A potential concern of the study is that the final study population represents 37% of those originally selected. The study protocol required collection of pathologic specimens in order to have unified review of histologic sections. The major limitation of this is that many hospitals routinely destroy specimens after 5 or 10 years. As a result, many potential cases and controls were excluded, which could result in potential selection bias. There were no significant differences in reasons for loss comparing cases to controls, indicating that any differences that do occur are likely to be due to chance.
Conclusion
In this study, we observed that women with proliferative BBD experienced no additional risk of breast cancer with increased alcohol consumption. It remains unclear whether alcohol acts early, acts late, or has multiple effects on breast cancer risk. Future studies with greater numbers should examine menopausal and hormone receptor status of the cancers in order to provide additional information on both timing of relevant exposure and biologic mechanisms.
Abbreviations
BBD = benign breast disease; CI = confidence interval; NHS (I, II) = Nurses' Health Study (I, II); OR = odds ratio; RR = relative risk.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
RMT conducted the analysis and wrote the manuscript. CB participated in the concept design and data collection. HJB participated in the writing and editing of the manuscript. BR provided statistical support. SJS participated in the concept and design, data collection, and manuscript editing. JLC participated in the concept and design, data collection, and manuscript editing. GAC provided funding for the study and participated in the concept and design, data collection, and manuscript editing. All authors read and approved the final manuscript.
Acknowledgements
Supported by Public Health Service Grants CA087969, CA046475, and CA050385 from the National Cancer Institute, National Institutes of Health, US Department of Health and Human Services. Dr Graham Colditz is supported in part by an American Cancer Society Cissy Hornung Clinical Research Professorship. We thank participants of the Nurses' Health Studies I and II for their outstanding dedication and commitment to the study and Ms Barbara DeSouza and Mr Gregory Kirkner for their assistance in specimen retrieval and tracking.
Figures and Tables
Figure 1 Schematic description of relation between alcohol consumption, proliferative benign breast disease (BBD), and breast cancer. Solid lines indicate well-established associations and dotted lines indicate less-established relations.
Table 1 Cases and controls included in benign breast disease pathologic specimen collectiona
Cases Controls Selection of participants
1,080 4,353 Participants initially selected
-29 -85 Participant died or withdrew from study subsequent to selection
-265 -1,090 We did not pursue collection of specimens, because:
• participant denied having biopsy or had cyst aspiration only
• participant could not recall hospital or hospital no longer exists
• participant denied permission to seek specimens
• biopsy contained cancer
-379 -1,406 We were unable to obtain requested samples
• specimens were no longer available or destroyed
• hospital had no record of the patient or biopsy
• hospital had no record of the patient or biopsy
-33 -113 Specimens received were not usable for study
• slides were not of good quality/not evaluable
• specimens did not contain breast tissue
• were from wrong person or incorrect date
374 1,659 Eligible participants for whom we collected usable specimens
aNurses' Health Study I (1976 to 1994) and II (1989 to 1995).
Table 2 Characteristics of controls in the nested case–control study according to benign breast disease category
Category of benign breast disease
Descriptive characteristic Nonproliferative Proliferative without atypia Atypical hyperplasia
Mean
Age at biopsy (y) 42.5 43.7 50.1
Year of biopsyb 1979 1980 1983
Age at menarche (y) 12.7 12.5 12.9
Age at first birthc (y) 24.8 24.9 25.0
Age at menopaused 46.6 46.5 47.2
Body mass index, kg/m2 25.1 24.6 24.5
Parity 2.8 2.7 2.9
Recent alcohol consumption, g/day 4.8 5.4 6.6
Frequency, %
Premenopausal at biopsy 76.5 76.1 57.1
Family history of breast cancer 11.2 13.3 18.3
Nulliparous 6.5 8.2 5.6
Natural menopause d 46.8 50.7 48.5
Bilateral oophorectomyd 20.1 16.8 17.5
aNurses' Health Study I (1976 to 1994) and II (1989 to 1995). bMedian. cAmong parous women only. dAmong postmenopausal women.
Table 3 Effect of category of benign breast disease histology on breast cancer riska
Benign breast disease category Cases No. (%) Controls No. (%) OR (95% CI)b OR (95% CI)c
Nonproliferative 69(24.5) 474 (38.8) 1.00 (Ref) 1.00 (Ref)
Proliferative without atypia 140 (49.7) 623 (50.9) 1.59 (1.16–2.18) 1.54 (1.12–2.11)
Atypical hyperplasia 73 (25.9) 126 (10.3) 4.54 (3.03–6.78) 4.43 (2.93–6.69)
Total 282 1,223
aNurses' Health Study I (1976 to 1994) and II (1989 to 1995). bAdjusted for age, year of biopsy, and follow-up interval. cAdjusted for age, year of biopsy, and follow-up interval, age at menarche, body mass index, weight change since age 18, age at first birth/parity, menopausal status/type of menopause, duration of postmenopausal hormone use. CI, confidence interval; OR, odds ratio; Ref, reference group for statistical comparisons.
Table 4 Combined effect of alcohol and category of benign breast disease histology on breast cancer riska
Jointly-classifiedb Stratifiedc>
Category of BBD Recent alcohol consumption (g/day) Cases Controls OR (95% CI)d OR (95% CI)e OR (95% CI)d OR (95% CI)e
Nonproliferative
0 21 178 1.00 (Ref) 1.00 (Ref) 1.00 (Ref) 1.00 (Ref)
<5 27 165 1.41 (0.76–2.60) 1.42 (0.77–2.65) 1.34 (0.72–2.50) 1.46 (0.75–2.85)
5–15 14 95 1.27 (0.62–2.62) 1.25 (0.60–2.60) 1.31 (0.63–2.74) 1.15 (0.53–2.53)
≥ 15 7 36 1.67 (0.66–4.27) 1.67 (0.65–4.34) 1.85 (0.71–4.84) 1.90 (0.67–5.39)
P for trendf 0.13 0.20
Proliferative without atypia
0 46 205 2.04 (1.17–3.57) 1.99 (1.13–3.50) 1.00 (Ref) 1.00 (Ref)
<5 52 235 1.91 (1.11–3.29) 1.82 (1.05–3.16) 0.99 (0.63–1.55) 0.99 (0.62–1.56)
5–15 34 123 2.43 (1.34–4.41) 2.40 (1.31–4.39) 1.17 (0.70–1.94) 1.14 (0.67–1.91)
≥ 15 8 60 1.15 (0.48–2.75) 1.09 (0.45–2.62) 0.59 (0.26–1.33) 0.61 (0.26–1.41)
P for trendf 0.49 0.58
Proliferative with atypia
0 23 40 5.57 (2.77–11.21) 5.58 (2.74–11.38) 1.00 (Ref) 1.00 (Ref)
<5 30 45 6.65 (3.43–12.90) 6.48 (3.31–12.70) 1.28 (0.62–2.67) 1.29 (0.60–2.78)
5–15 14 25 5.35 (2.39–12.01) 5.26 (2.30–12.04) 1.03 (0.42–2.51) 1.02 (0.39–2.68)
≥ 15 6 16 3.65 (1.27–10.47) 3.46 (1.19–10.09) 0.66 (0.22–1.97) 0.59 (0.19–1.89)
P for trendf 0.21 0.17
aNurses' Health Study I (1976 to 1994) and II (1989 to 1995). b Joint classification of BBD category and alcohol consumption where the reference group is women with nonproliferative BBD and 0 alcohol consumption. c The analysis is stratified by type of BBD where the reference group is women with 0 alcohol consumption within each category of BBD. dAdjusted for age at diagnosis, year of biopsy, and follow-up interval. eAdjusted for age at diagnosis, year of biopsy, and follow-up interval, age at menarche, body mass index quartiles, family history of breast cancer, parity/age at first birth, menopausal status/type of menopause, duration of postmenopausal hormone use. fTest for trend based on Wald test for continuous alcohol consumption. BBD, benign breast disease; CI, confidence interval; OR, odds ratio; Ref, reference group for statistical comparisons.
==== Refs
Smith-Warner SA Spiegelman D Yaun SS van den Brandt PA Folsom AR Goldbohm RA Graham S Holmberg L Howe GR Marshall JR0 Alcohol and breast cancer in women: a pooled analysis of cohort studies JAMA 1998 279 535 540 9480365 10.1001/jama.279.7.535
Longnecker MP Alcoholic beverage consumption in relation to risk of breast cancer: meta-analysis and review Cancer Causes Control 1994 5 73 82 8123780
Singletary KW Gapstur SM Alcohol and breast cancer: review of epidemiologic and experimental evidence and potential mechanisms JAMA 2001 286 2143 2151 11694156 10.1001/jama.286.17.2143
Longnecker MP Berlin JA Orza MJ Chalmers TC A meta-analysis of alcohol consumption in relation to risk of breast cancer JAMA 1988 260 652 656 3392790 10.1001/jama.260.5.652
Ellison RC Zhang Y McLennan CE Rothman KJ Exploring the relation of alcohol consumption to risk of breast cancer Am J Epidemiol 2001 154 740 747 11590087 10.1093/aje/154.8.740
Brisson J Merletti F Sadowsky NL Twaddle JA Morrison AS Cole P Mammographic features of the breast and breast cancer risk Am J Epidemiol 1982 115 428 437 7064977
Boyd NF Lockwood GA Byng JW Tritchler DL Yaffe MJ Mammographic densities and breast cancer risk Cancer Epidemiol Biomarkers Prev 1998 7 1133 1144 9865433
Boyd NF Lockwood GA Martin LJ Byng JW Yaffe MJ Tritchler DL Mammographic density as a marker of susceptibility to breast cancer: a hypothesis IARC Sci Publ 2001 154 163 169 11220655
Atkinson C Warren R Bingham SA Day NE Mammographic patterns as a predictive biomarker of breast cancer risk: effect of tamoxifen Cancer Epidemiol Biomarkers Prev 1999 8 863 866 10548313
Boyd NF Connelly P Byng J Yaffe M Draper H Little L Jones D Martin LJ Lockwood G Tritchler D Plasma lipids, lipoproteins, and mammographic densities Cancer Epidemiol Biomarkers Prev 1995 4 727 733 8672989
Vachon CM Kuni CC Anderson K Anderson VE Sellers TA Association of mammographically defined percent breast density with epidemiologic risk factors for breast cancer (United States) Cancer Causes Control 2000 11 653 662 10977110 10.1023/A:1008926607428
Vachon CM Kushi LH Cerhan JR Kuni CC Sellers TA Association of diet and mammographic breast density in the Minnesota breast cancer family cohort Cancer Epidemiol Biomarkers Prev 2000 9 151 160 10698475
Byrne C Webb PM Jacobs TW Peiro G Schnitt SJ Connolly JL Willett WC Colditz GA Alcohol consumption and incidence of benign breast disease Cancer Epidemiol Biomarkers Prev 2002 11 1369 1374 12433713
Rohan TE Jain M Miller AB Alcohol consumption and risk of benign proliferative epithelial disorders of the breast: a case-cohort study Public Health Nutr 1998 1 139 145 10933411
Rohan TE Cook MG Alcohol consumption and risk of benign proliferative epithelial disorders of the breast in women Int J Cancer 1989 43 631 636 2703271
van den Brandt PA Goldbohm RA van 't Veer P Alcohol and breast cancer: results from The Netherlands Cohort Study Am J Epidemiol 1995 141 907 915 7741120
Horn-Ross PL Canchola AJ West DW Stewart SL Bernstein L Deapen D Pinder R Ross RK Anton-Culver H Peel D Patterns of alcohol consumption and breast cancer risk in the California Teachers Study cohort Cancer Epidemiol Biomarkers Prev 2004 13 405 411 15006916
Willett WC Stampfer MJ Colditz GA Rosner BA Hennekens CH Speizer FE Moderate alcohol consumption and the risk of breast cancer N Engl J Med 1987 316 1174 1180 3574368
Page DL Dupont WD Rogers LW Rados MS Atypical hyperplastic lesion of the female breast. A long-term follow-up study Cancer 1985 55 2698 2708 2986821
London ST Connolly JL Schnitt SJ Colditz GA A prospective study of benign breast disease and the risk of breast cancer JAMA Med Assoc 1992 267 941 944 10.1001/jama.267.7.941
Carter CL Corle DK Micozzi MS Schatzkin A Taylor PR A prospective study of the development of breast cancer in 16,692 women with benign breast disease Am J Epidemiol 1988 128 467 477 3414655
Colditz GA The Nurses' Health Study: findings during 10 years of follow-up of a cohort of U.S. women Curr Probl Obstet Gynecol Fertil 1990 13 129 174
Dupont WD Page DL Risk factors for breast cancer in women with proliferative breast disease N Engl J Med 1985 312 146 151 3965932
Giovannucci E Colditz G Stampfer MJ Rimm EB Litin L Sampson L Willett WC The assessment of alcohol consumption by a simple self-administered questionnaire Am J Epidemiol 1991 133 810 817 2021148
Byrne C Connolly JL Colditz GA Schnitt SJ Biopsy confirmed benign breast disease, postmenopausal use of exogenous female hormones, and breast carcinoma risk Cancer 2000 89 2046 2052 11066044 10.1002/1097-0142(20001115)89:10<2046::AID-CNCR3>3.0.CO;2-F
Holmberg L Baron JA Byers T Wolk A Ohlander EM Zack M Adami HO Alcohol intake and breast cancer risk: effect of exposure from 15 years of age Cancer Epidemiol Biomarkers Prev 1995 4 843 847 8634655
Freudenheim JL Marshall JR Graham S Laughlin R Vena JE Swanson M Ambrosone C Nemoto T Lifetime alcohol consumption and risk of breast cancer Nutr Cancer 1995 23 1 11 7739910
Tjonneland A Christensen J Thomsen BL Olsen A Stripp C Overvad K Olsen JH Lifetime alcohol consumption and postmenopausal breast cancer rate in Denmark: a prospective cohort study J Nutr 2004 134 173 178 14704313
| 15987462 | PMC1175067 | CC BY | 2021-01-04 16:54:50 | no | Breast Cancer Res. 2005 May 16; 7(4):R555-R562 | utf-8 | Breast Cancer Res | 2,005 | 10.1186/bcr1039 | oa_comm |
==== Front
Breast Cancer ResBreast Cancer Research1465-54111465-542XBioMed Central London bcr10401598745810.1186/bcr1040Research ArticleCYP17 genetic polymorphism, breast cancer, and breast cancer risk factors: Australian Breast Cancer Family Study Chang Jiun-Horng [email protected] Dorota M [email protected] Xiaoqing [email protected] Gillian S [email protected] Mark A [email protected] Roger L [email protected] Melissa C [email protected] Margaret RE [email protected] Graham G [email protected] Georgia [email protected] John L [email protected] Amanda B [email protected] Centre for Genetic Epidemiology, The University of Melbourne, Victoria, Australia2 Cancer and Cell Biology Division, Queensland Institute of Medical Research, Herston, Queensland, Australia3 Department of Pathology, The University of Melbourne, Parkville, Victoria, Australia4 Cancer Epidemiology Research Unit, The Cancer Council of New South Wales, Sydney, New South Wales, Australia5 Department of Preventive and Social Medicine, University of Otago, Dunedin, New Zealand6 Cancer Epidemiology Centre, The Cancer Council Victoria, Carlton, Victoria, Australia2005 12 5 2005 7 4 R513 R521 16 12 2004 24 2 2005 30 3 2005 13 4 2005 Copyright © 2005 Chang et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Introduction
Because CYP17 can influence the degree of exposure of breast tissues to oestrogen, the interaction between polymorphisms in this gene and hormonal risk factors is of particular interest. We attempted to replicate the findings of studies assessing such interactions with the -34T→C polymorphism.
Methods
Risk factor and CYP17 genotyping data were derived from a large Australian population-based case-control-family study of 1,284 breast cancer cases and 679 controls. Crude and adjusted odds ratio (OR) estimates and 95% confidence intervals (CIs) were calculated by unconditional logistic regression analyses.
Results
We found no associations between the CYP17 genotype and breast cancer overall. Premenopausal controls with A2/A2 genotype had a later age at menarche (P < 0.01). The only associations near statistical significance were that postmenopausal women with A1/A1 (wild-type) genotype had an increased risk of breast cancer if they had ever used hormone replacement therapy (OR 2.40, 95% CI 1.0 to 5.7; P = 0.05) and if they had menopause after age 47 years (OR 2.59, 95% CI 1.0 to 7.0; P = 0.06). We found no associations in common with any other studies, and no evidence for interactions.
Conclusion
We observed no evidence of effect modification of reproductive risk factors by CYP17 genotype, although the experiment did not have sufficient statistical power to detect small main effects and modest effects in subgroups. Associations found only in subgroup analyses based on relatively small numbers require cautious interpretation without confirmation by other studies. This emphasizes the need for replication in multiple and large population-based studies to provide convincing evidence for gene–environment interactions.
==== Body
Introduction
The association between exposure to endogenous and exogenous steroid hormones and breast cancer risk is well established [1]. Consequently, genetic polymorphisms in genes involved with hormone-metabolizing pathways have been widely studied for evidence of their contribution to breast cancer risk [2,3]. One such candidate gene is CYP17 on chromosome 10q24.3, which encodes the enzyme cytochrome P450c17α (17α-hydroxylase; 17/20 lyase). P450c17α functions at two different points in the steroid biosynthesis pathway; the 17α hydroxylase activity can convert progesterone to 17α-hydroxyprogesterone, and the 17/20 lyase function may further convert 17α-hydroxyprogesterone to androstenedione (the precursor of both oestrone and testosterone) [4].
One common polymorphism in CYP17 has been extensively studied [5-23]. It is a T→C nucleotide substitution 34 base pairs upstream of the translation initiation site in the 5' promoter region. A subset of the literature refers to the wild-type T allele as A1, and the variant C allele as A2. The C allele creates an additional Sp1-type (CCACC box) promoter site, and although it was initially suggested to increase expression of the gene [9], a subsequent study did not observe binding to the human transcription factor Sp-1 [16]. There is conflicting evidence indicating that the CYP17 -34T→C polymorphism might influence endogenous steroid hormone levels [11,24-31], and the CC genotype has also been reported to be associated with the relative abundance of the 2OHE and 16α OHE forms of oestrogen [32]. A recent study also found the polymorphism associated with higher levels of DHEAS in premenopausal women and higher levels of oestradiol in postmenopausal women [33].
Although a few studies have found evidence for an association between this polymorphism and risk of breast cancer [7,9,19,23], these positive associations were observed for specific subgroups of cases defined by tumour aggressiveness, age at onset, or family history of breast cancer. Two recent meta-analyses [3,34] showed no overall association of breast cancer with the C (A2) variant, when comparing allele frequencies, or genotypes defined by these alleles under a dominant or recessive model. Results were consistently null in different ethnic groups [34].
As CYP17 may influence the degree of exposure of breast epithelial cells to oestrogen, the possibility that the effects of different hormonal risk factors is dependent on different CYP17 genotype is of particular interest. Some studies have suggested that CYP17 genotype is associated with hormonal risk factors, and/or that the association between breast cancer and hormonal risk factors depends on CYP17 genotypes. That is, CYP17 genotype may be an effect modifier. The hormonal risk factors examined in this manner have included age at menarche, age at first birth, use of oral contraceptives, age at menopause, and hormonal replacement therapy. So far, studies examining these gene–environment interactions or effect modifications have generally been small and have reported conflicting results [5,8,9,12,13,15,18,19,21,22,27,35-37]. For example, the CYP17 variant was significantly associated with earlier age at menarche in only two of eight reports, and an effect of later age at menarche (at least 13 years) limited to women with the wild-type T (A1) homozygous genotype was evident in only 4 of 11 reports. Moreover, as alluded to in a recent review of studies of the CYP17 polymorphisms and hormone levels [38], the published literature is compromised because many reported 'associations' were not statistically significant, and the possibility of publication bias in selective reporting of such data cannot be excluded.
We previously published Australian data on the overall relationship between this CYP17 genetic polymorphism and the risk of breast cancer before the age of 40 years [23]. In the present study of a larger sample of women under 60 years, we considered the issue of effect modification. We studied reproductive and hormone-related factors previously documented to be putative effect modifiers [5,9,11,19,24,35]. We used the same categorization as in the Western New York Breast Cancer Study (WNYBCS) [5], the most comprehensive study assessing CYP17 genotype as a potential effect modifier, in an attempt to replicate findings with an independent data set.
Materials and methods
Subjects
The Australian Breast Cancer Family Study (ABCFS) is a population-based case-control-family study of breast cancer before the age of 60 years [39-41]. Sampling of cases was stratified by age at onset, and half were diagnosed before age 40 years, so the study is predominantly of premenopausal women. Cases were women living in Melbourne or Sydney diagnosed with a first primary breast cancer, identified through the Victoria and New South Wales cancer registries. Controls were women with no previous breast cancer selected from the electoral roll (adult registration for voting is compulsory in Australia) by a stratified random sampling, frequency-matched for age. Questionnaires used to measure exposure to risk factors and family cancer history have been described previously [23,40,41]. Ethical approvals for the ABCFS and this genotyping study were obtained from The University of Melbourne, the Cancer Councils of Victoria and New South Wales, and the Queensland Institute of Medical Research.
For the purpose of these analyses, subjects were restricted to the women who identified themselves as being white/Caucasian (details in [23]). Molecular analyses (see [41]) have so far identified 41 Caucasian cases carrying a deleterious germline mutation in either BRCA1 or BRCA2, and these subjects have been excluded from the analyses.
Molecular analysis
As described in detail previously [23], the CYP17 -34T→C polymorphism was measured in DNA extracted from cases and controls with the use of the AIB1 Prism 7700 Sequence Detection System. DNA was extracted from peripheral blood cells by using salt extraction methods for those recruited before 1995 [42] and with the use of spin columns (Mini blood spin columns; Qiagen, Hilden, Germany) for those recruited from 1995 onwards.
Statistical methods
As one purpose of this study was to try to replicate the effect modification effects previously reported in the literature that are consistent with an increased exposure to endogenous oestrogen associated with genotypes defined by the C (A2) variant [5,9,11,19,24,35], we categorized risk factors collected in our study in an identical manner to the most comprehensive study assessing effect modification, the WYNBCS [5]. As far as possible, analyses were adjusted for the same factors and the results presented similarly. Consequently, reproductive variables were categorized as follows: age at menarche (less than 13 years, 13 years or more); age at first birth (less than 25 years, 25 years or more); ever use of oral contraceptives (yes, no); family history of breast cancer (yes, no for any first-degree relative reported to have had breast cancer); age at menopause (less than 48 years, 48 years or more); and ever use of hormone replacement therapy (HRT; yes, no).
The Hardy–Weinberg equilibrium assumption was assessed by comparing the genotype frequencies with those expected on the basis of the observed allele frequencies and random mating by using the Pearson χ2 distribution with one degree of freedom. The associations between risk of breast cancer and risk factors and CYP17 genotypes were assessed by multiple linear logistic regression, adjusting for the potential confounders reference age, body mass index, family history, education level, country of birth, benign breast disease, and age at menopause in postmenopausal women. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated with and without adjustment for measured risk factors. Logistic regression was also used to assess, in controls, the associations of hormone-related risk factors with genotype after adjusting for the above potential confounders. The statistical significance of interaction terms was assessed by the likelihood ratio test. All analyses were conducted with Stata version 8.0. All statistical tests were two-sided and the P values quoted are nominal; that is, no attempt was made to adjust for multiple comparisons, either in terms of the number of covariates or in terms of the number of modes of inheritance being considered.
A visual comparison between the results of the ABCFS and those of the WNYBCS was conducted by plotting the corresponding log OR estimates from each study against one another, using R version 1.6.2. The size of the points was proportional to the average of the inverse of standard errors of the two studies for that particular risk factor's estimate, so that larger points were those for which there was more precision. A positive correlation between the data points would be evidence for replication of findings.
Results
Analyses were conducted for the 1,284 cases and 679 controls genotyped for CYP17, including 1,572 premenopausal women (mean age 38.3 years) and 391 postmenopausal women (mean age 52.3 years). There was no evidence of deviation from Hardy–Weinberg equilibrium in controls (P = 0.95), but in cases there was marginally significant evidence of such deviation, with a deficiency of heterozygotes (565 observed versus 605.3 expected; P = 0.02). The C (A2) allele frequency in cases was 0.38 (SEM 0.001) in cases, and 0.37 (SEM 0.01) in controls (P for difference = 0.8).
Table 1 shows that there were no associations between breast cancer risk and CYP17 genotypes under codominant inheritance, among either postmenopausal or premenopausal women, with or without adjusting for covariates. For dominant inheritance, the adjusted OR estimates were 0.94 (95% CI 0.76 to 1.18) and 1.31 (95% CI 0.84 to 2.05) among premenopausal and postmenopausal women, respectively. For recessive inheritance, the adjusted OR estimates were 1.15 (95% CI 0.85 to 1.56) and 1.12 (95% CI 0.60 to 2.09) among premenopausal and postmenopausal women, respectively. We also examined risk for women stratified by report of family history, because our previous examination of a subset of the data found evidence for increased risk associated with the CC (A2/A2) genotype among women reporting a first-degree or second-degree family history of breast cancer. The OR for the CC (A2/A2) genotype was 1.10 (95% CI 0.81 to 1.49) for no family history and 2.46 (95% CI 0.70 to 8.60) for family history under a codominant model of inheritance, and 1.13 (95% CI 0.85, 1.50) for no family history and 2.37 (95% CI 0.72, 7.79) for family history under a recessive model of inheritance.
We excluded the 41 Caucasian cases with a known BRCA1 or BRCA2 germline mutation; because these mutations are associated with at least a 10-fold increased breast cancer risk [43], for more than 90% of them the cause of their disease was their germline mutation (that is, less than 10% are likely to be phenocopies). It is possible that their CYP17 genotype could have a modifying effect on their disease risk; however, the frequencies of the TT (A1/A1), TC (A1/A2) and CC (A2/A2) genotypes were 41% (n = 17), 41% (17) and 18% (7), respectively, in carriers, similar to those of 40% (513), 44% (565), and 16% (206) observed in the non-carrier cases (P = 0.95). The C (A2) allele frequency was 0.38 in both carriers and non-carriers.
Table 2 shows the distribution among controls of the CYP17 genotypes defined by the polymorphism, under a dominant model, in relation to reproductive and hormonally related risk factors, and family history status based on first-degree relatives, stratified by menopause status. The only significant association was between age at menarche and CYP17 genotype in the premenopausal women (P = 0.002), such that controls with the TT (A1/A1) genotype were more likely to have an age at menarche of less than 13 years. There was no evidence that women with a T (A1) allele were more likely to have used HRT. The strengths of the estimated associations between genotype and risk factor were little changed by also adjusting for potential confounders.
Table 3 shows that there was nominally significant evidence that ever use of HRT was associated with an increased risk of breast cancer among all women (OR 1.86; 95% CI 1.11 to 3.12; P = 0.02). This effect was significant among women homozygous for the T (A1) allele (OR 2.40; 95% CI 1.01 to 5.70; P = 0.05), but not significant for women with at least one C (A2) allele (OR 1.93; 95% CI 0.93 to 4.02; P = 0.08); the two estimates were not significantly different from one another (P = 0.7). There was at best marginally significant evidence that later age at menopause was associated with an increased breast cancer risk in women homozygous for the T (A1) allele (OR 2.59; 95% CI 0.97 to 6.95; P = 0.06), but not in women with at least one C (A2) allele, and the difference in risk estimate by genotype was also not statistically significant (P = 0.06).
We compared the adjusted OR estimates of each risk factor and breast cancer risk for each CYP17 genotype presented in Table 3 with those reported in the literature. A visual comparison between the results of the ABCFS and those of the WNYBCS [5] is shown in Fig. 1. This graphical presentation reveals no association between the point estimates and therefore no evidence for consistency in the estimates overall, or that any one or more findings were strong or statistically significant in both studies. The most consistent finding was for an increased risk in postmenopausal women with older age at menopause and A1/A1 genotype, but evidence was weak in the present study and not significant in the WNYBCS [5]. Other studies have examined association between genotype and menopausal status [15] or age at menopause [37], or risk associated with genotype in subgroups stratified by menopausal status [12], and none have reported statistically significant findings. Similarly, we failed to confirm other positive reports of associations with hormonal risk factors, or of effect modification, as detailed below.
Discussion
Consistent with most previous studies and a recent meta-analysis was our failure to find any evidence for an association between the CYP17 genotype defined by the – 34 promoter region T→C nucleotide-substitution polymorphism and risk of breast cancer overall, or within premenopausal or postmenopausal women, whether the genotype be defined under a codominant or a recessive mode of inheritance. Specifically, we failed to confirm our own positive finding [23] of a significantly increased risk associated with the CC (A2/A2) genotype in women reporting a positive family history.
When we examined associations between CYP17 genotypes and hormone-related risk factors in controls, the only significant finding was an increased risk for older age at menarche among premenopausal women with the C (A2) allele. This finding is not consistent with previous positive reports, or with the hypothesis that the C (A2) allele is associated with increased endogenous oestrogen levels and an earlier age at menarche. The point estimates for the effects on breast cancer risk of later age at menopause and HRT use were stronger among women carrying the TT (A1/A1) genotype, but the interaction terms were not statistically significant.
Other studies investigating the influence of CYP17 genotype on hormonal risk factors have found varying results. For example, the CYP17 variant was associated with earlier age at menarche in only two of eight reports [8,9,13,18,21,27,36,37]. Positive association with early age at first birth was observed in two of three reports [5,8,9,11,21], and only single studies have reported significant associations with decreased use of HRT [35] and with decreased difficulty in becoming pregnant [5].
Results from previous studies on the effect modification of reproductive risk factors by CYP17 genotypes have also been conflicting. The most consistent association reported was an effect modification of age at menarche. Four studies have presented evidence that the protective effect for a later age at menarche (at least 13 years) was mainly limited to women with the wild-type homozygous genotype [5,9,11,19], but in two of these studies the effect was observed only in premenopausal women [5,19], and seven other studies have failed to confirm these results [8,12,13,18,21,27,36]. Reports of significant associations between risk and age at first birth within strata of CYP17 genotype were in opposing directions from two studies reporting such results [11,21].
Our larger study found no association between breast cancer and age at menarche or age at first birth within any of the variant genotypes, either overall or within postmenopausal or premenopausal women. Although we have found at best marginally significant associations between breast cancer risk and both HRT use and age at menopausal status within some CYP17 genotype groups, these must be interpreted with caution. We considered both premenopausal and postmenopausal women, seven risk factors, and two genotype groups, so by chance alone we would expect to find a few significant results even if there were no real effects.
Perhaps more convincing evidence against a role of this CYP17 variant in modulating endogenous oestrogen levels and associated breast cancer risk factors is the results of a recent study of 1,975 postmenopausal women, which found no association between CYP17 genotypes defined by several polymorphisms and mean levels of sex hormones, in particular oestradiol, oestrone and sex-hormone-binding globulin [25]. This suggests that there might be little if any functional effect of the common CYP17 polymorphisms, at least among postmenopausal women, although there is a possibility that there might exist other variants and haplotypes associated with hormone levels and risk of breast cancer.
Although we have conducted a relatively large study, there are some limitations. Because the functionality, if any, of the polymorphism we have studied is not well established, it was not possible to specify a priori hypotheses about the likely existence and direction of interactions with risk factors. However, the few positive associations with serum hormone levels reported in the literature would suggest that the C (A2) variant would be associated with increased endogenous hormone levels, and we have observed significant effects both contradicting and in support of such an association. In addition, in this and other similar studies, there are multiple tests being conducted; the quoted P values are only nominal and should be interpreted accordingly. Consequently, we cannot claim with confidence that any of our 'significant' findings represent true effects. As there seems to be no overall effect of CYP17 genotype on breast cancer risk, finding any true interactions with breast cancer risk factors (should they exist) will require massive individual studies, or pooling of studies. The marginally significant deviation from Hardy–Weinberg equilibrium is unlikely to be due to genotyping error because cases and controls were genotyped at the same time on the same PCR plates, and thus any genotyping bias (and deviation from Hardy–Weinberg equilibrium) would be expected to be seen equally in both cases and controls, but this was not so. Furthermore, our PCR success rate was more than 99.5% for both cases and controls, and results were fully concordant for a subset of 168 duplicate DNAs for which PCR was successful.
The lack of significant associations in our data could be a consequence of not having sufficient statistical power to detect real effects. In terms of detecting a real effect on breast cancer risk associated with the homozygote A2/A2 (CC) genotype (whose frequency in controls is 14%; see Table 1), with the total sample sizes we studied we would have had 80% power at the 0.05 level of statistical significance to detect effects greater than the threshold of 1.5-fold. If analyses were restricted to postmenopausal women, this detection threshold would become about twofold, whereas if we were to consider only women with a family history the threshold would be about threefold to fourfold. Within the two genotype groups (for example A1/A1 compared with A1/A2 and A2/A2, which subdivides controls 39:61; see Table 1), the detection thresholds for effects associated with the risk factors (most of which are divided about 40:60 into two groups; see Table 3) would be a minimum of 1.8-fold, and much greater for the smaller subgroupings.
Our study sample was in general younger than that of the other studies reporting on possible effect modification by CYP17 genotype. Given that the younger the age at onset of breast cancer the stronger are the familial effects [41], one might expect the effects of genetic factors to be more pronounced in earlier onset disease. We therefore interpret our essentially null results as further support for the increasing body of evidence suggesting that there are no true associations or effect modifications, or at most weak ones, associated with this specific genetic variant of CYP17.
Conclusion
In summary, there are no known data to support a functional effect of this CYP17 polymorphism, and although lack of a demonstrated association with serum sex hormones does not exclude a possible functional effect, it does decrease enthusiasm for a possible modifying role of this CYP17 polymorphism on breast cancer risk. We have found little evidence to support previous reports of gene–environment interaction, in particular those of the most comprehensive study assessing relationships between CYP17 genotype and breast cancer hormonal risk factors [5]. Our post hoc power calculations show that we cannot exclude small main effects, or modest effects within subgroups. It is sobering to note that, if the aim of a study is to detect interactions, the size of the study will have to be at least four times larger than if attention were confined to detecting main effects of the same magnitude [44]. Given the concomitant issues of multiple testing, very large studies of tens of thousands of subjects will be required to evaluate gene-environment interactions, should they exist, and these will probably require the pooling of data from multiple studies (such as has been done by the Collaborative Group on Hormonal Factors in Breast Cancer).
Despite early enthusiasm and numerous reports of positive associations between genotypes, risk factors and breast cancer risk from small studies, few of these have held up under further scrutiny through being reproduced in larger studies. One educational example is of the association between breast cancer and a protein-truncating variant in CHK2 carried by 2.1% of women. Even in a study of more than 10,000 cases and 9,000 controls reporting a 2.3-fold increased risk for the CHK2 mutation and a 1.4-fold increased risk of family history, no nominally significant interactions were found between CHK2 genotypes and family history (P = 0.1) [45]. Future studies evaluating gene–environment interaction and cancer risk will need to be very large to produce credible evidence.
Abbreviations
ABCFS = Australian Breast Cancer Family Study; CI = confidence interval; HRT = hormone replacement therapy; OR = odds ratio; PCR = polymerase chain reaction; WNYBCS = Western New York Breast Cancer Study.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
J-HC conducted the analyses and led the manuscript preparation. DMG led the literature review and interpretation of the data. XC conducted the genotyping assays. GSD was responsible for data management. MAJ, RLM and JLH were responsible for development and supervision of the statistical analyses. MCS was responsible for the management of laboratory staff and biospecimens and participated in the quality control of the genotyping and data cleaning. MREMcC, GGG and JLH initiated the ABCFS and have been instrumental in the ongoing execution of these studies. GC-T initiated and obtained funding for the genotyping component of this study, and was involved in the supervision of laboratory work. ABS was responsible for designing and supervising the genotyping and laboratory work, genotype data cleaning, literature review and manuscript preparation. All authors read and approved the final manuscript.
Acknowledgements
We are grateful to the physicians, surgeons and oncologists who endorsed this project, to the interviewing staff, and to the many women and their relatives who participated in this research. The ABCFS was supported by the National Health and Medical Research Council (NHMRC) of Australia, the New South Wales Cancer Council, the Victorian Health Promotion Foundation (Australia), the Inkster-Ross Memorial Fund of the University of Otago, and the US National Cancer Institute, National Institutes of Health, under Request for Application CA-95-003 as part of the Breast Cancer Family Registries (CFRs). This work is supported by a NHMRC Program grant. ABS is funded by an NHMRC Career Development Award, and GC-T and JLH are NHMRC Senior and Senior Principal Research Fellows, respectively. JLH is a Group Leader of the Victorian Breast Cancer Research Consortium.
Figures and Tables
Figure 1 Log odds ratio estimates versus corresponding estimates from the Western New York Breast Cancer Study.
Table 1 CYP17 genotypes and breast cancer risk by menopausal status (BRCA1 and BRCA2 mutation carriers excluded)
Genotype Cases, n (%) Controls, n (%) Crude OR (95% CI) Adjusted OR (95% CI)a
Premenopausal
A1/A1 (TT) 414 (40) 201 (38) 1.00 1.00
A1/A2 (TC) 460 (44) 253 (48) 0.88 (0.70, 1.11) 0.90 (0.71, 1.14)
A2/A2 (CC) 169 (16) 75 (14) 1.09 (0.79, 1.51) 1.08 (0.78, 1.51)
Postmenopausal
A1/A1 (TT) 99 (41) 66 (44) 1.00 1.00
A1/A2 (TC) 105 (44) 64 (43) 1.09 (0.70, 1.70) 1.32 (0.82, 2.12)
A2/A2 (CC) 37 (15) 20 (13) 1.23 (0.66, 2.31) 1.29 (0.66, 2.52)
Pooled
A1/A1 (TT) 513 (40) 267 (39) 1.00 1.00
A1/A2 (TC) 565 (44) 317 (47) 0.93 (0.76, 1.14) 0.96 (0.77, 1.18)
A2/A2 (CC) 206 (16) 95 (14) 1.13 (0.85, 1.50) 1.12 (0.83, 1.51)
No family historyb
A1/A1 (TT) 462 (40) 249 (39) 1.00 1.00
A1/A2 (TC) 501 (44) 296 (47) 0.91 (0.74, 1.13) 0.95 (0.76, 1.18)
A2/A2 (CC) 182 (16) 91 (14) 1.08 (0.80, 1.45) 1.10 (0.81, 1.49)
With family historyb
A1/A1 (TT) 51 (37) 18 (42) 1.00 1.00
A1/A2 (TC) 64 (46) 21 (49) 1.08 (0.52, 2.23) 1.08 (0.49, 2.36)
A2/A2 (CC) 24 (17) 4 (9) 2.12 (0.65, 6.94) 2.46 (0.70, 8.60)
aOdds ratio (OR) adjusted for reference age, body mass index, family history defined by any first-degree relative who had breast cancer, state, education level, country of birth, benign breast lump and age at menopause in postmenopausal women.
bFamily history defined by any first-degree relative who had breast cancer.
Table 2 Breast cancer risk factors by CYP17 genotype (BRCA1 and BRCA2 mutation carriers excluded)
Factor Premenopausal (n = 529) Postmenopausal (n = 150) Pooled (pre- and postmenopausal) (n = 679)
CYP17 genotype... A1/A1 A1/A2 and A2/A2 A1/A1 A1/A2 and A2/A2 A1/A1 A1/A2 and A2/A2
Age at menarche
<13 years 95 (48) 112 (34)** 28 (42) 35 (42) 123 (46) 147 (36)***
≥13 years 105 (52) 216 (66) 38 (58) 49 (58) 143 (54) 265 (64)
Age at first birth
<25 years 52 (37) 93 (40) 37 (61) 40 (54) 89 (44) 133 (43)
≥25 years 89 (63) 142 (60) 24 (39) 34 (46) 113 (56) 176 (57)
Ever use oral contraceptives
No 12 (6) 20 (6)
Yes 189 (94) 308 (94)
Family history of breast cancer
No 189 (94) 309 (94) 60 (91) 78 (93) 249 (93) 387 (94)
Yes 12 (6) 19 (6) 6 (9) 6 (7) 18 (7) 25 (6)
Age at menopause
<48 years 40 (61) 45 (54)
≥48 years 26 (39) 39 (46)
Ever use HRT
No 28 (42) 33 (48)
Yes 38 (58) 51 (61)
Data are presented as n (%). **P = 0.002, ***P = 0.006. Information on difficulty getting pregnant was unavailable. HRT, hormone replacement therapy.
Table 3 Breast cancer risk and risk factors by CYP17 genotype (BRCA1 and BRCA2 mutation carriers excluded)
Factor All data A1/A1 A1/A2 and A2/A2
Case, n (%) Control, n (%) OR (95% CI) Case, n (%) Control, n (%) OR (95% CI) Case, n (%) Control, n (%) OR (95% CI)
Premenopausal
Age at menarche
<13 years 435 (42) 207 (39) 1.00 183 (45) 95 (47) 1.00 252 (40) 112 (34) 1.00
≥13 years 603 (58) 321 (61) 0.93 (0.7–1.2) 228 (55) 105 (53) 1.18 (0.8–1.8) 375 (60) 216 (66) 0.79 (0.6–1.1)
Age at first birth
<25 years 318 (40) 145 (39) 1.00 131 (42) 52 (37) 1.00 187 (39) 93 (40) 1.00
≥25 years 479 (60) 231 (61) 1.13 (0.9–1.5) 183 (58) 89 (63) 0.93 (0.6–1.5) 296 (61) 142 (60) 1.30 (0.9–1.8)
Ever use OC
No 87 (8) 32 (6) 1.00 32 (8) 12 (6) 1.00 55 (9) 20 (6) 1.00
Yes 955 (92) 497 (94) 0.82 (0.5–1.4) 382 (92) 189 (94) 0.75 (0.3–1.9) 573 (91) 308 (94) 0.82 (0.4–1.7)
Postmenopausal
Age at menarche
<13 years 101 (42) 63 (42) 1.00 44 (44) 28 (42) 1.00 57 (40) 35 (42) 1.00
≥13 years 140 (58) 87 (58) 0.92 (0.6–1.5) 55 (55) 38 (58) 0.83 (0.4–1.9) 85 (60) 49 (58) 0.84 (0.4–1.7)
Age at first birth
<25 years 122 (60) 77 (57) 1.00 47 (57) 37 (61) 1.00 75 (62) 40 (54) 1.00
≥25 years 79 (40) 58 (43) 0.98 (0.6–1.6) 35 (43) 24 (39) 1.16 (0.5–2.7) 44 (38) 34 (46) 0.84 (0.4–1.7)
Ever use HRT
No 79 (33) 61 (41) 1.00 31 (31) 28 (42) 1.00 45 (32) 33 (39) 1.00
Yes 165 (67) 89 (59) 1.86 (1.1–3.1)1 68 (69) 38 (58) 2.40 (1.0–5.7)2 97 (68) 51 (61) 1.93 (0.9–4.0)3
Age at menopause
<48 years 136 (56) 85 (57) 1.00 50 (51) 40 (61) 1.00 83 (58) 45 (37) 1.00
≥48 years 107 (44) 65 (43) 1.25 (0.7–2.2) 48 (49) 26 (39) 2.59 (1.0–7.0)4 59 (42) 39 (63) 0.80 (0.4–1.7)
Adjusted odds ratios (ORs) for premenopausal women are adjusted for age at menarche, age at first birth, ever use of OC, reference age, body mass index, family history defined by any first-degree relative who had breast cancer, state, education level, country of birth and benign breast lump disease. The same adjustment was performed for postmenopausal women except for ever use of hormone replacement therapy (HRT) instead of ever use of OC; adjustment was also made for age at menopause. 1P = 0.02, 2P = 0.05, 3P = 0.08, 4P = 0.06.
==== Refs
Pike MC Pearce CL Wu AH Prevention of cancers of the breast, endometrium and ovary Oncogene 2004 23 6379 6391 15322512 10.1038/sj.onc.1207899
Dunning A Healey C Pharoah P Teare M Ponder B Easton D A systematic review of genetic polymorphisms and breast cancer risk Cancer Epidemiol Biomarkers Prev 1999 8 843 854 10548311
de Jong MM Nolte IM te Meerman GJ van der Graaf WT Oosterwijk JC Kleibeuker JH Schaapveld M de Vries EG Genes other than BRCA1 and BRCA2 involved in breast cancer susceptibility J Med Genet 2002 39 225 242 11950848 10.1136/jmg.39.4.225
Miller WL Early steps in androgen biosynthesis: from cholesterol to DHEA Baillieres Clin Endocrinol Metab 1998 12 67 81 9890062 10.1016/S0950-351X(98)80461-8
Ambrosone CB Moysich KB Furberg H Freudenheim JL Bowman ED Ahmed S Graham S Vena JE Shields PG CYP17 genetic polymorphism, breast cancer, and breast cancer risk factors Breast Cancer Res 2003 5 R45 R51 12631398 10.1186/bcr570
Artamonov VV Liubchenko LN Shabanov MA Babenko OV Nemtsova MV Zaletaev DV Association of polymorphism of genetic markers of CYP19 and CYP17 with sporadic breast cancer. [In Russian] Mol Biol (Mosk) 2003 37 975 982 14714492
Bergman-Jungestrom M Gentile M Lundin AC Wingren S Association between CYP17 gene polymorphism and risk of breast cancer in young women Int J Cancer 1999 84 350 353 10404084 10.1002/(SICI)1097-0215(19990820)84:4<350::AID-IJC3>3.0.CO;2-L
Dunning AM Healey CS Pharoah PD Foster NA Lipscombe JM Redman KL Easton DF Day NE Ponder BA No association between a polymorphism in the steroid metabolism gene CYP17 and risk of breast cancer Br J Cancer 1998 77 2045 2047 9667690
Feigelson HS Coetzee GA Kolonel LN Ross RK Henderson BE A polymorphism in the CYP17 gene increases the risk of breast cancer Cancer Res 1997 57 1063 1065 9067272
Gudmundsdottir K Thorlacius S Jonasson JG Sigfusson BF Tryggvadottir L Eyfjord JE CYP17 promoter polymorphism and breast cancer risk in males and females in relation to BRCA2 status Br J Cancer 2003 88 933 936 12644832 10.1038/sj.bjc.6600839
Haiman CA Hankinson SE Spiegelman D Colditz GA Willett WC Speizer FE Kelsey KT Hunter DJ The relationship between a polymorphism in CYP17 with plasma hormone levels and breast cancer Cancer Res 1999 59 1015 1020 10070957
Hamajima N Iwata H Obata Y Matsuo K Mizutani M Iwase T Miura S Okuma K Ohashi K Tajima K No association of the 5' promoter region polymorphism of CYP17 with breast cancer risk in Japan Jpn J Cancer Res 2000 91 880 885 11011114
Helzlsouer KJ Huang HY Strickland PT Hoffman S Alberg AJ Comstock GW Bell DA Association between CYP17 polymorphisms and the development of breast cancer Cancer Epidemiol Biomarkers Prev 1998 7 945 949 9796641
Huang CS Chern HD Chang KJ Cheng CW Hsu SM Shen CY Breast cancer risk associated with genotype polymorphism of the estrogen-metabolizing genes CYP17, CYP1A1, and COMT: a multigenic study on cancer susceptibility Cancer Res 1999 59 4870 4875 10519398
Kuligina ES Togo AV Suspitsin EN Grigoriev MY Pozharisskiy KM Chagunava OL Berstein LM Theillet C Hanson KP Imyanitov EN CYP17 polymorphism in the groups of distinct breast cancer susceptibility: comparison of patients with the bilateral disease vs. monolateral breast cancer patients vs. middle-aged female controls vs. elderly tumor-free women Cancer Lett 2000 156 45 50 10840158 10.1016/S0304-3835(00)00436-5
Nedelcheva Kristensen V Haraldsen EK Anderson KB Lonning PE Erikstein B Karesen R Gabrielsen OS Borresen-Dale AL CYP17 and breast cancer risk: the polymorphism in the 5' flanking area of the gene does not influence binding to Sp-1 Cancer Res 1999 59 2825 2828 10383140
Tan W Qi J Xing DY Miao XP Pan KF Zhang L Lin DX Relation between single nucleotide polymorphism in estrogen-metabolizing genes COMT, CYP17 and breast cancer risk among Chinese women. [In Chinese] Zhonghua Zhong Liu Za Zhi 2003 25 453 456 14575568
Weston A Pan CF Bleiweiss IJ Ksieski HB Roy N Maloney N Wolff MS CYP17 genotype and breast cancer risk Cancer Epidemiol Biomarkers Prev 1998 7 941 944 9796640
Mitrunen K Jourenkova N Kataja V Eskelinen M Kosma VM Benhamou S Vainio H Uusitupa M Hirvonen A Steroid metabolism gene CYP17 polymorphism and the development of breast cancer Cancer Epidemiol Biomarkers Prev 2000 9 1343 1348 11142420
Miyoshi Y Iwao K Ikeda N Egawa C Noguchi S Genetic polymorphism in CYP17 and breast cancer risk in Japanese women Eur J Cancer 2000 36 2375 2379 11094312 10.1016/S0959-8049(00)00334-8
Wu AH Seow A Arakawa K Van Den Berg D Lee HP Yu MC HSD17B1 and CYP17 polymorphisms and breast cancer risk among Chinese women in Singapore Int J Cancer 2003 104 450 457 12584742 10.1002/ijc.10957
Feigelson HS McKean-Cowdin R Coetzee GA Stram DO Kolonel LN Henderson BE Building a multigenic model of breast cancer susceptibility: CYP17 and HSD17B1 are two important candidates Cancer Res 2001 61 785 789 11212283
Spurdle AB Hopper JL Dite GS Chen X Cui J McCredie MR Giles GG Southey MC Venter DJ Easton DF CYP17 promoter polymorphism and breast cancer in Australian women under age forty years J Natl Cancer Inst 2000 92 1674 1681 11036113 10.1093/jnci/92.20.1674
Feigelson HS Shames LS Pike MC Coetzee GA Stanczyk FZ Henderson BE Cytochrome P450c17alpha gene (CYP17) polymorphism is associated with serum estrogen and progesterone concentrations Cancer Res 1998 58 585 587 9485002
Haiman CA Hankinson SE Colditz GA Hunter DJ De Vivo I A polymorphism in CYP17 and endometrial cancer risk Cancer Res 2001 61 3955 3960 11358812
Dunning AM Dowsett M Healey CS Tee L Luben RN Folkerd E Novik KL Kelemen L Ogata S Pharoah PD Polymorphisms associated with circulating sex hormone levels in postmenopausal women J Natl Cancer Inst 2004 96 936 945 15199113
Garcia-Closas M Herbstman J Schiffman M Glass A Dorgan JF Relationship between serum hormone concentrations, reproductive history, alcohol consumption and genetic polymorphisms in pre-menopausal women Int J Cancer 2002 102 172 178 12385014 10.1002/ijc.10651
Miyoshi Y Ando A Ooka M Shiba E Taguchi T Tamaki Y Noguchi S Association of CYP17 genetic polymorphism with intra-tumoral estradiol concentrations but not with CYP17 messenger RNA levels in breast cancer tissue Cancer Lett 2003 195 81 86 12767515 10.1016/S0304-3835(02)00211-2
Gharani N Waterworth DM Williamson R Franks S 5' polymorphism of the CYP17 gene is not associated with serum testosterone levels in women with polycystic ovaries J Clin Endocrinol Metab 1996 81 4174 8923880 10.1210/jc.81.11.4174
Diamanti-Kandarakis E Bartzis MI Zapanti ED Spina GG Filandra FA Tsianateli TC Bergiele AT Kouli CR Polymorphism T→C (– 34 bp) of gene CYP17 promoter in Greek patients with polycystic ovary syndrome Fertil Steril 1999 71 431 435 10065777 10.1016/S0015-0282(98)00512-3
Daneshmand S Weitsman SR Navab A Jakimiuk AJ Magoffin DA Overexpression of theca-cell messenger RNA in polycystic ovary syndrome does not correlate with polymorphisms in the cholesterol side-chain cleavage and 17alpha-hydroxylase/C(17–20) lyase promoters Fertil Steril 2002 77 274 280 11821083 10.1016/S0015-0282(01)02999-5
Jernstrom H Vesprini D Bradlow HL Narod SA Re: CYP17 promoter polymorphism and breast cancer in Australian women under age forty years J Natl Cancer Inst 2001 93 554 555 11287451 10.1093/jnci/93.7.554
Hong C Thompson H Jiang C Hammond G Tritchler D Yaffe M Boyd N Association between the T27C polymorphism in the cytochrome P450 c17alpha (CYP17) gene and risk factors for breast cancer Breast Cancer Res Treat 2004 88 217 230 15609124 10.1007/s10549-004-0780-7
Ye Z Parry JM The CYP17 MspA1 polymorphism and breast cancer risk: a meta-analysis Mutagenesis 2002 17 119 126 11880540 10.1093/mutage/17.2.119
Feigelson HS McKean-Cowdin R Pike MC Coetzee GA Kolonel LN Nomura AM Le Marchand L Henderson BE Cytochrome P450c17alpha gene (CYP17) polymorphism predicts use of hormone replacement therapy Cancer Res 1999 59 3908 3910 10463580
Lai J Vesprini D Chu W Jernstrom H Narod SA CYP gene polymorphisms and early menarche Mol Genet Metab 2001 74 449 457 11749050 10.1006/mgme.2001.3260
Gorai I Tanaka K Inada M Morinaga H Uchiyama Y Kikuchi R Chaki O Hirahara F Estrogen-metabolizing gene polymorphisms, but not estrogen receptor-alpha gene polymorphisms, are associated with the onset of menarche in healthy postmenopausal Japanese women J Clin Endocrinol Metab 2003 88 799 803 12574216 10.1210/jc.2002-020353
Sharp L Cardy AH Cotton SC Little J CYP17 gene polymorphisms: prevalence and associations with hormone levels and related factors. A HuGE review Am J Epidemiol 2004 160 729 740 15466495 10.1093/aje/kwh287
Hopper JL Giles GG McCredie MRE Boyle P Background, rationale and protocol for a case-control-family study of breast cancer Breast 1994 3 79 86 10.1016/0960-9776(94)90003-5
McCredie MR Dite GS Giles GG Hopper JL Breast cancer in Australian women under the age of 40 Cancer Causes Control 1998 9 189 198 9578296 10.1023/A:1008886328352
Dite GS Jenkins MA Southey MC Hocking JS Giles GG McCredie MR Venter DJ Hopper JL Familial risks, early-onset breast cancer, and BRCA1 and BRCA2 germline mutations J Natl Cancer Inst 2003 95 448 457 12644538
Southey MC Batten LE McCredie MRE Giles GG Dite G Hopper JL Venter DJ Estrogen receptor polymorphism at codon 325 and risk of breast cancer in women before age forty J Natl Cancer Inst 1998 90 532 536 9539249 10.1093/jnci/90.7.532
Antoniou A Pharoah PD Narod S Risch HA Eyfjord JE Hopper JL Loman N Olsson H Johannsson O Borg A Average risks of breast and ovarian cancer associated with BRCA1 or BRCA2 mutations detected in case series unselected for family history: a combined analysis of 22 studies Am J Hum Genet 2003 72 1117 1130 12677558 10.1086/375033
Smith PG Day NE The design of case-control studies: the influence of confounding and interaction effects Int J Epidemiol 1984 13 356 365 6386716
Consortium CBCC-C CHEK2*1100delC and susceptibility to breast cancer: a collaborative analysis involving 10,860 breast cancer cases and 9,065 controls from 10 studies Am J Hum Genet 2004 74 1175 1182 15122511 10.1086/421251
| 15987458 | PMC1175068 | CC BY | 2021-01-04 16:54:50 | no | Breast Cancer Res. 2005 May 12; 7(4):R513-R521 | utf-8 | Breast Cancer Res | 2,005 | 10.1186/bcr1040 | oa_comm |
==== Front
Breast Cancer ResBreast Cancer Research1465-54111465-542XBioMed Central London bcr10421598745910.1186/bcr1042Research ArticleConnexin 43 mediated gap junctional communication enhances breast tumor cell diapedesis in culture Pollmann Mary-Ann [email protected] Qing [email protected] Dale W [email protected] Martin [email protected] Department of Anatomy and Cell Biology, The University of Western Ontario, London, Ontario, Canada2005 13 5 2005 7 4 R522 R534 13 1 2005 21 2 2005 31 3 2005 13 4 2005 Copyright © 2005 Pollmann et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Introduction
Metastasis involves the emigration of tumor cells through the vascular endothelium, a process also known as diapedesis. The molecular mechanisms regulating tumor cell diapedesis are poorly understood, but may involve heterocellular gap junctional intercellular communication (GJIC) between tumor cells and endothelial cells.
Method
To test this hypothesis we expressed connexin 43 (Cx43) in GJIC-deficient mammary epithelial tumor cells (HBL100) and examined their ability to form gap junctions, establish heterocellular GJIC and migrate through monolayers of human microvascular endothelial cells (HMVEC) grown on matrigel-coated coverslips.
Results
HBL100 cells expressing Cx43 formed functional heterocellular gap junctions with HMVEC monolayers within 30 minutes. In addition, immunocytochemistry revealed Cx43 localized to contact sites between Cx43 expressing tumor cells and endothelial cells. Quantitative analysis of diapedesis revealed a two-fold increase in diapedesis of Cx43 expressing cells compared to empty vector control cells. The expression of a functionally inactive Cx43 chimeric protein in HBL100 cells failed to increase migration efficiency, suggesting that the observed up-regulation of diapedesis in Cx43 expressing cells required heterocellular GJIC. This finding is further supported by the observation that blocking homocellular and heterocellular GJIC with carbenoxolone in co-cultures also reduced diapedesis of Cx43 expressing HBL100 tumor cells.
Conclusion
Collectively, our results suggest that heterocellular GJIC between breast tumor cells and endothelial cells may be an important regulatory step during metastasis.
==== Body
Introduction
Tumor metastasis is a multi-step process that involves the dissociation of tumor cells from the primary tumor followed by their entry into the circulation, extravasation from the vascular system and proliferation into tumor masses at secondary tissue sites. Both cell-cell and cell-matrix interactions are important regulators at different stages of this metastatic cascade [1-4]. For instance, loss of E-cadherin in bladder, prostate, breast and colorectal cancers, or integrins such as α2β1 and α5β1 in breast cancer and α6β1 in melanoma cells correlates with increased malignancy of tumor cells, indicating that the integrity of the primary tumor depends on signals from adjacent cells and from appropriate extracellular matrix ligands [1,2,5-8].
Extravasation of malignant cells often involves transendothelial migration (diapedesis) into tissues prior to forming secondary tumors. In contrast to diapedesis of leukocytes during inflammatory responses, little is known about the molecular mechanisms that regulate tumor cell diapedesis. The integrity of the endothelium depends mainly on the organization of interendothelial junctions, and tumor cells must traverse these junctions to extravasate across the endothelial barrier. Our previous studies demonstrated that diapedesis of melanoma cells is in part regulated by adhesion receptors present on both the vascular endothelium and the tumor cell [9]. A localized disruption in the endothelium of vascular endothelial (VE)-cadherin, α-catenin and platelet endothelial cell adhesion molecule-1 (PECAM-1) occurs at the site of tumor cell penetration, which is restored once the tumor cell completes diapedesis [9]. We further demonstrated that endothelial N-cadherin is necessary for the completion of melanoma cell diapedesis, suggesting that endothelial cells actively participate in the process of tumor cell diapedesis [9]. Integrins appear to play an important role during extravasation of tumor cells by binding to components in the extracellular matrix and on endothelial cells. The integrin αvβ3 regulates diapedesis of melanoma cells by binding to the immunoglobulin like adhesion receptor L1 on endothelial cells [10]. In prostate tumor cells, β3 integrins regulate diapedesis by binding to matrix components underneath the endothelium [11] and, in breast tumor cells, integrins β1, α5 and αvβ3 are involved in adhesion and migration through the extracellular matrix components vitronectin and fibronectin [12]. In addition, it has been reported that integrins mediate invasion and metastasis in murine melanoma cells and gliomas [13,14].
These previous studies suggest that signaling via cell-cell and cell-matrix adhesion receptors facilitates tumor cell diapedesis. Another possible mechanism by which tumor cells may communicate with endothelial cells to cross the endothelial barrier involves gap junctions [15-17]. Gap junctions are intercellular channels that mediate the direct intercellular exchange of secondary messengers, small metabolites or inorganic ions [18]. These channels are located at the plasma membrane and are composed of two hemichannels called connexons, each of which is assembled from six oligomerized protein subunits called connexins (Cx) [18,19]. To date the connexin family includes 20 members named according to their predicted molecular weights [20]. Connexin expression varies between cell types, and individual cells often express more than one connexin family member [18,21]. Although endothelial cells are heterogeneous according to the size of the vessel or the vascular bed of origin, they all express three different types of connexin, Cx43, Cx37 and Cx40 [22-26]. In normal human mammary glandular epithelium, Cx43 is found most often between mammary myoepithelial cells and in luminal cells, while adjacent luminal cells predominantly assemble Cx26 gap junctions [27,28]. Studies have demonstrated a decrease in gap junctional intercellular communication (GJIC) and connexin expression or aberrant connexin trafficking and assembly in primary mammary tumors [29-31]. Little is known, however, about connexin expression and GJIC in tumor cells en route to secondary metastatic sites outside the site of the primary diseased tissue. Previous studies have observed apparent gap junction formation between mammary tumor and vascular endothelial cells but the consequences of this interaction were not established [15,16]. In another study, Cx26 expressing BL6 mouse melanoma cells were found to form tumor/endothelial cell heterocellular gap junctions and these connexin expressing tumor cells had increased metastatic properties in vivo [17]. These findings are puzzling as Cx26 is not generally capable of forming heterotypic gap junctions with any of the endothelial expressed connexins, which include Cx40, Cx37 and Cx43 [32].
Previous studies investigating GJIC between tumor cells and endothelial cells focused primarily on tumor cell adhesion to the endothelium during the initial stages of diapedesis [15-17]. To address the hypothesis that GJIC may affect diapedesis, we used a non-metastatic cell line (HBL100) that is GJIC-deficient and does not express any known connexins in order to better evaluate a possible contribution of GIJC to diapedesis. Because endothelial cells abundantly express Cx43, we engineered HBL100 cells to express either wild-type Cx43 or a non-functional chimeric mutant of Cx43 to determine if the potential of HBL100 cells to cross the endothelium would change under both or either of these conditions. In this report we show that Cx43 expression up-regulates tumor cell diapedesis via a GJIC-dependent mechanism.
Materials and methods
Cell lines and culture conditions
Human microvascular endothelial cells (HMVECs) derived from the lung were purchased from Cambrex Biosciences Inc. (Walkersville, MD, USA) and cultured according to supplier's instructions in endothelial growth medium (EGM; Cambrex Biosciences Inc.), supplemented with 100 units/ml penicillin G sodium, and 100 μg/ml streptomycin sulphate (Invitrogen, Burlington, ON, Canada). Wild-type transformed mammary epithelial cells (HBL100) and HBL100 cells expressing Cx43 or Cx43 with green fluorescent protein (GFP) tagged to the amino terminal (GFP-Cx43) [33] were routinely cultured in D-MEM supplemented with 100 units/ml penicillin G sodium, 100 μg/ml streptomycin, 2 mM L-glutamine (Invitrogen) and 10% FBS (Sigma-Aldrich, Oakville, ON, Canada). All cells were maintained in a humidified chamber at 37°C containing 5% CO2.
Constructs, infection and transfection
HBL100 cells expressing Cx43 (HBL100Cx43) or empty vector (HBL100v) were engineered by retroviral infection as described by Qin et al. [34]. Briefly the cDNA of Cx43 inserted into the AP-2 retroviral vector was transfected into the 293GPG packaging cell line and 48 h later retroviral supernatant was collected and filtered through a 0.45 μm filter and used to infect HBL100 cells. Twenty-four hours after infection, medium was replaced with D-MEM and cells were subcultured as normal. More than 90% of tumor cells expressed Cx43 after three rounds of viral infection as determined by immunofluorescence [34].
HBL100 cells expressing non-functional GFP-Cx43 (HBL100GFP-Cx43) [33] were generated by transfecting 2.5 × 105 HBL100 cells with 2 μg of GFP-Cx43 using Metafectene (Biontex Laboratories GmbH, Munich, Germany). After 24 h, approximately 40% of cells expressed GFP-Cx43 and these cells were used for diapedesis assays.
Diapedesis assay
The diapedesis assay was carried out as previously described [9]. Briefly, 7.5 × 105 HMVECs were seeded onto 12 mm glass coverslips coated with matrigel (Becton-Dickenson, Bedford, MA, USA). HMVEC cells were allowed to settle for 2–3 h and form monolayers in a 37°C humidified chamber containing 5% CO2. Coverslips were then transferred to a 24-well plate and cultured for 48 h in EGM. The number of cells seeded produces an endothelial monolayer without requiring cell proliferation. Monolayers have intact adherens junctions as previously evaluated by VE-cadherin staining [9,35]. HBL100 tumor cells were labeled with 10 μg/ml 1,1'-dioctadecyl-3, 3,3', 3'-tetramethylindocarbocyanine percholate (DiI; Molecular Probes, Eugene, OR, USA), and harvested non-enzymatically using 2 mM EDTA (EMD Chemicals Inc., Gibbstown, NJ, USA). Approximately 7.5 × 104 cells were added to endothelial monolayers at a ratio of around 1:10 (tumor cell:endothelial cell). Cells were then co-cultured for different times (1, 5, or 7 h) before live observation or fixation and immunostaining. Diapedesis was quantified blinded to the investigators using a Leica IRE2-DM epifluorescence microscope (Leica Microsystems Inc., Toronto, ON, Canada) as previously described [11]. Briefly, co-cultures were fixed and stained for F-actin. Diapedesis was quantified by counting the number of tumor cells in contact with the endothelium classified into three stages according to their position relative to the endothelium: round on top (tumor cells with a spherical shape located on the apical surface of the endothelium); migrating (tumor cells have penetrated through the endothelial monolayer at endothelial cell junctions where part of the cell body is located above the endothelium and part of it has spread on the matrigel, beneath the endothelial cell f-actin stress fibers); and underneath (all of the tumor cell body is below the plane of the endothelial cell stress fibers). Migrating and underneath cells were scored together as transmigrating. For each coverslip, approximately 100 cells adherent to the endothelium were scored and counted and all experiments were repeated three times with triplicate coverslips.
Pre-loading dye coupling assay
GJIC was evaluated using the preloading dye coupling assay as described previously [36] with the following modifications. Tumor cells were labeled for 15 minutes at 37°C with 2 ml of Opti-MEM (Invitrogen) containing 10 μg/ml calcein-AM and 10 μg/ml DiI. Calcein-AM is a fluorescent substrate that is cleaved in viable cells into a membrane impermeable form able to pass through functional gap junctions but not through other plasma membrane channels. Labeled tumor cells were washed twice with Hank's balanced salt solution (HBSS), harvested non-enzymatically using 2 mM EDTA, and added to HMVEC monolayers on matrigel. Dye transfer from tumor cells to endothelial cells was observed live by epifluorescence microscopy after 30 minutes of co-culture and the number of adherent tumor cells that transferred dye to adjacent endothelial cells was scored and expressed as percent of total number of tumor cells counted. For each coverslip, approximately 70 cells adherent to the endothelium were scored. Unless specified, all experiments were repeated three times with triplicate coverslips.
GJIC as assessed by fluorescence recovery after photobleaching
To determine whether carbenoxolone (CBX) blocks gap junctional coupling, HMVEC monolayers were pretreated for 7 h with 150 μM CBX or the inactive analog glycyrrhizic acid (GZA). HMVEC were then loaded for 15 minutes with 10 μg/ml calcein-AM (Molecular Probes) in an Opti-MEM solution containing CBX or GZA. Cells were rinsed twice with Opti-MEM containing CBX or GZA and immersed in fresh media containing CBX or GZA. Because the fluorescent dye can only pass through gap junctions, recovery of fluorescence in the bleached cell will only occur if dye passes through gap junctions with adjoining unbleached cells. For each experimental condition, three individual cells were bleached for approximately 30 s using the Zeiss laser scanning confocal microscope (LSM) 410 and an argon laser at full power. To observe recovery of fluorescence, images were obtained every five minutes following photobleaching using the argon laser at 33% power until maximum recovery was reached (15 minutes). In control experiments, monolayers were left untreated or were treated with the inactive analog of CBX, GZA. Mean fluorescence intensity of bleached/unbleached cells in the same area was measured over time using scion imaging software and expressed as mean relative recovery (arbitrary units). Data represent three independent experiments.
Immunofluorescence staining and laser scanning confocal microscopy
Cells were fixed at room temperature for 20 minutes with 2% or 4% (w/v) paraformaldehyde (EMD Chemicals Inc.) and washed in cation free phosphate buffered saline (PBS). Cells were permeabilized for five minutes at 4°C using a buffer containing 15 mM Tris-HCL, 120 mM sodium chloride (NaCl), 25 mM potassium chloride (KCl), 2 mM EDTA, 2 mM EGTA, 0.5% (v/v) Triton X-100, pH 7.5. To label F-actin, co-cultures were incubated for 30 minutes at room temperature with Alexa Fluor® 488 or Texas Red conjugated-phalloidin (Molecular Probes) diluted 1:50 in PBS with 1% (w/v) bovine serum albumin (BSA) followed by three washes in PBS. To examine the localization of F-actin and Cx43, cells were labeled with Texas red conjugated-phalloidin (1:50 dilution) and with monoclonal anti-Cx43 (clone P4G9 from the Fred Hutchinson Cancer Research Center Antibody Development Group, Seattle, WA, USA) diluted 1:10 in 1% (w/v) BSA/PBS. Cells were incubated with Alexa Fluor® 488-conjugated goat-anti-mouse secondary antibody (Molecular Probes). Nuclei were stained using 10 μg/ml Hoechst 33342 (Sigma-Aldrich) diluted in PBS. Coverslips were mounted on glass slides using vectashield mounting media (Vector Laboratories, Burlington, ON, Canada) and sealed with nail enamel. The inclusion of spacers made from plastic coverslips cut into strips prevented contact of co-cultures with the slides as described by Voura et al. [37].
Co-cultures were imaged using a Zeiss LSM 410 laser scanning confocal microscope system (Carl Zeiss Canada Ltd, Toronto, ON, Canada). Serial 0.7 μm optical sections of selected areas were recorded and evaluated with the LSM 410 software (Carl Zeiss Canada Ltd). Representative serial optical sections are presented in an apical to basal direction.
Adhesion assay
Adhesion to the endothelium was evaluated as previously described for leukocyte adhesion to the endothelium [38] with the following modifications. HMVEC monolayers were established in 1:20 matrigel coated wells (2.1 × 104 cells per well) and cultured for 48 h. HBL100, HBL100v, or HBL100Cx43 cells (2.1 × 103), pre-labeled with DiI, were added to HMVEC monolayers and co-cultured for 1 h. Co-cultures were fixed with 2% (w/v) paraformaldehyde and washed twice with PBS to remove non-adherent cells. The PBS was replaced with water and fluorescence intensity of DiI was measured for each well at 549 nm excitation and 565 nm emission wavelengths using a SAFIRE fluorescent microplate reader (Tecan US Inc., Durham, NC, USA). The number of adherent tumor cells was calculated and normalized based on a standard curve generated with known numbers of DiI-labeled tumor cells present in duplicate wells in the same microplate. Experiments were repeated three times with eight sample replicates per cell type.
Protein extraction and western blot analysis
HBL100, HBL100v, HBL100Cx43, HBL100GFP-Cx43, and HMVEC cells grown in culture dishes were lysed using the following buffer for protein collection: 10 mM Tris (pH 7.4), 150 mM NaCl, 1 mM EDTA, 0.1% (w/v) SDS, and 0.5% (v/v) Triton X-100. Protein concentration was determined with bicinchoninic acid (BCA) protein assay reagent (Pierce Chemical Co., Rockford, IL, USA). Protein (50 μg) for each cell type was separated in 12% SDS-polyacrylamide gels and transferred to nitrocellulose membranes. Nitrocellulose membranes were blocked with 5% non-fat dry milk in tris buffered saline (TBS) containing 5% Tween 20. Membranes were then probed for Cx43 protein, using a monoclonal anti-Cx43 antibody against amino-terminal amino acid residues 1–20 (clone P1E11 from the Fred Hutchinson Cancer Research Center Antibody Development Group, Seattle, WA, USA) at a dilution of 1:500. Immunoblots were probed with appropriate secondary antibodies conjugated to horse radish peroxidase (Pierce Chemical Co.) at a dilution of 1:10,000. The protein was visualized using Supersignal West Pico Chemiluminescent substrate (Pierce Chemical Co.) by exposing the immunoblot to Amersham Hyperfilm (Amersham Biosciences, Piscataway, NJ, USA) for 30 s.
Statistical analysis
Comparisons between cell types and different time points or treatments were evaluated using a two-way analysis of variance followed by Bonferroni post-hoc test. A p-value of < 0.05 was considered to be significant. Comparisons between cell types at a single time point were evaluated using one-way analysis of variance followed by Tukey's multiple comparison post-hoc test. Data are presented as means +/- SEM.
Results
Morphological evaluation of tumor cell diapedesis
Previous work suggested that GJIC between tumor cells and endothelial cells may affect diapedesis during metastasis [15-17]. To analyze and evaluate the efficiency of breast tumor cell diapedesis we used an in vitro assay in which tumor cells were allowed to migrate through a monolayer of endothelial cells grown on matrigel as we have previously described [9,39-41]. This assay mimics aspects of the vessel wall and has been used by us [35,42] and others [39-41] to analyze leukocyte diapedesis. Diapedesis was assessed by evaluating the spatial location of tumor cells as they migrated through an endothelial monolayer. Tumor cells pre-labeled with DiI were seeded for 1, 5 or 7 h onto endothelial monolayers before fixation and stained for F-actin. Tumor cells were scored according to their morphology and location with respect to the endothelium. We defined three distinct stages of migration observable by confocal microscopy (Fig. 1): top; migrating; and underneath. Tumor cells located on top of the endothelium prior to diapedesis tended to have a round or oval shape with small filopodia extending from the periphery of the cell (Fig. 1b, arrows). This shape was maintained from the apical (Fig. 1b) to the basal surface (Fig. 1d) of the tumor cell, differing from melanoma cells, which tend to have a fibroblastic shape and spread on the apical endothelial surface [37]. At sites of tumor cell-endothelial cell interactions, prominent microfilament bundles within the underlying endothelial cell were seen, indicating no disruption of the endothelium (Fig. 1d). Tumor cells in the process of diapedesis (migrating) had a unique morphology with distinct shapes at various focal levels (Fig. 1f–h) not previously seen in melanoma cells [37] but very similar to transmigrating monocytes and prostate tumor cells [11,35]. The apical aspect of migrating tumor cells located above the endothelium was round in shape and lacking filopodial extensions, but contained smooth cortical actin staining (Fig. 1f). At the level of the endothelium, tumor cells maintained close contact with the endothelial cells, which formed a circular opening surrounded by thick bundles of F-actin (Fig. 1g, arrow). The basal aspect of the tumor cells was located underneath the endothelium and spread along the matrigel by extending processes in various directions (Fig. 1h, arrows). Tumor cells located completely underneath the endothelium (Fig. 1j–l) had endothelial microfilament bundles extended over the top of the tumor cell as previously seen in melanoma cells [37]. No opening above the tumor cell remained, indicating complete closure of the transmigration passage (Fig. 1j). Tumor cells underneath the endothelium spreading on the surface of the matrigel had an irregular shape and stress fibers (Fig. 1l, arrows). The morphological features of HBL100v or HBL100Cx43 cells (not shown) during diapedesis were not noticeably different from those of HBL100 wild-type cells. We used the morphological criteria shown in Fig. 1 to quantify various stages and assess the efficiency of diapedesis.
Expression and distribution of exogenous Cx43 in HBL100 cells and HMVEC monolayers
To explore if Cx43 expression in HBL100 cells affects any aspects of diapedesis, cells were engineered to express functional Cx43, non-functional GFP-Cx43 or the empty retroviral vector as a control. Cx43 was absent in both wild-type and empty vector HBL100v control cell lines used (Fig. 2a,c,d) but expressed in HMVEC (Fig. 2a,b) and cells engineered to express Cx43 (Fig. 2a,e) or GFP-Cx43 (Fig. 2a,f). Importantly, both Cx43 and GFP-Cx43 were routinely observed to localize to the cell surface, consistent with the formation of gap junction plaques (Fig. 2b,e,f). Intracellular populations of Cx43 likely reflecting different stages of assembly or degradation in HMVEC, HBL100Cx43 and HBLGFP-Cx43 were also seen. These results suggest that exogenous or endogenously expressed Cx43 assembles into gap junctional plaques at sites of homocellular endothelial or tumor cell-cell contacts.
HBL100vCx43 cells form functional gap junctional channels with HMVEC monolayers
To determine whether HBL100 cells engineered to exogenously express Cx43 formed gap junctions and established GJIC with HMVEC cells, we co-cultured HBL100Cx43 with HMVEC monolayers, localized Cx43 in these cultures (Fig. 3a–c), and assayed for dye transfer between tumor and endothelial cells (Fig. 3d–j). Confocal microscopy revealed Cx43 at tumor cell-endothelial cell interfaces (Fig. 3b, arrows) where tumor cells were wedged between endothelial cells and where cortical F-actin was found (Fig. 3b,c). These results indicate that during diapedesis HBL100Cx43 cells form heterocellular gap junctions with adjacent endothelial cells.
Preloading dye transfer studies to assess whether Cx43 is being assembled into functional gap junctions at tumor/endothelial cell interfaces revealed that dye spread extensively from Cx43 expressing HBL100 cells to endothelial cells (Fig. 3i), whereas dye failed to spread from control HBL100v cells to endothelial cells (Fig. 3f). These results suggest that exogenous Cx43 in HBL100Cx43 can form functional gap junctions with HMVEC early on during diapedesis. Live quantification of coupling performed after 30 minutes of co-culture revealed that ten times more Cx43 expressing HBL100 cells coupled to the endothelium than those lacking Cx43 (Fig. 3j). The low occurrence of dye transfer between HBL100v control cells and HMVEC is likely due to nonspecific dye uptake from damaged tumor cells rather than functional gap junctional coupling due to connexin expression.
Cx43 enhances HBL100 diapedesis
To compare the efficiency of diapedesis of Cx43 deficient tumor cells with those that exogenously express Cx43, we co-cultured HBL100, HBL100v or HBL100Cx43 cells with HMVEC for 1, 5 and 7 h. At all time points HBL100Cx43 cells showed an up to two-fold increase in diapedesis efficiency compared to cells lacking Cx43 (Fig. 4a). This increase in migration appeared to plateau at 5 h. Conversely, at all time points no statistical difference in migration efficiency was observed between wild-type (HBL100) and control (HBL100v) cells (Fig. 4a). The relative proportion of cells at different stages of migration was evaluated after 7 h of co-culture (Fig. 4b). About 50% of wild-type and vector control cells encountered were located on top of the endothelium, 40% were migrating across the endothelium, and about 10% were completely located underneath. In contrast, only 20% of HBL100Cx43 cells were located on top of the endothelium, while over 60% migrated across. No significant difference in the relative proportion of cells underneath the endothelium was observed between HBL100 cells with or without Cx43 expression (Fig. 4b). These results suggested that Cx43 facilitated tumor cell diapedesis but it was unclear if this was due to increased adhesiveness or whether GJIC was required.
To determine whether the increase in diapedesis efficiency was due to increased adhesive properties of Cx43 expressing tumor cells, we co-cultured HBL100, HBL100v and HBL100Cx43 cells on endothelial cells and compared the number of DiI labeled tumor cells adherent to HMVEC monolayers using fluorescence intensity measurements. No significant difference in the ability of tumor cells to adhere to the endothelium was observed between HBL100 breast tumor cells expressing Cx43 and tumor cells lacking Cx43 expression (Fig. 5), suggesting that the increased efficiency of diapedesis was not due to increased cellular adhesion.
To determine whether GJIC was required to increase HBL100 tumor cell diapedesis, we chemically blocked gap junctional coupling using CBX. HMVEC monolayers were preloaded with calcein dye and fluorescence recovery after photobleaching (FRAP) analysis was first used to determine whether CBX effectively blocked gap junctions in endothelial cells. Individual cells within the calcein-AM loaded monolayer were bleached and recovery of fluorescence was quantified. Our results demonstrated that, in the presence of 150 μM CBX, fluorescence recovery in the bleached cell was prevented (Fig. 6a, triangles), indicating lack of homocellular gap junctional coupling. In GZA treated monolayers (negative control), however, recovery of fluorescence occurred at levels similar to those in untreated cells (Fig. 6a, rectangles and circles, respectively), indicating functional homocellular gap junctional coupling between adjacent endothelial cells.
To determine whether CBX would impair the Cx43-induced increase in diapedesis of HBL100Cx43 cells, HMVEC monolayers were pretreated with 150 μM CBX, or 150 μM GZA prior to adding tumor cells (Fig. 6b). In the presence of GZA HBL100Cx43 cells were more efficient at diapedesis compared to HBL100v control cells. A significant decrease in diapedesis efficiency in the presence of the gap junctional blocker CBX, however, was observed for both HBL100v and HBL100Cx43 cells. This finding suggests that gap junctional coupling between endothelial cells partially regulates diapedesis efficiency of tumor cells regardless of their Cx43 expression.
To determine if GJIC was necessary for Cx43-linked increase in diapedesis and, more specifically, if heterocellular GJIC was necessary between tumor and endothelial cells, we compared diapedesis of HBL100 cells that expressed either the functional or non-functional Cx43 (GFP-Cx43). Previous studies showed that GFP fused to the amino terminus of Cx43 (GFP-Cx43) results in a chimeric connexin that can traffic and form gap junction-like structures at cell-cell contacts, but these structures are not functional [33] and also functional hemichannels do not form [43]. HBL100, HBL100v, HBL100Cx43, or HBL100GFP-Cx43 cells were co-cultured with HMVEC for 7 h and scored according to the described criteria for stages of migration (Fig. 1). After 7 h of co-culture the percentage of migrating HBL100GFP-Cx43 cells was similar to that of migrating HBL100 and HBL100v cells, whereas diapedesis efficiency of HBL100Cx43 cells was more than 50% higher (Fig. 7). This result suggests that the presence of Cx43 protein in the tumor cell alone is not responsible for the observed increase in diapedesis, but that heterocellular GJIC between tumor cells and endothelial cells is required to augment the efficiency of diapedesis of HBL100 cells.
Discussion
Previous work by Kramer and Nicolson [44] evaluated melanoma cell diapedesis through endothelial monolayers by time-lapse, phase-contrast, scanning and transmission electron microscopy. In our study we used an in vitro cell culture model to evaluate the role of Cx43 expression in diapedesis of breast tumor cells through the endothelium by confocal microscopy. Our in vitro cell culture model enables us to dissect various stages of diapedesis and to examine the molecules or mechanisms involved in each of these stages [9,35,37,39-41]. Detailed morphological analysis by confocal microscopy allowed us to classify HBL100 breast tumor cell diapedesis into three stages (Fig. 1). First HBL100 breast cancer cells adhered to the endothelium but maintained a round morphology while often extending finger-like processes. This is similar to the behavior of PC3 prostate tumor cells [11] but differs from our previous findings with WM239 melanoma cells, which tended to spread along the endothelial cell surface before diapedesis and extended clusters of blebs at the heterotypic contact site between tumor cells and endothelial cells [9]. It is, therefore, likely that the adhesive interactions between tumor cells and the mechanisms of communication between endothelial cells and tumor cells are cell type specific and may regulate important aspects of diapedesis. During diapedesis, HBL100 cells, like melanoma cells, sent out projections through the endothelium, resulting in a localized retraction of the endothelial cells [9]. Compared to melanoma cells, the width of the opening through which HBL100 cells migrate appeared to be smaller and is lined by F-actin. HBL100 cells that came into contact with the underlying basement membrane spread and flattened out, forming stress fibers on the extracellular matrix, while the endothelium reformed cell-cell junctions over the top of the tumor cell. This stage of diapedesis is very similar in WM239 melanoma cells and PC3 prostate tumor cells [9,11]. Exogenous expression of Cx43 in HBL100 cells did not appear to have an effect on their adhesive properties to endothelial surfaces nor their morphology at any of the stages of diapedesis.
Interestingly, glioma cells engineered to express Cx43 formed bigger homocellular aggregates compared to control cells [45], suggesting that increased aggregation may be due to increased adhesive properties of the cells as a result of Cx43 expression.
In our study we demonstrated that HBL100 breast cancer cells engineered to express Cx43 were able to localize Cx43 protein to cell-cell contacts during diapedesis, forming communication competent gap junctions with endothelial cells (Fig. 3b,g–j). Previous studies by El-Sabban et al. [15] using melanoma cells expressing Cx43 support these findings, reporting that highly metastatic melanoma cells with higher levels of Cx43 mRNA coupled more efficiently to endothelial cells than low metastatic tumor cells with low amounts of Cx43 mRNA [15]. Although these and other reports demonstrated gap junctional coupling between tumor cells and endothelial cells [15-17] it was unclear if GJIC was required for increased tumor cell diapedesis. Whereas previous reports have shown Cx43 localized at contact sites between tumor cells and endothelial cells [16], our study demonstrated for the first time that Cx43 protein can be localized at the cell-cell interface between endothelial cells and tumor cells during the process of diapedesis. This localization of Cx43 to tumor/endothelial cell boundaries and the establishment of functional heterocellular coupling indicated that GJIC facilitates diapedesis. This finding is surprising because previous work has demonstrated a downregulation of connexin expression and gap junctional communication in primary tumor [29,46-49]. In addition, other studies have demonstrated reintroduction of connexins in primary tumor cells acts as a tumor suppressor in several cancers [34,50-52]. It is conceivable that tumor cells with low levels of connexins may encounter factors within the circulation such as growth factors or cytokines that could restore connexin expression previously lost in the primary tumor. For example it has been shown that vascular endothelial growth factor (VEGF), IL-1β and IL-6 are upregulated in serum of cancer patients [53,54] and that VEGF and IL-1β can increase Cx43 expression [55,56]. Increasing connexin expression in poorly invasive tumor cells, therefore, may increase their ability to migrate through the endothelium, allowing them to invade other tissues where they may develop into more aggressive tumors.
Our study shows that HBL100 breast tumor cells expressing Cx43 migrated more efficiently through the endothelium than their counterparts lacking Cx43 expression. This finding suggests that gap junctions and/or Cx43 expression in these cells may affect the motile behavior of cells by changing the kinetics of diapedesis of individual cells. The fact that Cx43 expression can influence the motile behavior of cells has been previously demonstrated [57]. For example, HeLa cells transfected with Cx43 formed functional gap junctions with embryonic chicken heart cells, and were more invasive into the ventricle of the heart compared to their wild-type counterparts, supporting the hypothesis that Cx43 expressing cells can be more invasive [57]. In addition, Cx26 expressing BL6 mouse melanoma cells had increased metastatic properties in vivo [17].
Cellular motility may also be affected by changes in cadherin expression. Down regulation of E-cadherin expression or function correlates with tumor development and malignancy [58,59], while N-cadherin expression appears to enhance cell motility [60,61]. Interestingly, both Cx43 and N-cadherin appear to regulate mouse neural crest cell motility, perhaps by engaging p120ctn signaling [62]. At present, a functional relationship between cadherin expression and Cx43 expression in breast tumor cell motility is largely unknown and the role of ZO-1 or p120ctn, which interact with both cadherins and connexins, in this process needs to be examined.
It was important to distinguish whether the observed increase in diapedesis efficiency was due to enhanced expression of Cx43 in the tumor cell or due to the exchange of signals between tumor and endothelial cells via functional gap junctions. Two approaches were used to examine this issue. First, we blocked homocellular and heterocellular GJIC. In the presence of the gap junctional blocker, diapedesis efficiencies were significantly reduced in both the control HBL100v cells and the Cx43 expressing HBL100 cells (HBL100Cx43) (Fig. 6b), suggesting gap junctional coupling between the adjacent endothelial cells is important in regulating tumor cell diapedesis efficiency regardless of their connexin content. Disruption of homocellular endothelial GJIC may compromise the normal function of the endothelium required for blood borne cells to cross this barrier. Our findings therefore support the hypothesis that the endothelium plays an active role in the process of diapedesis in line with previous work demonstrating that during diapedesis endothelial cells require N-cadherin to reform the junctions above melanoma cells [9]. In the second approach, heterocellular coupling between tumor cells and endothelial cells was specifically disrupted in HBL100 breast tumor cells by expressing a non-functional Cx43 in which GFP was tagged to the amino terminus of the Cx43 polypeptide. This chimeric connexin is capable of forming gap junction-like structures at the cell surface but incapable of forming functional gap junction channels or functional hemichannels as revealed by dye [33] and electrical [43] coupling assays. Our findings reveal that the diapedesis efficiency of tumor cells expressing the non-functional Cx43 was similar to that of wild-type and empty vector control cells (Fig. 7), suggesting that heterocellular GJIC between tumor and endothelial cells is necessary to enhance diapedesis. It is possible that fusing GFP to the amino terminus of Cx43 may be blocking Cx43 binding proteins that could be required for efficient tumor cell diapedesis, although, to date, no amino-terminal Cx43 binding proteins have been identified. Together with our finding that CBX decreases the efficiency of tumor cell diapedesis, our results support the hypothesis that GJIC facilitates tumor cell diapedesis.
Several possible scenarios could explain how GJIC might modulate tumor cell diapedesis. Tumor cells might send signals to the endothelium, resulting in the release of endothelial proteolytic enzymes, facilitating endothelial cell retraction. A recent study by Bazarbachi et al. [63] demonstrated that neoplastic lymphocytes triggered increased matrix metalloproteinase activity in endothelial cells. This activity was reduced in the presence of the gap junctional blocker 18α glycyrrhetinic acid [63]. Tumor cells may also send signals, such as inositol triphosphate (IP3), through gap junctions that directly trigger the weakening of inter-endothelial junctions. IP3 can pass between adjacent cells through gap junctions [64] and regulate intracellular calcium levels [65]. Interestingly, the depletion of calcium stores in the endothelium inhibited extravasation of MCF-7 tumor cells [66] and elevated levels of intracellular calcium resulted in the transient phosphorylation of myosin light chain [67,68], which is known to increase endothelial permeability [69]. Endothelial cells may also send signals through gap junctions to tumor cells, resulting in the upregulation of proteins involved in cellular motility (such as Rho GTPases, protein kinase C, and phosphoinositide 3 kinase).
Although it has been suggested that extravasation may not be a limiting factor for metastasis [70-72], metastatic ability appears to be correlated with the capability of tumor cells to communicate with endothelial cells as previously demonstrated in metastatic B16-F10 melanoma cells [15-17]. Transient re-expression of connexins in tumor cells, and the exchange of molecules between tumor cells and endothelial cells through gap junctions, might also affect tumor cells in their behavior following extravasation when these cells interact with the extracellular matrix and form secondary tumors. For example, melanoma cells invading into the dermis (i.e. towards the endothelium) have an increase in Cx26 expression compared to those located in the basal layer of the epidermis [17]. These invading melanoma cells showed stronger heterocellular coupling to the endothelium than homocellular coupling with each other (in vitro). Furthermore, squamous cell carcinomas exhibited reduced levels of Cx26 and Cx43 expression at early stages of mouse skin carcinogenesis and sites of invasion, although Cx26 expression was partially restored in these cells at metastatic sites within the lymph nodes [73].
Conclusion
Our work supports the hypothesis that heterocellular gap junctional coupling between tumor cells and endothelium may regulate the extravasation of at least some poorly invasive tumor cells such as HBL100 breast tumor cells, and that homocellular GJIC amongst endothelial cells may play an active role in this process. These studies would suggest that even though extravasation may not necessarily be a rate limiting step in metastasis, there might be therapeutic advantages to designing drugs or reagents that inhibit vascular homocellular and heterocellular GJIC as a means of reducing metastatic spread of tumor cells.
Abbreviations
BSA = bovine serum albumin; CBX = carbenoxolone; Cx = connexin; DiI = dioctadecyl-3, 3,3', 3'-tetramethylindocarbocyanine percholate; D-MEM = Dulbecco's modified Eagles medium; EGM = endothelial growth media; FBS = fetal bovine serum; FRAP = fluorescence recovery after photobleaching; GFP = green fluorescent protein; GJIC = gap junctional intercellular communication; GZA = glycyrrhizic acid; HBL100Cx43 = HBL100 cells expressing Cx43; HBSS = Hanks balances salt solution; HMVEC = human microvascular endothelial cell; IL = interleukin; IP3 = inositol triphosphate; LSCM = laser scanning confocal microscopy; PBS = phosphate buffered saline; VE = vascular endothelial; VEGF = vascular endothelial growth factor.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MAP participated in the design of the study, carried out the imaging, immunocytochemistry, immunoblotting, and migration assays in this study, and drafted the manuscript. QS carried out the molecular biology of the study. DWL participated in the design of the study. MS conceived the study, and participated in its design and coordination and revised the manuscript. All authors read and approved the final manuscript.
Acknowledgements
This work was supported by grants from the Cancer Research Society to MS and DWL, Canadian Breast Cancer Research Alliance to DWL, the Canadian Institutes of Health Research to MS and graduate scholarships from the Ontario government to MAP.
Figures and Tables
Figure 1 Tumor cell diapedesis through human microvascular endothelial cell (HMVEC) monolayers. HBL100 cells were pre-labeled with dioctadecyl-3, 3,3', 3'-tetramethylindocarbocyanine percholate (DiI; red), co-cultured with HMVEC monolayers for various time periods, fixed with paraformaldehyde and labeled for F-actin (green). Optical sections obtained by laser scanning confocal microscopy at the focal levels indicated in (a,e,i) identified three major stages of diapedesis: round on top of (a-d), migrating through (e-h), or located underneath (i-l) the endothelium. (b) Cells round on top had filopodial extensions present on the apical surface of a cell (arrows). (g) Thick bundles of F-actin were present at the interface between endothelial cells and the migrating tumor cell (arrow). (h) Migrating tumor cells with portions spread underneath the endothelium (arrows) and (i) those that completed diapedesis often contained prominent stress fibers underneath the endothelium (arrows). No major morphological differences between wild-type and connexin43 (Cx43) expressing HBL100 cells were observed. Bar = 10 μm.
Figure 2 Exogenous connexin43 (Cx43) in HBL100 cells forms gap junctional plaques. To explore if Cx43 expression and/or gap junctional intercellular communication (GJIC) affects diapedesis of breast tumor cells we used GJIC-deficient HBL100 cells to engineer cells expressing functional Cx43 (HBL100Cx43) and non-functional Cx43 (HBL100GFP-Cx43). HBL100v, containing the empty expression vector, served as a control. (a) Cx43 expression of these cells was compared by western blot analysis to that of human microvascular endothelial cell (HMVEC) grown on matrigel used as a control for endogenous Cx43 expression. Whereas Cx43 was undetectable in HBL100 and HBL100v cells, HMVEC grown on matrigel, HBL100Cx43, and HBL100GFP-Cx43 cells expressed Cx43. Immunocytochemistry localized Cx43 at cell-cell borders in (b) HMVEC monolayers (arrows), (e) between HBL100Cx43 cells (arrow) or (f) HBL100GFP-Cx43 cells (arrows), whereas (c) HBL100 and (d) HBL100v cells lacked Cx43 staining. Bar = 10 μm.
Figure 3 Connexin43 (Cx43) localizes to tumor/endothelial cell interfaces and assembles into functional gap junction channels. To examine Cx43 distribution during tumor cell diapedesis, cells expressing functional Cx43 (HBL100Cx43) were co-cultured with human microvascular endothelial cell (HMVEC) monolayers, fixed and labeled for (b) Cx43 (green) and (c) F-actin (red). The optical section obtained by confocal microscopy revealed (b) Cx43 (arrows) enriched at tumor cell (T)/endothelial cell (e) contact site, where it partially co-localizes with cortical F-actin. (a) Schematic diagram indicating the level of the optical section where the tumor cell is wedged between the endothelial cells and (b) the location of the Cx43-containing plaques (arrows). To evaluate if expression of exogenous Cx43 in HBL100 cells resulted in formation of functional gap junctions with HMVEC, (f) HBL100v or (i) HBL100Cx43 cells (arrows) were preloaded with (e,h) dioctadecyl-3, 3,3', 3'-tetramethylindocarbocyanine percholate (DiI; red) and (f,i) calcein-AM (green) and co-cultured with HMVEC monolayers for 30 minutes. Live observation of the co-cultures revealed that in contrast to (d-f) HBL100v cells, (g-i) HBL100Cx43 cells allowed the calcein dye to spread to adjacent HMVEC cells. (j) Gap junctional intercellular communication (GJIC) between tumor cells and endothelial cells was quantified live after 30 minutes of co-culture, revealing a 10-fold increase in the number of connexin expressing tumor cells coupled to the endothelium compared to tumor cells lacking Cx43 (*p < 0.05). Bar = (c) 10 μm and (f,i) 50 μm.
Figure 4 Connexin43 (Cx43) enhances tumor cell diapedesis. HBL100 and HBL100v cells or cells expressing functional Cx43 (HBL100Cx43) were co-cultured with HMVEC monolayers for 1, 5 and 7 h. (a) Adherent cells in the process of diapedesis and cells that had completed transmigration were scored together as migrating. Compared to HBL100v or HBL100 wild-type cells a two-fold increase in diapedesis of HBL100Cx43 cells was seen at 1, 5 and 7 h. (b) At 7 h the relative proportion of cells at different stages of migration was evaluated. Compared to control cells, the observed two-fold increase in the number of transmigrating HBL100Cx43 cells correlated with a two-fold decrease in the number of HBL100Cx43 cells located on top of the monolayer. No significant difference in the proportion of cells underneath the endothelium was observed between the Cx43 expressing and non-expressing HBL100 cells. Data are expressed as the mean percentage ± SEM of three independent experiments performed in triplicate (*p < 0.05).
Figure 5 Connexin43 (Cx43) expression does not affect tumor cell adhesion to the endothelium. HBL100, HBL100v and Cx43 expressing HBL100 cells (HBL100Cx43) were labeled with dioctadecyl-3, 3,3', 3'-tetramethylindocarbocyanine percholate (DiI), added for 1 h to endothelial monolayers and co-cultures were washed to remove non-adherent cells. Fluorescence intensity measurements revealed no significant difference in the number of tumor cells attached to the endothelium regardless of cell type. Data are expressed as mean percentage ± SEM of three independent experiments with eight replicates.
Figure 6 Carbenoxolone (CBX) blocks gap junctional intercellular communication (GJIC) in human microvascular endothelial cell (HMVEC) monolayers and decreases tumor cell diapedesis. (a) To determine the extent by which GJIC in HMVEC was blocked by CBX, monolayers were pre-treated with 150 μM CBX or the inactive analog glycyrrhizic acid (GZA) and analyzed by fluorescence recovery after photo-bleaching (FRAP). FRAP analysis revealed that fluorescence recovery by homocellular gap junctional coupling was blocked by 150 μM CBX (triangles), while recovery of fluorescence was observed in untreated (circles) or GZA treated monolayers (rectangles). Data represent three independent experiments. (b) To determine the effect of blocking gap junctional coupling on diapedesis, HMVEC monolayers were pre-treated with 150 μM CBX, or the inactive analog GZA and HBL100v cells or cells expressing functional Cx43 (HBL100Cx43) were co-cultured for 7 h with HMVEC monolayers, fixed and labeled for F-actin. Tumor cells were scored according to the criteria specified in Fig 1. In comparison to the inactive analog (GZA), treatment with CBX resulted in a significant decrease in the number of migrating HBL100v and HBL100Cx43 cells. Data are expressed as the mean percentage of three independent experiments ± SEM performed in duplicate or triplicate. Identical letters over bars indicate no statistical significance while different letters signify statistical significance (p < 0.01).
Figure 7 Non-functional Connexin43 (Cx43) does not enhance tumor cell diapedesis. HBL100 cells were transfected with Cx43 to which green fluorescent protein (GFP) was fused at the amino terminus (GFP-Cx43) to make it non-functional. Diapedesis of HBL100v cells and cells expressing functional (HBL100Cx43) or non-functional (HBL100GFP-Cx43) Cx43 was scored following co-culture with human microvascular endothelial cell (HMVEC) monolayers for 7 h. Analysis revealed that HBL100 cells transfected with GFP-Cx43 had an efficiency of diapedesis similar to HBL100v control cells in contrast to HBL100Cx43 cells, which showed significantly greater numbers of migrating cells. Data are expressed as the mean percentage ± SEM of five independent experiments in duplicate or triplicate (* p < 0.05).
==== Refs
Ahmad A Hart IR Mechanisms of metastasis Crit Rev Oncol Hematol 1997 26 163 173 9481522
Meyer T Hart IR Mechanisms of tumour metastasis Eur J Cancer 1998 34 214 221 9741324 10.1016/S0959-8049(97)10129-0
Couzin J Medicine. Tracing the steps of metastasis, cancer's menacing ballet Science 2003 299 1002 1006 12586919 10.1126/science.299.5609.1002
Friedl P Wolf K Tumour cell invasion and migration: diversity and escape mechanisms Nat Rev Cancer 2003 3 362 374 12724734 10.1038/nrc1075
Umbas R Schalken JA Aalders TW Carter BS Karthaus HF Schaafsma HE Debruyne FM Isaacs WB Expression of the cellular adhesion molecule E-cadherin is reduced or absent in high-grade prostate cancer Cancer Res 1992 52 5104 5109 1516067
Bringuier PP Umbas R Schaafsma HE Karthaus HF Debruyne FM Schalken JA Decreased E-cadherin immunoreactivity correlates with poor survival in patients with bladder tumors Cancer Res 1993 53 3241 3245 8324734
Dorudi S Hanby AM Poulsom R Northover J Hart IR Level of expression of E-cadherin mRNA in colorectal cancer correlates with clinical outcome Br J Cancer 1995 71 614 616 7880746
Gui GP Wells CA Yeomans P Jordan SE Vinson GP Carpenter R Integrin expression in breast cancer cytology: a novel predictor of axillary metastasis Eur J Surg Oncol 1996 22 254 258 8654607 10.1016/S0748-7983(96)80013-8
Sandig M Voura EB Kalnins VI Siu CH Role of cadherins in transendothelial migration of melanoma cells in culture Cell Motil Cytoskeleton 1997 38 351 364 9415377 10.1002/(SICI)1097-0169(1997)38:4<351::AID-CM5>3.0.CO;2-6
Voura EB Ramjeesingh RA Montgomery AM Siu CH Involvement of Integrin alpha(v)beta(3) and cell adhesion molecule L1 in transendothelial migration of melanoma cells Mol Biol Cell 2001 12 2699 2710 11553709
Wang X Ferreira AM Shao Q Laird DW Sandig M Beta(3) integrins facilitate matrix interactions during transendothelial migration of PC3 prostate tumor cells The Prostate 2005 63 65 80 15468167 10.1002/pros.20168
Bartsch JE Staren ED Appert HE Adhesion and migration of extracellular matrix-stimulated breast cancer J Surg Res 2003 110 287 294 12697413 10.1016/S0022-4804(03)00004-0
Humphries MJ Olden K Yamada KM A synthetic peptide from fibronectin inhibits experimental metastasis of murine melanoma cells Science 1986 233 467 470 3726541
Ohnishi T Arita N Hiraga S Taki T Izumoto S Fukushima Y Hayakawa T Fibronectin-mediated cell migration promotes glioma cell invasion through chemokinetic activity Clin Exp Metastasis 1997 15 538 546 9247256 10.1023/A:1018422926361
El Sabban ME Pauli BU Cytoplasmic dye transfer between metastatic tumor cells and vascular endothelium J Cell Biol 1991 115 1375 1382 1955478 10.1083/jcb.115.5.1375
El Sabban ME Pauli BU Adhesion-mediated gap junctional communication between lung- metastatic cancer cells and endothelium Invasion Metastasis 1994 14 164 176 7657509
Ito A Katoh F Kataoka TR Okada M Tsubota N Asada H Yoshikawa K Maeda S Kitamura Y Yamasaki H Nojima H A role for heterologous gap junctions between melanoma and endothelial cells in metastasis J Clin Invest 2000 105 1189 1197 10791993
Bruzzone R White TW Paul DL Connections with connexins: the molecular basis of direct intercellular signaling Eur J Biochem 1996 238 1 27 8665925 10.1111/j.1432-1033.1996.0001q.x
White TW Paul DL Genetic diseases and gene knockouts reveal diverse connexin functions Annu Rev Physiol 1999 61 283 310 10099690 10.1146/annurev.physiol.61.1.283
Willecke K Eiberger J Degen J Eckardt D Romualdi A Guldenagel M Deutsch U Sohl G Structural and functional diversity of connexin genes in the mouse and human genome Biol Chem 2002 383 725 737 12108537 10.1515/BC.2002.076
Willecke K Hennemann H Dahl E Jungbluth S Heynkes R The diversity of connexin genes encoding gap junctional proteins Eur J Cell Biol 1991 56 1 7 1666038
Pepper MS Montesano R el Aoumari A Gros D Orci L Meda P Coupling and connexin 43 expression in microvascular and large vessel endothelial cells Am J Physiol 1992 262 C1246 C1257 1317103
Bruzzone R Haefliger JA Gimlich RL Paul DL Connexin40, a component of gap junctions in vascular endothelium, is restricted in its ability to interact with other connexins Mol Biol Cell 1993 4 7 20 8382974
Little TL Beyer EC Duling BR Connexin 43 and connexin 40 gap junctional proteins are present in arteriolar smooth muscle and endothelium in vivo Amer J Physiol-Heart Circ Phy 1995 37 H729 H739
Van Rijen H van Kempen MJ Analbers LJ Rook MB van Ginneken AC Gros D Jongsma HJ Gap junctions in human umbilical cord endothelial cells contain multiple connexins Am J Physiol 1997 272 C117 C130 9038818
Yeh HI Lai YJ Chang HM Ko YS Severs NJ Tsai CH Multiple connexin expression in regenerating arterial endothelial gap junctions Arterioscler Thromb Vasc Biol 2000 20 1753 1762 10894813
Monaghan P Clarke C Perusinghe NP Moss DW Chen XY Evans WH Gap junction distribution and connexin expression in human breast Exp Cell Res 1996 223 29 38 8635493 10.1006/excr.1996.0055
Jamieson S Going JJ D'Arcy R George WD Expression of gap junction proteins connexin 26 and connexin 43 in normal human breast and in breast tumours J Pathol 1998 184 37 43 9582525 10.1002/(SICI)1096-9896(199801)184:1<37::AID-PATH966>3.0.CO;2-D
Laird DW Fistouris P Batist G Alpert L Huynh HT Carystinos GD Alaoui-Jamali MA Deficiency of connexin43 gap junctions is an independent marker for breast tumors Cancer Res 1999 59 4104 4110 10463615
Saunders MM Seraj MJ Li Z Zhou Z Winter CR Welch DR Donahue HJ Breast cancer metastatic potential correlates with a breakdown in homospecific and heterospecific gap junctional intercellular communication Cancer Res 2001 61 1765 1767 11280719
Carystinos GD Bier A Batist G The role of connexin-mediated cell-cell communication in breast cancer metastasis J Mammary Gland Biol Neoplasia 2001 6 431 440 12013532 10.1023/A:1014787014851
Elfgang C Eckert R Lichtenberg-Frate H Butterweck A Traub O Klein RA Hulser DF Willecke K Specific permeability and selective formation of gap junction channels in connexin-transfected HeLa cells J Cell Biol 1995 129 805 817 7537274 10.1083/jcb.129.3.805
Laird DW Jordan K Thomas T Qin H Fistouris P Shao Q Comparative analysis and application of fluorescent protein-tagged connexins Microsc Res Tech 2001 52 263 272 11180619 10.1002/1097-0029(20010201)52:3<263::AID-JEMT1012>3.0.CO;2-Q
Qin H Shao Q Curtis H Galipeau J Belliveau DJ Wang T Alaoui-Jamali MA Laird DW Retroviral delivery of connexin genes to human breast tumor cells inhibits in vivo tumor growth by a mechanism that is independent of significant gap junctional intercellular communication J Biol Chem 2002 277 29132 29138 12042301 10.1074/jbc.M200797200
Sandig M Negrou E Rogers KA Changes in the distribution of LFA-1, catenins and F-actin during transendothelial migration of monocytes in culture J Cell Sci 1997 110 2807 2818 9427289
Goldberg GS Bechberger JF Naus CC A pre-loading method of evaluating gap junctional communication by fluorescent dye transfer Biotechniques 1995 18 490 497 7779401
Voura EB Sandig M Kalnins VI Siu CH Cell shape changes and cytoskeleton reorganization during transendothelial migration of melanoma cells Cell Tissue Res 1998 293 375 387 9716727 10.1007/s004410051129
Vaporciyan AA Jones ML Ward PA Rapid analysis of leukocyte-endothelial adhesion J Immunol Methods 1993 159 93 100 8445268 10.1016/0022-1759(93)90145-W
Worthylake RA Burridge K Leukocyte transendothelial migration: orchestrating the underlying molecular machinery Curr Opin Cell Biol 2001 13 569 577 11544025 10.1016/S0955-0674(00)00253-2
Carman CV Springer TA A transmigratory cup in leukocyte diapedesis both through individual vascular endothelial cells and between them J Cell Biol 2004 167 377 388 15504916 10.1083/jcb.200404129
Worthylake RA Lemoine S Watson JM Burridge K RhoA is required for monocyte tail retraction during transendothelial migration J Cell Biol 2001 154 147 160 11448997 10.1083/jcb.200103048
Ronald JA Ionescu CV Rogers KA Sandig M Differential regulation of transendothelial migration of THP-1 cells by ICAM-1/LFA-1 and VCAM-1/VLA-4 J Leukoc Biol 2001 70 601 609 11590197
Contreras JE Saez JC Bukauskas FF Bennett MV Gating and regulation of connexin 43 (Cx43) hemichannels Proc Natl Acad Sci U S A 2003 100 11388 11393 13130072 10.1073/pnas.1434298100
Kramer RH Nicolson GL Interactions of tumor cells with vascular endothelial cell monolayers: a model for metastatic invasion Proc Natl Acad Sci USA 1979 76 5704 5708 293673
Lin JH Takano T Cotrina ML Arcvino G Kang J Liu S Gao Q Jiang L Li F Lichtenberg-Frate H Connexin 43 enhances the adhesivity and mediates the invasion of malignant glioma cells J Neurosci 2002 22 4302 4311 12040035
Naus CC Bechberger JF Caveney S Wilson JX Expression of gap junction genes in astrocytes and C6 glioma cells Neurosci Lett 1991 126 33 36 1650934 10.1016/0304-3940(91)90364-Y
Cesen-Cummings K Fernstrom MJ Malkinson AM Ruch RJ Frequent reduction of gap junctional intercellular communication and connexin43 expression in human and mouse lung carcinoma cells Carcinogenesis 1998 19 61 67 9472694 10.1093/carcin/19.1.61
Yamasaki H Mesnil M Omori Y Mironov N Krutovskikh V Intercellular communication and carcinogenesis Mutat Res 1995 333 181 188 8538626
Saunders MM Seraj MJ Li Z Zhou Z Winter CR Welch DR Donahue HJ Breast cancer metastatic potential correlates with a breakdown in homospecific and heterospecific gap junctional intercellular communication Cancer Res 2001 61 1765 1767 11280719
Rose B Mehta PP Loewenstein WR Gap-junction protein gene suppresses tumorigenicity Carcinogenesis 1993 14 1073 1075 8389252
Mehta PP Perez-Stable C Nadji M Mian M Asotra K Roos BA Suppression of human prostate cancer cell growth by forced expression of connexin genes Dev Genet 1999 24 91 110 10079514 10.1002/(SICI)1520-6408(1999)24:1/2<91::AID-DVG10>3.0.CO;2-#
Tanaka M Grossman HB Connexin 26 induces growth suppression, apoptosis and increased efficacy of doxorubicin in prostate cancer cells Oncol Rep 2004 11 537 541 14719096
Coskun U Gunel N Toruner FB Sancak B Onuk E Bayram O Cengiz O Yilmaz E Elbeg S Ozkan S Serum leptin, prolactin and vascular endothelial growth factor (VEGF) levels in patients with breast cancer Neoplasma 2003 50 41 46 12687277
Dulger H Alici S Sekeroglu MR Erkog R Ozbek H Noyan T Yavuz M Serum levels of leptin and proinflammatory cytokines in patients with gastrointestinal cancer Int J Clin Pract 2004 58 545 549 15311551 10.1111/j.1368-5031.2004.00149.x
Tonon R D'Andrea P Interleukin-1 beta increases the functional expression of connexin 43 in articular chondrocytes: evidence for a Ca2+-dependent mechanism J Bone Miner Res 2000 15 1669 1677 10976987
Pimentel RC Yamada KA Kleber AG Saffitz JE Autocrine regulation of myocyte Cx43 expression by VEGF Circ Res 2002 90 671 677 11934834 10.1161/01.RES.0000014823.75393.4D
Graeber SH Hulser DF Connexin transfection induces invasive properties in HeLa cells Exp Cell Res 1998 243 142 149 9716458 10.1006/excr.1998.4130
Takeichi M Cadherins in cancer: implications for invasion and metastasis Curr Opin Cell Biol 1993 5 806 811 8240824 10.1016/0955-0674(93)90029-P
Behrens J Vakaet L Friis R Winterhager E Van RF Mareel MM Birchmeier W Loss of epithelial differentiation and gain of invasiveness correlates with tyrosine phosphorylation of the E-cadherin/beta- catenin complex in cells transformed with a temperature-sensitive v-SRC gene J Cell Biol 1993 120 757 766 8425900 10.1083/jcb.120.3.757
Nieman MT Prudoff RS Johnson KR Wheelock MJ N-cadherin promotes motility in human breast cancer cells regardless of their E-cadherin expression J Cell Biol 1999 147 631 643 10545506 10.1083/jcb.147.3.631
Hazan RB Phillips GR Qiao RF Norton L Aaronson SA Exogenous expression of N-cadherin in breast cancer cells induces cell migration, invasion, and metastasis J Cell Biol 2000 148 779 790 10684258 10.1083/jcb.148.4.779
Xu X Li WE Huang GY Meyer R Chen T Luo Y Thomas MP Radice GL Lo CW Modulation of mouse neural crest cell motility by N-cadherin and connexin 43 gap junctions J Cell Biol 2001 154 217 230 11449002 10.1083/jcb.200105047
Bazarbachi A Abou Merhi R Gessain A Tahouk R El-Khoury H Nasr R Gout O Sulhian R Homaidan F de The H El-Sabban ME Human T-cell lymphotropic virus type I-infected cells extravasate through the endothelial barrier by a local angiogenesis-like mechanism Cancer Res 2004 64 2039 2046 15026341
Carter TD Chen XY Carlile G Kalapothakis E Ogden D Evans WH Porcine aortic endothelial gap junctions: identification and permeation by caged InsP3 J Cell Sci 1996 109 1765 1773 8832399
Vaca L Kunze DL IP3-activated Ca2+ channels in the plasma membrane of cultured vascular endothelial cells Am J Physiol 1995 269 C733 C738 7573404
Lewalle JM Cataldo D Bajou K Lambert CA Foidart JM Endothelial cell intracellular Ca2+ concentration is increased upon breast tumor cell contact and mediates tumor cell transendothelial migration Clin Exp Metastasis 1998 16 21 29 9502074 10.1023/A:1006555800862
Moore TM Norwood NR Creighton JR Babal P Brough GH Shasby DM Stevens T Receptor-dependent activation of store-operated calcium entry increases endothelial cell permeability Am J Physiol Lung Cell Mol Physiol 2000 279 L691 L698 11000129
Garcia JG Schaphorst KL Regulation of endothelial cell gap formation and paracellular permeability J Investig Med 1995 43 117 126 7735915
Iino M Molecular basis of spatio-temporal dynamics in inositol 1,4,5-trisphosphate-mediated Ca2+ signalling Jpn J Pharmacol 2000 82 15 20 10874583 10.1254/jjp.82.15
Chambers AF Schmidt EE MacDonald IC Morris VL Groom AC Early steps in hematogenous metastasis of B16F1 melanoma cells in chick embryos studied by high-resolution intravital videomicroscopy J Natl Cancer Inst 1992 84 797 803 1573668
Koop S MacDonald IC Luzzi K Schmidt EE Morris VL Grattan M Khokha R Chambers AF Groom AC Fate of melanoma cells entering the microcirculation: over 80% survive and extravasate Cancer Res 1995 55 2520 2523 7780961
Koop S Schmidt EE MacDonald IC Morris VL Khokha R Grattan M Leone J Chambers AF Groom AC Independence of metastatic ability and extravasation: metastatic ras-transformed and control fibroblasts extravasate equally well Proc Natl Acad Sci U S A 1996 93 11080 11084 8855312 10.1073/pnas.93.20.11080
Kamibayashi Y Oyamada Y Mori M Oyamada M Aberrant expression of gap junction proteins (connexins) is associated with tumor progression during multistage mouse skin carcinogenesis in vivo Carcinogenesis 1995 16 1287 1297 7788845
| 15987459 | PMC1175070 | CC BY | 2021-01-04 16:54:50 | no | Breast Cancer Res. 2005 May 13; 7(4):R522-R534 | utf-8 | Breast Cancer Res | 2,005 | 10.1186/bcr1042 | oa_comm |
==== Front
Breast Cancer ResBreast Cancer Research1465-54111465-542XBioMed Central London bcr10431598746310.1186/bcr1043Research ArticleAlternative initiation and splicing in dicer gene expression in human breast cells Irvin-Wilson Charletha V [email protected] Gautam [email protected] Division of Cancer Biology, Department of Biomedical Sciences, Meharry Medical College, Nashville, TN 37208, USA2005 16 5 2005 7 4 R563 R569 21 12 2004 7 3 2005 25 3 2005 14 4 2005 Copyright © 2005 Irvin-Wilson and Chaudhuri.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Introduction
Dicer is a ribonuclease that mediates RNA interference both at the transcriptional and the post-transcriptional levels. Human dicer gene expression is regulated in different tissues. Dicer is responsible for the synthesis of microRNAs and short temporal (st)RNAs that regulate the expression of many genes. Thus, understanding the control of the expression of the dicer gene is essential for the appreciation of double-stranded (ds)RNA-mediated pathways of gene expression. Human dicer mRNA has many upstream open reading frames (uORFs) at the 5'-leader sequences (the nucleotide sequence between the 5'-end and the start codon of the major ORF), and we studied whether these elements at the 5'-leader sequences regulate the expression of the dicer gene.
Method
We determined the 5'-leader sequences of the dicer mRNAs in human breast cells by 5'-RACE and S1-nuclease protection analysis. We have analyzed the functions of the 5'-leader variants by reporter gene expression in vitro and in vivo.
Results
We found that the dicer transcripts in human breast cells vary in the sequence of their 5'-leader sequences, and that alternative promoter selection along with alternative splicing of the 5'-terminal exons apparently generate these variations. The breast cell has at least two predominant forms of dicer mRNAs, one of which has an additional 110 nucleotides at the 5'-end. Sequence comparison revealed that the first 80 nucleotides of these mRNA isoforms are encoded by a new exon located approximately 16 kb upstream of the reported start site. There are 30 extra nucleotides added to the previously reported exon 1. The human breast cells studied predominantly express two 5'-leader variants of dicer mRNAs, one with the exons 2 and 3 (long form) and the other without them (short form). By reporter gene expression analysis we found that the exon 2 and 3 sequences at the 5'-leader sequences are greatly inhibitory for the translation of the mRNA into protein.
Conclusion
Dicer gene expression in human breast cells is regulated by alternative promoter selection to alter the length and composition of the 5'-leader sequence of its mRNA. Furthermore, alternative splicing of its exon 2 and 3 sequences of their pre-mRNA creates a more translationally competent mRNA in these cells.
==== Body
Introduction
RNA interference (RNAi), a process of silencing gene expression, involves the generation of short, double-stranded RNA (dsRNA) molecules by an enzyme called dicer, which cleaves RNA duplexes into 21–23 base-pair oligomers [1-8]. These oligomers are called, depending on their end-point functions, small interfering RNAs (siRNA), microRNA (miRNA) or short temporal RNA (stRNA) [9]. These small RNA molecules cause sequence-specific post-transcriptional gene silencing by guiding an endonuclease, the RNAi-induced silencing complex (RISC), to mRNA [10,11]. This ubiquitous process has also been recently reported [12-15] in human cells to induce transcriptional silencing through promoter methylation.
Dicer gene expression is regulated in different tissues in humans [16]. Because dicer catalyzes the biosynthesis of miRNAs and stRNAs that in turn regulate the expression of many genes, it is likely that the expression of the dicer gene itself is a highly regulated process [9-11]. While studying the published 5'-leader sequences of human dicer transcripts we noticed that it is infested with many upstream open reading frames (uORF) and out of frame AUG codons [16]. To evaluate whether the 5'-untranslated region of the dicer transcript is in part responsible for the regulation of the dicer transcripts in human breast cells we have amplified, cloned and functionally characterized the 5'-leader sequences of human dicer transcripts from these cells. We report here that the dicer gene in breast cells is transcribed from a far upstream promoter in chromosome 14 and the sequence of the 5'-leader sequences determines the translatability of the dicer transcript and is dictated by alternative splicing of the 5'-exons.
Materials and methods
Cell culture
We used a series of commercially available human lines of breast cells, including human mammary epithelial (HME) cells (Clonetics, purchased through Fisher Scientific, Pittsburgh, PA, USA), MDA-MB-231, MCF-7, MDA-MB-468, MCF-10A, and BT549. We also used non-breast cells such as HeLa and HepG2. All cells, other than the HME cells, were purchased from American Type Culture Collection (ATCC, Manassas, VA, USA). HMEC cells were grown in medium purchased from Clonetics under their recommended conditions. Human breast carcinoma MDA-MD-231 and MDA-MD-468 cells were maintained in Leibovitz's L-15 medium supplemented with 1% antibiotic/antimycotic and 10% fetal bovine serum. MCF-10A cells were maintained in a 1:1 mixture of Ham's F12 medium, Dulbecco's modified Eagle's medium supplemented with 1% antibiotic/antimycotic, 0.098 mg/ml cholera toxin, 0.02 μg/μl epidermal growth factor, 0.5 μg/ml hydrocortisone, and 10% horse donor heard serum. BT-549 cells were maintained following standard ATCC recommended media. All cells were maintained in a humidified CO2 (5%) incubator at 37°C. Other cells were maintained and grown in ATCC recommended media and conditions [17,18].
5'-RACE
Complementary DNAs (cDNA) were made from total RNA (5 μg) using random primers following standard protocols [18,19]. The cDNAs were dC-tailed and the 5'-ends of the dicer mRNAs were amplified using a dicer gene specific primer (5'-AGTTGACCAAGAACACCG-3'), and the Abridged Anchor Primer (AAP, Invitrogen, Carlsbad, CA, USA) following 5'-rapid amplification of cDNA ends (RACE) analysis protocols from Invitrogen. Amplicons were re-amplified successively using nested dicer gene specific primers (5'-TGACCAAGAACACCGTCC-3', and 5'-AAATGTCTTCCCTGAGCC-3'), and AUAP and UAP, respectively. All PCRs were done using suggested thermocycler conditions of the 5'-RACE protocol (Invitrogen). 5'-RACE products were cloned in the pCRII-Topo vector following TOPO-TA cloning protocols (Invitrogen) and the nucleotide sequences of the cloned inserts were determined by automated DNA sequencing [19].
S1-nuclease protection assay
Oligonucleotides that were biotinylated at the 5'-end (5'-CACAGCATGCCCAAGCTT CTGCTCTCAAAATGCTGATTCTAAGTTC-3', and 5'-GCATTTTTGTTCTAGCACAGC TTACCTTCCCACTCGCCTGCGTTTC-3') spanning the exon 2 and exon 3 boundary of the long 5'-variant and the exon 1 and exon 4 boundary of the short 5'-variant of the dicer mRNA, respectively, were custom synthesized and gel-purified (Invitrogen). The standard S1-nuclease protection protocol was followed [19] with the following modifications: total cellular RNA (10 μg) was co-precipitated with each probe (10 pmol) and hybridized at 65°C for 15 h. S1-nuclease (Promega, Madison, WI, USA) was used at a concentration of 500 units/ml for 90 min at 37°C. Products were analyzed in a 15% TBE-urea gel, electrophoretically transferred to Zeta Probe blotting membrane (BioRad, Hercules, CA. USA) and biotinylated protected probes were detected using the North2South HRP Detection protocol (Pierce, Rockford, IL, USA).
5'Leader sequence/Renilla luciferase constructs
The long and short form dicer 5'-leader sequences were amplified using the forward primer 5'-GCGGAAGTGGGTGTTTGTTATTTCC-3' and the reverse primer 5'-GGATCATAAACTTTCGAAGTCATTGCATTTTTGTTCTAGCACAGC-3'. Gene splicing by overlapping extension (SOE) was used to fuse dicer variants to Renilla luciferase that was amplified with forward primer 5'-GCTGTGCTAGAACAAAAATGCAATGACTTCGAAAGTTTATGATCC-3' and reverse primer 5'-CTCGAAGCGGCCGCTCTAG-3' [20]. The Renilla luciferase ORF alone was amplified using forward primer 5'-ATGACTTCGAAAGTTTATGATCC-3' and reverse primer 5'-CTCGAAGCGGCCGCTCTAG-3'. SOEing products were then ligated into the pCRII-Topo vector (Invitrogen), digested with EcoRI (Promega) and cloned at the EcoRI site of pcDNA3.1(+) (Invitrogen) and sequenced with T7 primer to determine the orientation of the cloned insert. It is anticipated from the vector information (Invitrogen) that the Renilla luciferase transcript from these constructs will have a 133 nucleotide vector derived sequence before the 5'-leader sequences from the dicer gene at their 5'-end. Thus, the nucleotide sequence that is common to the control and the experimental transcripts is 5'-TAGAGAACCCACTGCTTACTGGCTTATCGAAATTAATACGACTCACTATAGGGAGACCCAAGCTGGCTAGCGTTTAAACTTAAGCTTGGTACCGAGCTCGGATCCACTAGTCCAGTGTGGTGGAATTCGGCTT-3'. This sequence does not have any ATG codon.
Transfection and luciferase assay
Cells were seeded at 80% confluency in a 24-well plate for 24 h in their growth media before co-transfection with one of the pcDNA3.1(+) test plasmid constructs and pGL3-control plasmid. The latter was used as a transfection normalization control. Plasmids were mixed at 0.5 μg per well and transfection was done using the Lipofectamine Plus transfection reagents (Invitrogen) using the protocol suggested by the supplier. After 20 h of incubation in complete medium at 37°C, the cells were lysed in 100 μl passive lysis buffer (Promega) and 5–20 μl of the lysate was assayed [18] for firefly luciferase as well as Renilla luciferase activities using Dual Luciferase Assay Reagents (Promega) following suggested protocols [18]. Renilla luciferase activity was normalized with respect to firefly luciferase activity and presented as a ratio (relative light units). Protein contents of the extract, when needed, were determined using RC-DC reagents and protocol from BioRad Laboratories Hercules, CA, USA [18].
In vitro transcription and translation
The pcDNA3.1(+) constructs containing the cloned inserts also have a T7 RNA polymerase promoter (Invitrogen). To evaluate whether there is any effect of the dicer 5'-leader sequence insert at the EcoRI site in proper orientation on the translatability of the mRNA, we used the T7 RNA polymerase/rabbit reticulocyte lysate in vitro transcription/translation system (TNT, Promega). Each plasmid DNA (1 μg) was added to 40 μl of TNT Quick Master Mix containing 1 mM of methionine (Promega). The reactions were incubated at 30°C for 75 min then cooled to 4°C. The TNT reactions were diluted 1:2 with 1x passive lysis buffer (Promega) and incubated at room temperature for 15 min. An aliquot (5–20 μl) was assayed in 100 μl of Renilla Luciferase Assay Substrate (Promega).
Statistical procedures
Each data set is presented as mean ± SEM (N = 12). Statistical significance of a difference between two series of data was tested by determining the P value [21]. If the P value was less than 0.05, the difference was considered to be significant [21].
Results and discussion
Dicer mRNA has a new exon at the 5'-end
The expression of the dicer gene is differentially regulated in many organisms that have the RNAi mechanism [22-26]. In humans the expression of this gene is highly tissue specific [16] and the mechanisms that regulate it are not known. Initial characterization of dicer mRNA indicated an abundance of overlapping and non-overlapping uORFs at the 5'-leader sequences [16]. To evaluate whether the 5'-leader sequences of dicer mRNA may be a regulatory factor for dicer gene expression, we amplified, cloned and characterized the 5'-leader sequences of these mRNAs from human breast cells. We found that dicer mRNAs have at least two major 5'-leader sequences (Figs 1, 2, 3). Fig. 1 shows our 5'-RACE data from MDA-MB-231 and MDA-MB-468 cells. Similar experiments were done with BT549, MCF7, MCF10A and HMEC cells with similar results (not shown). In fact, we also found these two major forms in HepG2 and HeLa cells (data not shown). We purified the PCR products as a whole and cloned all the DNA molecules. Nucleotide sequencing of more than 40 clones revealed only two types of sequences from dicer mRNAs (Fig. 2). The other bands in the 5'-RACE products were thus PCR artifacts deriving from mis-priming. The 5'-leader sequences in the characterized dicer mRNAs are termed 'long' (320 nucleotides) and 'short' (149 nucleotides) forms because of their lengths. The human dicer gene was originally cloned from HepG2 cells and was shown to be located on chromosome 14q31 near marker D14S605 [16]. The gene was reported to have 28 exons and 27 introns [16]. Our characterization of the 5'-leader sequences of the dicer mRNA variants from breast cells indicate a new exon at the 5'-end and 30 extra nucleotides added to the original first exon (Figs 1, 2, 3). Alignment of the nucleotide sequence to the chromosome 14 sequence shows that the new exon is transcribed from a segment of the chromosome approximately 16 kb upstream of the transcription start site previously predicted for human dicer mRNA. Because we have not done extensive studies with different non-breast cell types, we cannot say at this point that the transcript originally reported for HepG2 or other types of cells does not exist. It is possible that human dicer gene is transcribed from different promoters in different tissues. Selection of this alternative promoter in breast cells leads to the formation of a new intron 1 that is 16,184 base pairs in length (Fig. 3). Alternative promoter selection is a known mode of regulation of the expression of mammalian genes [27-29].
The long form of the 5'-leader sequences contains nine upstream AUG codons (Fig. 2). The short form of the 5'-leader sequence is the alternatively spliced form of the long form. In the short form exon 1 is directly joined to exon 4 and exons 2 and 3 are spliced out (Figs 2 and 3); the number of upstream AUG codons is decreased to five due to this alternative splicing (Fig. 2). Upstream AUG codons often are reported to slow down the rate of translation of an mRNA [30-37]. We tested whether the decrease in the upstream AUGs has any effect on the translatability of the dicer mRNAs (see below).
To ensure that the existence of the different 5'-leader sequences in dicer mRNA as revealed by the 5'-RACE analysis was not a PCR artifact, and also to determine relative quantities of each of the mRNA isoforms, we verified the levels of each of the isoforms by S1-nuclease protection assays (Fig. 4). Data from RNAs isolated from MCF-10A, MDA-MB-231, MDA-MB-468 and BT-549 cells are shown in Fig. 4. We used beta-actin RNA level as a loading control. Both the short and long 5'-leader sequence variants were found in these breast cell lines (not shown). The relative amounts of the two forms vary from cell to cell, indicating that alternative splicing of dicer pre-mRNA could be a form of regulation of dicer gene expression.
Translation of the short 5'-leader form of the dicer mRNA is more efficient
To evaluate whether the altered length of the 5'-leader sequences has any effect on the ability of the mRNA to be translated we tested reporter gene expression from constructs that had either the short or the long 5'-leader sequence (Fig. 5). The data shown are with the control constructs, which have no added 5'-leader sequence. We also performed additional experiments with constructs that have other unrelated DNA segments of similar lengths. We found no significant differences between these controls (data not shown). Transient transfection of the breast or non-breast cells with the plasmid constructs was supposed to produce Renilla luciferase mRNAs with the general structure shown in Fig. 5. We lysed the transfected cells after 20 h and assayed for Renilla luciferase activity in the cell lysate. The plasmid that constitutively expresses firefly luciferase from SV40 late promoter was used as transfection control and the Renilla luciferase data were normalized with respect to the firefly luciferase activity in the extract. Tthe construct with the long form of the 5'-leader sequence expressed no significant Renilla luciferase activity compared to the controls (Fig. 6) in all the cells tested. On the other hand, the construct with the short form of the 5'-leader sequence expressed significant Renilla luciferase activity (Fig. 6). The activity from the construct with the short form of the 5'-leader sequence was much less than from the constructs with no added 5'-leader sequence, indicating that the residual AUG codons in the short form of the 5'-leader sequence may still be negatively regulating the translation and/or stability of the mRNA. To verify whether inclusion of different 5'-leader sequences in the reporter plasmid has any effect on the level of Renilla luciferase mRNA, we performed RT-PCR analysis. Our data suggest that irrespective of whether the construct expresses luciferase activity, there is no significant difference in the luciferase mRNA levels (Fig. 7). The effect of the added 5'-leader sequences thus may be at the translational level. We further verified this by in vitro transcription/translation with the plasmid constructs followed by reporter gene expression analysis (Fig. 8). This experiment also suggests that perhaps the consequence of altering the 5'-leader sequence of human dicer transcripts is the altered translation of the mRNA into dicer protein. Taken together, these observations imply that alternative promoter use and alternative splicing of 5'-exons are important for the regulation of human dicer gene expression.
Conclusion
In human breast cells, the dicer gene is transcribed from a promoter that is more than 16 kbp upstream of the initiation site reported for this gene from non-breast cells [16]. This alternative promoter selection modifies the length and composition of the 5'-leader sequences of its mRNA. Furthermore, alternative splicing of the exon 2 and 3 sequences of its pre-mRNA creates a more translationally competent mRNA in these cells. Breast cell dicer mRNAs have a high number of upstream AUG codons with in frame stop codons. Translational efficiency and/or stability of the RNA are decreased due to this type of 5'-leader sequence [34-37]. The non-sense mediated degradation pathway of mRNA decay may also be stimulated by these uORFs [38,39]. It is interesting to note that there is an association between active dicer mRNA expression and the invasiveness of the breast cell line. Recently we found that the zinc finger repressor protein SLUG may have the determinative role in the invasiveness of the breast cell [40,41]. We are exploring whether dicer regulates SLUG gene expression in human breast cells. The mechanisms by which the 5'-leader sequences of the dicer mRNAs in breast cells regulate the expression of this gene are yet to be determined.
Abbreviations
ATCC = American Type Culture Collection; dsRNA = double-stranded RNA; HME = human mammary epithelial; kb = kilobases; miRNA = microRNA; nt = nucleotide; PCR = polymerase chain reaction; RACE = rapid amplification of cDNA ends; RISC = RNAi-induced silencing complex; RNAi = RNA interference; siRNA = small interfering RNA; SOE = splicing by overlapping extension; stRNA = short temporal RNA; uORF = upstream open reading frame.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
CVIW executed the experiments described in the manuscript and wrote the initial drafts of the manuscript. GC conceived and coordinated the study. All authors read and approved the final manuscript.
Acknowledgements
Supported by the MMC/VICC cancer partnership grant #1U54CA091408-010003 from NCI and the DOD grant #DAMD17-00-1-0341 to GC and graduate training fellowship #1T32GM062758-01 from NIH to CVIW.
Figures and Tables
Figure 1 Ethidium bromide stained agarose gel showing the analysis of the 5'-RACE products from human dicer mRNAs from breast cells. Lane M, 1 kb plus DNA ladder (Invitrogen); lane 1, cDNA from MDA-MB-231; lane 2, cDNA from MDA-MB-468. The band at approximately 550 bp is for the unspliced long form of the 5'-leader sequence. The band at approximately 380 bp is the short spliced form.
Figure 2 Nucleotide sequences of the (a) long form and (b) short form of the 5'-leaders derived from the 5'-RACE products (Entrez Accession numbers AY845867 and AY845868). The exon junctions are indicated by two vertical lines. The upstream AUGs and corresponding stop codons are numbered and underscored. The AUG at the 3'-end is the putative start codon.
Figure 3 Cartoon showing the possible splicing of the 5'-exons of the human dicer pre-mRNA to produce the short and long forms of mRNAs.
Figure 4 S1-Nuclease protection analysis to verify the long and short forms of dicer mRNA. Autoradiograms showing the relative quantities of (a) short from, (b) long form and (c) β-actin mRNAs. See text for details. Lanes from left to right: RNA from MCF-10A cells; RNA from MDA-MB-231 cells; RNA from MDA-MB-468 cells; RNA from BT-549 cells; 50% of the probe input; and biotinylated oligonucleotide ladder. The lower band in the long form second and third lane from left may be the 3'-truncated (2 nt) form of the biotinylated probe due to the possible degradation of the fraying 3'-end of probe by S1 nuclease.
Figure 5 The map of the putative luciferase mRNA synthesized from the plasmid inside the transfected cells. The 5'-end 133 nt come from the vector. The lengths of the dicer 5'-leader sequence vary. The control does not have any insert. The 3'-untranslated region also comes from the vector. Nt, nucleotides; UTR, untranslated region.
Figure 6 Effect of dicer 5'-leader sequences upstream of the luciferase coding sequence on the expression of relative luciferase activities in BT549, MDA-MB-468, HeLa and HepG2 cells. Luciferase activity is expressed as relative light units (RLU) after normalization with firefly luciferase activity. Results are means of data from 12 different experiments ± standard errors. The differences were statistically significant (p < 0.001). UTR, untranslated region.
Figure 7 Verification of Renilla luciferase mRNA levels by RT-PCR in transfected BT549 cells. β-Actin mRNA levels were used as control. Lane M, 1 kb plus DNA ladder (Invitrogen; lowest band is 200 bp); lane 1–3, RNA isolated from BT549 cells transfected with pcDNA3.1(+), pcDNA3.1(+)-short form 5'-leader or pcDNA3.1(+)-long form plasmid DNA, respectively.
Figure 8 In vitro transcription and translation of the Renilla luciferase gene from the plasmid constructs and evaluation of the luciferase activity. Results are means of data from 12 different experiments ± standard errors. The differences were statistically significant (p < 0.001). UTR, untranslated region.
==== Refs
Campbell TN Choy FY RNA interference: past, present and future Curr Issues Mol Biol 2005 7 1 6 15580776
Nesterova M Cho-Chung YS Killing the messenger: antisense DNA and siRNA Curr Drug Targets 2004 5 683 689 15578949 10.2174/1389450043345137
Manoharan M RNA interference and chemically modified small interfering RNAs Curr Opin Chem Biol 2004 8 570 579 15556399 10.1016/j.cbpa.2004.10.007
Banan M Puri N The ins and outs of RNAi in mammalian cells Curr Pharm Biotechnol 2004 5 441 450 15544492 10.2174/1389201043376643
Silva J Chang K Hannon GJ Rivas FV RNA-interference-based functional genomics in mammalian cells: reverse genetics coming of age Oncogene 2004 23 8401 8409 15517022 10.1038/sj.onc.1208176
Stevenson M Therapeutic potential of RNA interference N Engl J Med 2004 351 1772 1777 15496626 10.1056/NEJMra045004
Lingel A Izaurralde E RNAi: finding the elusive endonuclease RNA 2004 10 1675 1679 15496518 10.1261/rna.7175704
Hannon GJ Rossi JJ Unlocking the potential of the human genome with RNA interference Nature 2004 431 371 378 15372045 10.1038/nature02870
Nelson P Kiriakidou M Sharma A Maniataki E Mourelatos Z The microRNA world: small is mighty Trends Biochem Sci 2003 28 534 540 14559182 10.1016/j.tibs.2003.08.005
Tijsterman M Plasterk RH Dicers at RISC; the mechanism of RNAi Cell 2004 117 1 3 15066275 10.1016/S0092-8674(04)00293-4
Carmell MA Hannon GJ RNase III enzymes and the initiation of gene silencing Nat Struct Mol Biol 2004 11 214 218 14983173 10.1038/nsmb729
Kawasaki H Taira K Induction of DNA methylation and gene silencing by short interfering RNAs in human cells Nature 2004 431 211 217 15311210 10.1038/nature02889
Morris KV Chan SWL Jacobson SE Looney DJ Small-interfering RNA-induced transcriptional gene silencing in human cells Science 2004 305 1289 1292 15297624 10.1126/science.1101372
Lippman Z Martienssen R The role of RNA interference in heterochromatic silencing Nature 2004 431 364 370 15372044 10.1038/nature02875
Matzke M Aufsatz W Kanno T Daxinger L Papp I Mette MF Matzke AJ Genetic analysis of RNA-mediated transcriptional gene silencing Biochim Biophys Acta 2004 1677 129 141 15020054
Matsuda S Ichigotani Y Okuda T Irimura T Nakatsugawa S Hamaguchi M Molecular cloning and characterization of a novel human gene (HERNA) which encodes a putative RNA-helicase Biochim Biophys Acta 2000 1490 163 169 10786632
Sharan C Hamilton N Parl AK Singh PK Chaudhuri G Identification and characterization of a transcriptional silencer upstream of human BRCA2 gene Biochem Biophys Res Commun 1999 265 285 290 10558858 10.1006/bbrc.1999.1652
Tripathi MK Chaudhuri G Down-regulation of UCRP and UBE2L6 in BRCA2 Knocked-Down human breast cells Biochem Biophys Res Commun 2005 328 43 48 15670748 10.1016/j.bbrc.2004.12.142
Sambrook J Russell DW Molecular Cloning: A Laboratory Manual 2001 Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY
Dieffenbach CW Dveksler GS PCR Primer: A Laboratory Manual 1995 Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press
Campbell MJ Machin D Medical Statistics: A Commonsense Approach 1994 2 New York: John Wiley
Hammond SM Bernstein E Beach D Hannon GJ An RNA-directed nuclease mediates post-transcriptional gene silencing in Drosophila cells Nature 2000 404 293 296 10749213 10.1038/35005107
Liu Q Rand TA Kalisas S Du F Kim HE Smith DP Wang X R2D2, a bridge between the initiation and effector steps of the Drosophila RNAi pathway Science 2003 301 1921 1925 14512631 10.1126/science.1088710
Banerjee D Slack F Control of developmental timing by small temporal RNAs: a paradigm for RNA-mediated regulation of gene expression Bioessays 2002 24 119 129 11835276 10.1002/bies.10046
Ketting RF Fischer SE Berstein E Sijen T Hannon GJ Plasterk RH Dicer functions in RNA interference and in synthesis of small RNA involved in developmental timing in C. elegans Genes Dev 2001 15 2654 2659 11641272 10.1101/gad.927801
Grishok A Pasquinelli AE Conte D Li N Parrish S Ha I Baillie DL Fire A Ruvkun G Mello CC Genes and mechanisms related to RNA interference regulate expression of small temporal RNAs that control C. elegans developmental timing Cell 2001 106 23 34 11461699 10.1016/S0092-8674(01)00431-7
Itani OA Campbell JR Herrero J Snyder PM Thomas CP Alternate promoters and variable splicing lead to hNedd4-2 isoforms with a C2 domain and varying number of WW domains Am J Physiol Renal Physiol 2003 285 916 929
Munoz-Sanjuan I Smallwood PM Nathans J Isoform diversity among fibroblast growth factor homologous factors is generated by alternative promoter usage and differential splicing J Biol Chem 2000 275 2589 2597 10644718 10.1074/jbc.275.4.2589
Ball HJ Shine J Herzog H Multiple promoters regulate tissue-specific expression of the human NPY-Y1 receptor gene J Biol Chem 1995 270 27272 27276 7592987 10.1074/jbc.270.45.27272
Jin X Turcott E Englehardt S Mize GJ Morris DR The two upstream open reading frames of oncogene mdm2 have different translational regulatory properties J Biol Chem 2003 278 25716 25721 12730202 10.1074/jbc.M300316200
Child SJ Miller MK Geballe AP Translational control by an upstream open reading frame in the HER-2/neu transcript J Biol Chem 1999 274 24335 24341 10446211 10.1074/jbc.274.34.24335
Morris DR Geballe AP Upstream open reading frames as regulators of mRNA translation Mol Cell Biol 2000 20 8635 8642 11073965 10.1128/MCB.20.23.8635-8642.2000
Vilela C McCarthy JE Regulation of fungal gene expression via short open reading frames in the mRNA 5'untranslated region Mol Microbiol 2003 49 859 867 12890013 10.1046/j.1365-2958.2003.03622.x
Lu PD Harding HP Ron D Translation reinitiation at alternative open reading frames regulates gene expression in an integrated stress response J Cell Biol 2004 167 27 33 15479734 10.1083/jcb.200408003
Muller EA Danner DJ Tissue-specific translation of murine branched-chain alpha-ketoacid dehydrogenase kinase mRNA is dependent upon an upstream open reading frame in the 5'-untranslated region J Biol Chem 2004 279 44645 44655 15302860 10.1074/jbc.M406550200
Sapetschnig A Koch F Rischitor G Mennenga T Suske G Complexity of translationally controlled transcription factor Sp3 isoform expression J Biol Chem 2004 279 42095 42105 15247228 10.1074/jbc.M404989200
Meijer HA Thomas AA Control of eukaryotic protein synthesis by upstream open reading frames in the 5'-untranslated region of an mRNA Biochem J 2002 367 1 11 12117416 10.1042/BJ20011706
Singh G Lykke-Andersen J New insights into the formation of active nonsense-mediated decay complexes Trends Biochem Sci 2003 28 464 466 13678954 10.1016/S0968-0004(03)00176-2
Holbrook JA Neu-Yilik G Hentze MW Kulozik AE Nonsense-mediated decay approaches the clinic Nat Genet 2004 36 801 808 15284851 10.1038/ng1403
Tripathi MK Misra S Chaudhuri G Negative regulation of the expressions of cytokeratins 8 and 19 by SLUG repressor protein in human breast cells Biochem Biophys Res Commun 2005 329 508 515 15737616 10.1016/j.bbrc.2005.02.006
Tripathi MK Misra S Khedkar SV Hamilton N Irvin-Wilson C Sharan C Sealy L Chaudhuri G Regulation of BRCA2 gene expression by the SLUG repressor protein in human breast cells J Biol Chem 2005 280 17163 17171 15734731 10.1074/jbc.M501375200
| 15987463 | PMC1175071 | CC BY | 2021-01-04 16:04:33 | no | Breast Cancer Res. 2005 May 16; 7(4):R563-R569 | utf-8 | Breast Cancer Res | 2,005 | 10.1186/bcr1043 | oa_comm |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.